This volume re-examines the topic of economic growth in Europe after the Second World War. The contributors approach th...
110 downloads
1686 Views
8MB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
This volume re-examines the topic of economic growth in Europe after the Second World War. The contributors approach the subject armed not only with new theoretical ideas, but also with the experience of the 1980s on which to draw. The analysis is based on both applied economics and economic history. Thus, while the volume is greatly informed by insights from growth theory, emphasis is given to the presentation of chronological and institutional detail. The case study approach and the adoption of a longer-run perspective than is normal for economists, allow new insights to be obtained. Individual chapters cover Belgium, Denmark, Germany, Spain, France, Ireland, Italy, the Netherlands, Portugal, Sweden and the UK. Further chapters explore general European institutional arrangements and historical circumstances. The result is a genuinely comparative picture of postwar growth, with insights that do not emerge from standard cross-section regressions based on the post-1960 period.
Economic growth in Europe since 1945
Centre for Economic Policy Research The Centre for Economic Policy Research is a network of over 250 Research Fellows, based primarily in European universities. The Centre coordinates its Fellows' research activities and communicates their results to the public and private sectors. CEPR is an entrepreneur, developing research initiatives with the producers, consumers and sponsors of research. Established in 1983, CEPR is a European economics research organization with uniquely wide-ranging scope and activities. CEPR is a registered educational charity. Institutional (core) finance for the Centre is provided by major grants from the Economic and Social Research Council, under which an ESRC Resource Centre operates within CEPR; the Esmee Fairbairn Charitable Trust; the Bank of England; 17 other central banks and 40 companies. None of these organizations gives prior review to the Centre's publications, nor do they necessarily endorse the views expressed therein. The Centre is pluralist and non-partisan, bringing economic research to bear on the analysis of medium- and long-run policy questions. CEPR research may include views on policy, but the Executive Committee of the Centre does not give prior review to its publications, and the Centre takes no institutional policy positions. The opinions expressed in this volume are those of the authors and not those of the Centre for Economic Policy Research. Executive Committee Chairman Vice-Chairman
Anthony Loehnis Guillermo de la Dehesa
Jan Bielecki Honor Chapman Quentin Davies Sheila Drew Smith Otmar Issing
Philippe Lagayette Peter Middleton David Miliband Mario Sarcinelli Catherine Sweet
Officers Director Deputy Director
Richard Portes Stephen Yeo 1 November 1995
Economic growth in Europe since 1945 Edited by
NICHOLAS CRAFTS and GIANNI TONIOLO
CAMBRIDGE UNIVERSITY PRESS
PUBLISHED BY THE PRESS SYNDICATE OF THE UNIVERSITY OF CAMBRIDGE
The Pitt Building, Trumpington Street, Cambridge, United Kingdom CAMBRIDGE UNIVERSITY PRESS
The Edinburgh Building, Cambridge CB2 2RU, UK 40 West 20th Street, New York, NY 10011-4211, USA 477 Williamstown Road, Port Melbourne, VIC 3207, Australia Ruiz de Alarcon 13, 28014 Madrid, Spain Dock House, The Waterfront, Cape Town 8001, South Africa http: //www. Cambridge. org © Centre for Economic Policy Research 1996 This book is in copyright. Subject to statutory exception and to the provisions of relevant collective licensing agreements, no reproduction of any part may take place without the written permission of Cambridge University Press. First published 1996 Reprinted 1999, 2002 A catalogue record for this book is available from the British Library Library of Congress Cataloguing in Publication data applied for ISBN 0 521 49627 6 hardback ISBN 0 521 49964 X paperback
Transferred to digital printing 2003
Contents
List of List of tables Preface List of contributors
figures
Postwar growth: an overview Nicholas Crafts and Gianni Toniolo 1 Introduction 2 Main trends in postwar European growth 3 Previous interpretations of European postwar growth 4 Recent developments in growth theory 5 Europe's postwar growth in a long-run perspective: some quantitative aspects 6 The years of high growth 7 Slowdown and instability 8 Factors in the performance of individual countries 9 Concluding comments Institutions and economic growth: Europe after World War II Barry Eichengreen 1 Introduction 2 The model 3 The structure of domestic institutions 4 The evolution of domestic institutions 5 The structure of international institutions 6 The decline of the postwar settlement 7 Conclusion
vn
page xii xv xxi xxiii
1 1 2 8 14 16 20 25 27 31 38 38 43 45 50 53 58 65
viii
Contents 3
4
5
6
7
The varieties of Eurosclerosis: the rise and decline of nations since 1982 Mancur Olson 1 Introduction 2 A recapitulation of the theory 3 Economic growth since Rise and Decline 4 Distinctive institutions and common processes 5 Distinctive institutions and inescapable logic Why the 1950s and not the 1920s? Olsonian and non-Olsonian interpretations of two decades of German economic history Karl-Heinz Paque 1 Introduction 2 An Olsonian interpretation of the two postwar records 3 A non-Olsonian mode of interpretation Convergence, competitiveness and the exchange rate Andrea Boltho 1 Introduction 2 Competitiveness and the exchange rate 3 Western Europe in the 'Golden Age' 4 Country experience 5 Conclusions Statistical appendix British economic growth since 1945: relative economic decline . . . and renaissance? Charles Bean and Nicholas Crafts 1 Introduction 2 The legacy of the 1930s and World War II 3 Reconstruction 4 The Golden Age 5 Shocks and stagflation 6 Recovery in the 1980s and its legacy 7 Bargaining models and productivity change 8 Human capital formation 9 Deindustrialization 10 Conclusions Appendix Economic growth in postwar Belgium Isabelle Cassiers, Philippe De Ville and Peter M. Solar 1 Introduction 2 Postwar economic growth: main features, structures and institutions, initial conditions 3 The phases of postwar economic growth 4 Structural change and the control of industry 5 Conclusion
73 73 74 78 81 91
95 95 96 100 107 107 108 110 116 124 126
131 131 138 140 142 147 149 153 158 159 160 162 173 173 174 180 198 201
Contents 8
9
10
11
12
France, 1945-92 Pierre Sicsic and Charles Wyplosz 1 Introduction 2 Aggregate performance 3 The legacy of the 1930s and reconstruction 4 The 'vingt glorieuses' (1954-76) 5 Shocks and stagflation in the 1970s 6 A partial recovery 7 Human capital 8 Institutions 9 Conclusion Economic growth and the Swedish model Magnus Henrekson, Lars Jonung and Joakim Stymne 1 Introduction 2 Aggregate performance 3 A review of macroeconomic policies 4 Ultimate causes of Swedish economic performance 5 Conclusions Characteristics of economic growth in the Netherlands during the postwar period Bart van Ark, Jakob de Haan and Herman J. de Jong 1 Introduction 2 The major facts 3 The legacy of the 1930s 4 World War II and the reconstruction years 5 The 'golden years' from 1950 to 1973 6 Shocks and sluggish growth during the 1970s 7 Continuity and change during the 1980s 8 Labour market and wage policies 9 The performance of the export sector and exchange rate policy 10 The public sector and economic growth 11 Conclusions
ix 210 210 211 217 219 223 226 229 232 236 240 240 242 252 256 280
290 290 292 301 302 305 309 311 312 315 317 322
Portuguese postwar growth: a global approach Jodo L. Cesar das Neves 1 Introduction 2 Postwar growth: the setting 3 Postwar growth: the phases 4 Some particular themes 5 Conclusions
329
Growth and macroeconomic performance in Spain, 1939-93 Leandro Prados de la Escosura and Jorge C. Sanz 1 Introduction 2 Spain's economic performance in the long run
355
329 329 335 346 352
355 355
x
Contents 3 The legacy of the 1930s and the Civil War (1936-9) 4 Reconstruction: Spain under autarky, 1939-59 5 The Golden Age: years of accelerated growth, 1959-75 6 Shocks and stagflation: the transition from dictatorship to democracy, 1975-85 7 Recovery of the late 1980s and its legacy: the integration of Spain into the EEC 8 Concluding remarks
361 362 369 372 375 378
13 Irish economic growth, 1945-88 Cormac OGrdda and Kevin O'Rourke 1 Introduction 2 Irish growth and the economic convergence debate 3 Irish economic history, 1945-92 4 Investment 5 Human capital and emigration 6 Trade policies 7 Rent seeking and interest groups: the political economy of growth 8 Some simple cross-section evidence 9 Conclusion
388
14 Italy Nicola Rossi and Gianni Toniolo 1 Introduction 2 The aggregate performance: an overview 3 Total factor productivity, market structure, scale economies and capacity utilization 4 The legacy of Fascism and the war 5 Setting the stage: reconstruction and stabilization 6 An 'economic miracle'? 7 The roots of the malaise, 1963-73 8 Productivity slowdown, 1973-92 9 Concluding remarks Appendix
427
15 West German growth and institutions, 1945-90 Wendy Carlin 1 Introduction 2 The growth weakness of the Weimar Republic 3 Reconstruction, 1945-61 4 Golden Age growth, 1961-73 5 Slow growth, 1973-90 6 Is there a West German economic model and, if so, has it affected growth? 7 West German growth - what does it tell us about the prospects for growth in East Germany?
388 389 398 404 407 409 413 416 419
427 428 430 438 439 441 442 445 449 450 455 455 460 463 468 473 482 489
Contents 16
17
18
An exercise in futility: East German economic growth and decline, 1945-89 Albrecht 0. Ritschl 1 Introduction 2 Macroeconomic performance: a brief overview 3 The legacy of the 1930s and the war: how bad a start? 4 The productivity gap in the making, 1945-50 5 East Germany's transition to Communism: a brief review 6 The 1950s: an East German Wirtschaftswunderi 7 2 fast 4 you: frustrated catching up, the Berlin Wall and attempted reform during the 1960s 8 The Golden Seventies: a belated Wirtschaftswundert 9 The road to bankruptcy, 1980-9 10 The aftermath of unification 11 Conclusion
xi
498 498 499 504 508 511 513 517 519 522 528 532
Postwar growth of the Danish economy Peder J. Pedersen 1 Introduction 2 The aggregate growth performance: an overview 3 The legacy from the 1930s and the war and reconstruction years 4 Growth factors and sectoral shifts in the postwar years 5 Growth and economic policy since 1950 6 Concluding comments
541
Reflections on the country studies Nicholas Crafts and Gianni Toniolo 1 Introduction 2 The idea of a 'Golden Age' 3 The accumulation of capital and the acquisition of knowledge 4 The impact of economic policy 5 Growth economics
576
Index
582
541 543 551 554 563 570
576 577 578 579 580
Figures
1.1 Trend growth of real output per person in Italy, 1870-1990 page 17 2.1 Initial (1950) GDP per man-hour and growth, 1950-73 39 2.2 Wartime and postwar (1938-50) change in GDP per man-hour and growth, 1950-73 39 2.3 Investment share and growth rates of GDP: 16 countries, 1950-89 40 2.4 Growth rates of GDP and exports: 16 countries, 1950-89 42 5.1 Effective and real exchange rate: Germany, 1950-70 115 5.2 Effective and real exchange rate: Italy, 1950-70 115 5.3 Effective and real exchange rate: France, 1950-70 116 5.4 Effective and real exchange rate: Spain, 1950-70 117 7.1 Sectoral growth rates of output: Belgium, 1953-86 185 7.2 Sectoral growth rates of capital formation: Belgium, 1953-86 185 7.3 Sectoral employment: Belgium, 1953-86 186 7.4 Sectoral shares of wages in value added: Belgium, 1953-85 189 7.5 Relative value-added prices, sheltered sector/open sector: Belgium, 1953-86 191 7.6 Indices of competitiveness: Belgium, 1969-91 192 7.7 Net financial need of the government as a share of GDP: Belgium, 1953-93 194 7.8 Labour force, employment capacity and actual employment: Belgium, 1953-88 196 8.1 GDP: France and Germany, 1950-92 213 8.2 Unemployment rates, 1973-93 215 8.3 Real franc exchange rate, 1949-92 216 8.4 Investment rate: France, 1890-1984 218 8.5 Factor reallocations 224 8.6 Inflation and unemployment: EU and France, 1971-92 225 8.7 Investment rates and non-wage share in national income: France, 1949-92 * 226 8.8 GDP growth rates: selected countries, 1974-83 227
xii
List of figures xiii 8.9 9.1 9.2 9.3 9.4 9.5 9.6 9.7 9.8 9.9 9.10 10.1 11.1 11.2 12.1 12.2 12.3 12.4 12.5 12.6 13.1 13.2 13.3 13.4 13.5 13.6 13.7 14.1 14.2 14.3 14.4 14.5 14.6 15.1 15.2 17.1
Real consumer wages: France, 1946-91 Industrial production in Sweden and OECD, 1960-92 Annual percentage changes in Swedish GDP and the Swedish business cycle, 1945-90 Structure of Swedish employment, 1945-91 Private and public sector gross saving as a percentage of GDP: Sweden, 1950-92 Swedish capital formation by category, 1950-92 The competitive and sheltered sectors of the Swedish economy, 1952-90 Cumulative changes of public and private employment in Sweden, 1960-93 Public employment as a share of total employment in Sweden, 1950-93 GDP and taxation in Sweden, 1950-93 Average hourly wages for blue-collar workers in Swedish manufacturing, 1990: II Growth rates of real GNP in the Netherlands and the world trade volume, 1951-94 GDP per capita relative to Portugal, average, 1950-90 Log per-capita output, HP trend and subperiods, 1945-91 Spain's real GDP per head: comparative performance, 1850-1993 Rate of inflation: Spain, 1929-93 Relative price of capital goods: Spain, 1950-90 Budget balance of central government and public administrations: Spain, 1940-93 Current account balance: Spain, 1940-93 Employment, labour force and the rate of unemployment: Spain, 1955-93 GDP per capita indices: Ireland, 1950-88 GDP per capita, initial level and growth: Europe, 1950-88 GNP per capita, initial level and growth: Europe, 1979-88 GDP per worker, initial level and growth: Europe, 1950-88 GDP per employed worker, initial level and growth: Europe, 1964-88 Investment shares: Ireland, UK and the rest of Europe, 1950-88 Equipment share and price: Europe, 1960-85 TFP levels: Italy, 1938-90 Profitability: Italy, 1938-90 Market power: Italy, 1938-90 Scale economies: Italy, 1938-90 Capacity utilization: Italy, 1938-90 Inequality and poverty: Italy, 1973-91 Manufacturing productivity levels, 1950-89 Profitability and capital accumulation in manufacturing: Germany, 1951-90 Log GNP per capita: Denmark, 1890-1990
228 249 250 252 259 261 263 267 269 271 276 316 333 336 357 364 365 366 367 372 392 393 394 396 397 405 406 434 436 436 437 437 449 458 471 543
xiv
List of figures 17.2 Growth in real GNP: Denmark, 1946-91 17.3 Change in aggregate real GNP per employed person: Denmark, 1949-91 17.4 Change in hourly productivity of labour, manufacturing industry: Denmark, 1949-91 17.5 Annual change in total factor productivity: Denmark, 1967-91 17.6 Factor quality corrected estimates of TFP change in sheltered and exposed sectors: Denmark, 1967-89 17.7 Gross investments to output ratio, aggregate economy: Denmark, 1948-91 17.8 Relative change in machine capital per employed person: Denmark, 1949-91 17.9 Relative price and output share of machinery investments: Denmark, 1948-91 17.10 Aggregate export and import ratios: Denmark, 1948-91 17.11 Total taxes relative to GNP: Denmark, 1948-91 17.12 Sector GDP deflators relative to aggregate GDP deflator: agriculture, manufacturing industry and private service industries in Denmark, 1948-91 17.13 Relative price between sheltered and exposed sectors: Denmark, 1948-91 17.14 Rate of unemployment: Denmark, 1948-91 17.15 Rate of increase in nominal and real wages: Denmark, 1949-91 17.16 Ratio between user cost of capital and industrial wages: Denmark, 1955-91
544 547 547 549 550 555 555 556 558 562
562 563 564 565 567
Tables
1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 1.10 1.11 1.12 1.13 1.14 1.15 2.1 5.1 5.2 5.3 5.4 5.5 5.6 5.7
European growth, 1890-1992 page 2 War damage and reconstruction 4 GDP per person growth in Eastern Europe 5 GDP per person in 1990 international dollars 6 Unemployment rates 7 Differences in the sources of growth 9 TFP growth in different periods 10 Demand fluctuations and growth 11 Income elasticities of demand for exports 13 Estimated trend rates of growth of output per person in different periods 16 Using the Levine-Renelt model to account for changes in European growth of output per head 18 Comparative levels of labour productivity in manufacturing 21 Actual and forecast convergence, 1950-86/7 28 Growth performance and its sources relative to the European average: an analysis based on Levine-Renelt 29 Some new growth variables 31 Intra-European trade: percentage liberalized at selected dates 59 Indicators of European competitiveness, 1950 113 Selected indicators of European competitiveness, 1920s and 1950 114 Changes in selected macroeconomic indicators: Germany and Italy, selected dates 119 Indices of revealed comparative advantage: Germany and Italy, selected dates 119 Trade liberalization in OEEC countries, 1952-8 121 Changes in selected macroeconomic indicators: France and Spain, selected dates 121 Indices of revealed comparative advantage: France and Spain, selected dates 122 xv
xvi
List of tables 5.8 Indicators of foreign trade changes: France and Spain 6.1 Growth of real output per head and productivity, 1950-89 6.2 Estimates of comparative levels of productivity, 1950-89 6.3 Differences in the sources of growth, 1950-87 6.4 Levine and Renelt accounting for UK growth shortfall, 1960-73 6.5 Some comparisons relating to investment 6.6 Inflation and unemployment: UK, 1950-89 6.7 National debt, money supply and balance of payments: UK, selected dates 6.8 Growth rates of output, capital stock and TFP: UK, selected dates 6.9 Revealed comparative advantage rankings: UK and USA, selected dates 6.10 The short-term UK macroeconomic position, 1948 and 1951 6.11 Research and development and patenting, 1938-89 6.12 Labour force with intermediate vocational qualifications, 1979 and 1988 6.13 Labour productivity comparisons in international companies: reasons for differentials, 1972 6.14 Productivity outcomes in UK nationalized/privatized sector, selected dates 6.15 Productivity of labour in manufacturing, 1987: sources of German lead over UK 6.16 Panel regressions of total factor productivity growth in British industry 6.17 Value added in high-technology industry, 1979 and 1988 7.1 Major growth indicators: Belgium and north-west Europe, 1913-90 7.2 Final demand growth: Belgium, 1953-86 7.3 Sectoral cost, price and productivity growth: Belgium, 1953-86 8.1 Aggregate performance: France, 1950-92 8.2 Ratio of variances: France relative to other countries 8.3 Ratio of variances: France relative to other countries 8.4 Capital growth: France, 1930-58 8.5 TFP growth by industries, and relative prices of value added: France, 1950-92 8.6 Accounting decomposition of TFP growth: France, 1950-92 8.7 Estimates by Maddison (1987) of the effect of education on growth: France and Germany, 1973-84 8.8 Highest degree obtained and unemployment rate: Germany and France, 1989 8.9 Proportion of employees without degree according to relative earnings: France, 1970 and 1985 8.10 Credit outstanding in France, 1956 8.11 Share of French exports by destination, 1952-84 9.1 Growth in GDP per man-hour in 16 OECD countries, 1870-1970
123 133 134 135 136 137 138 139 139 140 141 144 144 145 149 152 155 160 175 186 189 212 214 214 217 220 223 230 231 232 233 235 242
List of tables xvii 9.2 Growth in GDP per man-hour in 16 OECD countries, 1950-70 9.3 Average annual growth rate of GDP, GDP per person employed and GDP per capita, 1950-70 9.4 Average annual growth rate of GDP, GDP per person employed and GDP per capita, 1970-92 9.5 Growth rate of total factor productivity in the private sector in 14 OECD countries, 1970-85 9.6 Proximate sources of economic growth in the non-government sector, 1950-90 9.7 Composition of GDP: Sweden, 1950-90 9.8 Macroeconomic outcomes: Sweden, 1940-93 9.9 Saving as a percentage of GDP: Sweden, 1950-92 9.10 Gross investment as a share of GDP: Sweden, 1950-89 9.11 Growth in productivity per hour worked in the competitive and sheltered sectors of the Swedish economy, 1951-90 9.12 The public sector share as a percentage of GDP in Sweden and in the OECD, selected dates 9.13 Total tax wedge on labour income: Sweden, 1952-92 9.14 Share of labour force having completed secondary education, 1987 9.15 Share of labour force having completed tertiary education, 1987 9.16 Before-tax educational premiums (Mincerian rates of return) in Sweden, 1968-91 9.17 Percentage increase in the hourly wage attributable to an additional year of labour market experience: Sweden, 1968-91 10.1 Population, gross domestic product, GDP per capita and GDP per hour worked: Netherlands, north-western Europe and the USA, 1913-90 10.2 GDP per head of the population and GDP per hour worked: Netherlands, north-western Europe and the USA, 1913-90 10.3 Distribution of spendable income in the Netherlands, 1959-90 10.4 Sectoral shares of employment in the total economy of the Netherlands, 1909-90 10.5 Value added and labour productivity by sector of the economy in the Netherlands, 1950-90 10.6 Value added per person employed in agriculture, manufacturing and the rest of the economy, 1950-90 10.7 Labour force, employment, hours worked and labour force participation: Netherlands, 1913-90 10.8 Average unemployment rates and growth rates of wages in the Netherlands and north-western Europe, 1929-90 10.9 Average growth rates of investment and annual compound growth rates of capital stock, capital intensity and the capital-output ratios: Netherlands, 1947-90 10.10 Capital intensity and total factor productivity in the Netherlands as a percentage of the average for France/Germany/UK and the United States, 1950-90
243 243 244 245 246 247 249 258 261 264 269 272 274 274 275 276
291 292 293 294 294 295 296 297
298
299
xviii
List of tables
10.11 Decomposition on growth of GDP and the contribution of total factor productivity growth in the Netherlands, 1950-90 10.12 Educational distribution of the Dutch labour force, 1960, 1971 and 1987 10.13 Labour costs, labour productivity and unit labour costs in manufacturing: Netherlands, 1950-90 10.14 Exports and export surplus as a percentage of GDP, export volume and terms of trade in the Netherlands, 1929-90 10.15 Government expenditure in the Netherlands as a percentage of net national income, 1955-90 10.16 Tax revenues in the Netherlands as a percentage of gross domestic product, 1965-89 11.1 Average per-capita product of the world countries in relation to Portuguese per-capita product 11.2 Growth accounting for phases: Portugal, 1952-91 11.3 Sources of output growth in Portugal, 1959-74 11.4 Growth, unemployment and political instability in Portugal, 1946-91 11.5 Portuguese external relations, 1948-91 11.6 Portuguese exchange rates, 1946-91 11.7 Portuguese government debt, 1946-91 11.8 Portuguese demand shares, 1952-90 11.9 The Pianos de Fomento 11.10 Portuguese exports by sector, 1938-90 11.11 Portuguese imports by sector, 1938-90 11.12 Portuguese exports by country, 1938-90 11.13 Portuguese imports by country, 1938-90 11.14 Postwar emigration from Portugal, 1946-91 11.15 Portuguese adult illiteracy rate, 1940-91 11.16 Rates of return to education and job training in Portugal, 1977 and 1985 12.1 Comparative performance of Spain's real GDP per head, 1850-1993 12.2 Sources of Spain's economic growth, 1965-90 12.3 Relative growth of Spain's real GDP per head, 1950-85 12.4 Differential growth of Spain's real GDP per head, 1965-90 12.5 GDP and employment composition: Spain, 1929-93 13.1 Economic growth in Ireland, the UK and Europe, 1950-88 13.2 Maddison and Heston-Summers GDP per capita 13.3 Gross investment, 1961-90 13.4 Investment by use: Ireland, 1953-90 13.5 Returns on investment in transport equipment, selected years 13.6 Net migration in Western Europe, 1950-87 13.7 A measure of two-way trade, North and South, 1926-91 13.8 Calmfors-Driffill indicators of economic performance 13.9 Cross-section analysis, 1950-60 13.10 Cross-section analysis, 1960-73
300 300 313 315 318 320 332 333 334 337 337 337 338 338 340 348 348 349 349 350 351 351 356 359 359 360 368 391 399 405 406 407 409 411 416 418 418
List of tables
xix
13.11 Cross-section analysis, 1973-88 13.12 Cross-section analysis, 1950-88 14.1 Levels and growth rates of real per-capita GDP: G7 countries, 1938-89 14.2 Main economic indicators: Italy, 1951-90 14.3 Value added per labour unit: Italy, 1951-89 14.4 Regional value added per labour unit: Italy, 1951-89 14.5 The sources of growth: Italy, 1938-90 (revenue side) 14.6 The sources of growth: Italy, 1938-90 (cost side) 15.1 The proximate causes of growth: Germany, 1870-1989 15.2 Convergence of value added per hour in manufacturing and the total economy: Germany, 1950-90 15.3 Growth accounting for manufacturing: Germany, France, Italy and UK, 1950-88 15.4 The balance between productivity and wages in industry: Germany, 1948-52 15.5 Sources of growth of non-agricultural employment: Germany, 1955-73 15.6 Educational and vocational qualifications of German and foreign employees, 1984 15.7 Apprentices as a share of employment in Germany, 1950-88 15.8 Comparative levels of labour productivity per hour, manufacturing: Germany, USA and Japan, 1973-90 15.9 German export competitiveness, 1979 and 1990 15.10 World market shares of Germany and its competitors, 1980 and 1990 15.11 Comparisons of productivity measures in the biscuit industry: quality effects 15.12 The transition from school to work: comparison between Germany and the USA 16.1 Growth of per-capita output and productivity: East Germany and West Germany, 1950-89 16.2 Population and the labour force: East Germany, 1950-89 16.3 Human capital investment in East Germany, 1950-89 16.4 Estimates of comparative East German levels of productivity 16.5 Output of GDR industry as a percentage share of total output of Potsdam Germany, 1944 16.6 Capacity losses due to war damage and dismantling: East Germany and West Germany, 1936-50 16.7 Population in Potsdam Germany, 1939-50 16.8 Capital-labour ratios in German manufacturing, 1950 16.9 Output in East German manufacturing, 1945-50 16.10 East German manufacturing output per person employed, 1936 and 1950 16.11 Company data on East German productivity, 1936-50 16.12 East German reparations to the USSR 16.13 Per-capita consumption in early postwar Germany
419 420 428 428 429 430 433 434 456 457 457 464 468 469 475 476 477 478 480 486 500 502 502 503 505 506 507 507 509 509 510 513 513
xx
List of tables 16.14 Composition of output and investment by major sectors: East Germany, 1936-60 16.15 The efficiency of investment: East Germany and West Germany, 1950-60 16.16 The efficiency of investment: East Germany, 1960-70 16.17 Output of consumer durables by categories: East Germany, 1950-70 16.18 Growth in the East German economy, 1970-80 16.19 The efficiency of investment: East Germany, 1970-80 16.20 East Germany's cumulative trade balance in the 1970s 16.21 The efficiency of investment: East Germany, 1980-9 16.22 World market and COMECON prices for crude oil, 1972-89 16.23 Growth in the East German economy, 1980-9 16.24 Valuta mark/mark exchange rate for East German exports, 1970-89 16.25 East Germany's trade balance and foreign debt, 1980-9 16.26 East German trade with Western countries, 1980-9 17.1 Long-run growth performance in international perspective 17.2 Annual change in GDP per work-hour in Denmark and average for 16 countries, decades 1870-1989 17.3 Average annual rate of growth of real GNP: Denmark, EC4 and other Nordic countries, 1950-92 17.4 Annual change in total factor productivity: Denmark, 1966-91 17.5 Average change per year in total factor productivity, labour productivity and capital productivity in the private sector, 1966-91 17.6 Indicators of sectoral shifts: Denmark, 1945-50 17.7 Sectoral GNP growth indicators: Denmark, 1950-89
515 515 519 520 521 522 522 523 524 525 526 527 528 545 546 546 548
549 554 561
Preface
Our principal aim in organizing this volume has been to present case studies of postwar growth in individual countries written in such a way that the book as a whole gives the reader a genuinely comparative picture. The collection also contains papers on important aspects of the general experience of growth, and an overview chapter which is intended to provide the context into which the country studies fit. The work which is reported in this book is the outcome of a network which operated during 1992-3 sponsored by the SPES programme of the European Commission, directed by Crafts and Toniolo and administered by CEPR. This organization permitted the establishment of a common format for the country studies and facilitated exchange of ideas and mutual criticism of initial drafts, with the result that the papers are more comparable than would otherwise have been the case. The reawakening of economists' interest in growth since the mid-1980s has been one of the principal reasons for this project. At the same time, it is now quite a while since the last exercise of a similar kind, the much used and respected collection edited by A. Boltho (The European Economy: Growth and Crisis, Oxford University Press, 1982). The authors in this volume are able to revisit the topic of postwar growth armed not only with some new theoretical ideas, but also with the experience of the 1980s on which to draw. Moreover, on this occasion the analysis has been based not only on applied economics, but also on economic history. Thus, while the book is greatly informed by insights from growth theory, the approach in the country studies emphasizes the presentation of chronological and institutional detail. The case study approach and the adoption of a longer-run perspective than is normal for economists allow insights to be obtained which do not emerge from standard cross-section regressions based on the post-1960 period. These various themes are developed in the overview chapter by Crafts and Toniolo. This paper focuses particularly on the so-called Golden Age of growth conventionally dated as 1950-73. Two important points which stand out are that the catch-up of leading countries by followers was much stronger in this period than xxi
xxii
Preface
in any other, and that this catch-up was by no means automatic, but was based on particular institutional arrangements and historical circumstances. These are explored at a general level in the contributions by Boltho, Eichengreen and Olson, while Olson's well-known ideas relating to Eurosclerosis are scrutinized for the central case of Germany by Paque. Authors of the country studies were asked to consider the role that catch-up and convergence played in each case, and also to examine more generally the determinants of productivity performance and to consider the legacy of the interwar and wartime experience. Clearly, investment in 'broad capital' was expected to be a central concern, but this would be seen in its institutional context, which could be expected to have influenced incentives to invest. The extent of institutional differences across Western European countries is broad enough to make this a potentially rewarding exercise, but the research required to address this issue fully remains a large project for the future. Each country study contains an overview of performance and a brief chronological account of growth divided into similar subsections. There is also an opportunity to take up interesting aspects of growth specific to individual countries. The set of countries comprises most of Western Europe together with East Germany. East Germany has been included not only because of unification, but also because it allows further insights into what the West got right after World War II. Members of the network on European growth also devoted a good deal of effort to exploring quantitative aspects of economic growth, with a view to improving understanding of and facilitating access to the key raw materials of a database for studying postwar growth. In addition, the evidence on key topics such as convergence, investment in broad capital and trend growth was evaluated on a cross-sectional basis. This work should be seen as complementary to the present volume, in terms of providing further interpretation of postwar economic history and also factual material for reference, and it is to be published as B. van Ark and N.F.R. Crafts (eds.), Quantitative Aspects of Postwar European Economic Growth in 1996.
List of contributors
Charles Bean Centrefor Economic Performance, London School of Economics, and CEPR Andrea Boltho Magdalen College, University of Oxford Wendy Carlin University College London Isabelle Cassiers Universite catholique de Louvain and FNRS Nicholas Crafts London School of Economics and CEPR Joao Cesar das Neves Universidade Catolica Portuguesa, Lisbon Jakob de Haan University of Groningen Herman J. de Jong University of Groningen Leandro Prados de la Escosura Universidad Carlos III, Madrid Phillippe De Ville Universite catholique de Louvain Barry Eichengreen University of California, Berkeley, and CEPR Magnus Henrekson Institute for Industrial and Social Research, Stockholm Lars Jonung Stockholm School of Economics Cormac 6 Grada University College Dublin and CEPR Mancur Olson University of Maryland Kevin O'Rourke University College Dublin Karl-Heinz Paque Institutfur Weltwirtschaft, Kiel Peder J. Pedersen University of Aarhus Albrecht O. Ritschl Universitat Pompeu Fabra, Barcelona Nicola Rossi II Universita degli Studi di Roma Jorge C. Sanz Ministerio de Economia y Hacienda, Madrid Pierre Sicsic Banque de France and NBER Peter Solar Vesalius College, Vrije Universiteit, Brussels Joakim Stymne Stockholm School of Economics Gianni Toniolo Universita degli Studi di Venezia and CEPR Bart van Ark University of Groningen Charles Wyplosz Graduate Institute of International Studies, Geneva, and INSEAD, Fontainebleau
xxm
1
Postwar growth: an overview NICHOLAS CRAFTS and GIANNI TONIOLO
1
Introduction One of the most difficult and intriguing tasks of a theory of economic growth is to combine both the disruptive and the integrative, the qualitatively changing and the quantitatively steady, aspects of the process. (Kuznets, 1965: 23)
We still refer to the past fifty years as the 'postwar': this is perhaps the best tribute to the fact that the 'second Thirty Years War' (1914-1945) marked a major watershed in the history of mankind. So much so that it proved to be a major intellectual watershed as well. In fact, until fairly recently, 1945 often marked the borderline of historical research, more recent decades being considered as the playing ground for journalists, political scientists and sociologists. Only the boldest, or most inconsiderate, scholars entered thefield,and they did so at their own risk. The same can be said of economic historians: with few exceptions, they have been reluctant to apply the tools of their trade to the 'postwar' period, more often than not leaving it as the domain of applied economists. Things are changing, however, and the half century following the end of the Second World War is now increasingly seen as being ripe for historical investigation, much beyond the Marshall Plan years that have attracted much recent attention. This chapter aims at reviewing the performance of the European economy since 1945 in a longer-run perspective, which sees the period from 1913 to 1973 as being an exceptional one in the history of 'modern economic growth', in that it departed from the secular trendfirst(1913-45) by under- and then (1945-73) by overperforming. In this chapter, the European economy is tentatively seen as an aggregate, at least in fieri. If the interwar years of slow and volatile growth were characterized by nationalism and wars that crystallized national economic peculiarities, the subsequent period of high growth resulted in a convergence of per-capita incomes that can be seen and understood as being both the cause and the effect of a broader social and institutional convergence. 1
2
Nicholas Crafts and Gianni Toniolo
Table 1.1. European growth, 1890-1992 (average annual growth) Period
Real GDP
Population
1890-1992 1890-1913 1913-50 1950-73 1973-92
(1) 2.5 2.6 1.4 4.6 2.0
(2)
Real GDP per capita (3)
Real GDP per person-hour (4)
0.6 0.8 0.5 0.7 0.3
1.9 1.7 1.0 3.8 1.7
2.6° 1.6 1.9 4.7 2.7"
a
1890-1987. 1973-87. Note: GDP and population are aggregates for 12 countries (Austria, Belgium, Denmark, Finland, France, Germany, Italy, Netherlands, Norway, Sweden, Switzerland, United Kingdom, all adjusted for boundary changes). Sources: 1870-1989, Maddison (1991); 1989-92, OECD (1993). Countries are those for which col. (4) can be calculated.
b
2
Main trends in postwar European growth
By 1870 all European countries were experiencing that 'epochal innovation' which Kuznets called 'modern economic growth'. Some of them, the pioneers, had already proceeded a long way along the road leading to ever increasing material welfare; others, the late-comers, were just taking their first steps. Since then, the aggregate real GDP of Western Europe1 has grown at an average annual rate of 2.3 per cent, or about 1.7 per cent per annum, per capita. The secular trend fits well with Kuznets' definition of'modern economic growth' as a process characterized by 'rates of growth in per capita income rang[ing] mostly from about 10 percent to over 20 percent per decade' (Kuznets, 1965: 18). While the long-run trends follow the steady quantitative change predicted by such authors as Colin Clark, Kuznets, Abramovitz and Chenery, Table 1.1 shows that the performance of the European economy during individual relevant subperiods diverged considerably from the secular trend. The perspective of secular trends in 'modern economic growth' is a useful starting point in considering the economic history of Europe during the half century following the end of the Second World War. In particular, such a perspective is helpful when considering the standard subdivision into two distinct periods, the first, to about 1973, being characterized by very high growth rates and near-full employment, the second showing a rather sluggish performance in terms both of output and employment. Here, the longer-run view conveys two messages: (1) the period 1950-73 was truly exceptional in the process of'modern economic growth', (2) the subsequent growth record can hardly be regarded as unsatisfactory. The exceptional character of the 'high-growth years' is better judged in the light of the poor European performance during the previous three decades. Whether or not one agrees with Kuznets2 that major wars are somehow endogenous to the process of modern economic growth, they certainly coincide with a considerable slowdown in European growth between 1913 and 1950, plausibly not unrelated to the boom of
Postwar growth: an overview
3
the following two decades. As we shall see, the catch-up for ground lost in two world wars and in the most severe economic depression to date is one of the reasons explaining the much above average growth rates of the 1950s and 1960s: other reasons, discussed in section 6, relate to domestic and international factors likely to be exceptional in the history of modern economic growth. Post-1973 growth looked uncomfortably low in the light of expectations created during the previous quarter of a century. However, if both the longer run and the predictions of the then-prevailing theory are taken into account, the picture looks considerably less dismal. In particular, Europe's per-capita output growth is (1) just slightly below the secular trend,3 (2) equal to that achieved during the 'belle epoque' of the world economy (1890-1913), (3) distinctly better than the belle epoque if seen in terms of GDP/hour worked, (4) much above the pre-1890 rates.4 It is hard to pass a negative judgement on such a record, particulary if it is confronted with the neoclassical growth theory assumption of diminishing returns to factor inputs. Having assessed the postwar European performance against the longer-run background of'modern economic growth', we may now turn to a brief appraisal of some quantitative features of the main subperiods. 2.1
War damage and reconstruction
The damage induced by war on the individual national economies is relevant to our analysis of long-run growth in that: (1) its sheer dimensions provided a rationale for American aid, with the attendant new 'international order' and technology flow, and (2) domestic reconstruction was often accompanied by important institutional changes. Table 1.2 provides an impressionistic picture both of the extent of war-induced economic setbacks and of the speed of reconstruction. In 1945 the GDP of France, of the three Axis countries and of the Netherlands had fallen back to late nineteenth-century or early twentieth-century levels. One or two generations of work and accumulation had been lost. Belgium fared a little better. Of the belligerent countries, only the United Kingdom managed to contain the post-1943 economic collapse to minor proportions; it could not, however, avoid a serious post-war slump. Neutral countries in our sample fared distinctly better and, in fact, managed to grow throughout the conflict. If the economic effect of the war was devastating, the speed of recovery was so impressive as to take by surprise informed observers, trade unions, entrepreneurs and policy-makers alike. In five years at most, Europe recovered the ground lost relative to the highest prewar income levels.5 It is, thus, quite safe to place the end of the first phase of reconstruction and the beginning of a new era in the history of European economic growth in 1950. 2.2
The high-growth years
As we have seen, the years 1950-73 witnessed a unique episode in the history of European 'modern economic growth' - so much so that the period is now often referred to in the literature as the 'Golden Age7. Not only were growth rates exceptionally high (Table 1.1), but cyclical fluctuations were mild6 and inflation rates socially acceptable.7
4
Nicholas Crafts and Gianni Toniolo
Table 1.2. War damage and reconstruction
Austria Belgium0 Denmark Finland France Germany Italy Netherlands Norway Sweden Switzerland United Kingdom
Pre-war year when G D P was the same as in 1945
Year when G D P recovered the highest prewar level
(1) 1886 1924 1936 1938 1891 1908* 1909 1912 1937 never never never
(2) 1951 1948 1946 1945 1949 1951 1950 1947 1946
Annual rate of GDP growth during 'reconstruction' (1945 to year in col. (2)) (3) 15.2 6.0 13.5 19.0 13.5 11.2 39.8 9.7
a
Interpolations. ^Relative to 1946. Source: Data from Maddison (1991: 208-19).
Productivity growth is, obviously, the core phenomenon to be investigated in the present context: detailed measures and analyses of its determinants are, therefore, provided in the individual country studies included in this book. In the last column of Table 1.1, an attempt is made at estimating Europe-wide long-run changes in labour productivity. Although aggregation and index number problems make estimates in Table 1.1 more questionable than those for individual countries, the table itself leaves little doubt as to the order of magnitude of the general acceleration in productivity that took place in Europe during the Golden Age, making it a unique episode in the 'modern economic growth history' of the Old Continent. Four features of this extraordinary event are worth noting: • It is a distinctly European phenomenon, Japan being quantitatively and otherwise a case of its own,8 North America and Australia showing a much less pronounced deviation from the secular 'norm', 9 and other successful countries such as the NICs entering their phases of high growth in later decades. • High growth rates, relative to previous and subsequent records, characterize almost all the individual European economies, regardless of their social, political and economic institutions. • The period is strongly characterized by both /?- and a-convergence, to use Barro and Sala-i-Martin (1991) terminology. In other words, during 1950-73 we observe both a pronounced tendency for initially poorer countries to grow faster than richer ones and a decline in the cross-sectional scatter of per-capita output growth.
Postwar growth: an overview 5 Table 1.3. GDP per person growth in Eastern Europe
Czechoslovakia Hungary Poland Bulgaria Romania
Czechoslovakia Hungary Poland Bulgaria Romania
GDP/person 1950 (1990 international dollars)
GDP/person 1973 (1990 international dollars)
Growth of real GDP/person 1950-73 (% per year)
3480 2482 2447 1654 1175
6995 5601 5334 5294 3458
GDP/person 1973 (1990 international dollars)
GDP/person 1992 (1990 international dollars)
Growth of real GDP/person 1973-92 (% per year)
6995 5601 5334 5294 3458
6457 5639 4819 4132 2531
-0.4 0.0 -0.5 -1.1 -1.3
3.1 3.6 3.4 5.2 4.8
Source: Maddison (1995).
• Exceptionally low unemployment rates prevailed. In most countries, the 'triumph of full employment' was a historically distinct feature of the period. While we leave a discussion of the above-mentioned features of the period to sections 5 and 6, dealing with the causes of growth, a brief quantification of the last three points might be in order here. Table 1.3 provides some evidence showing that 1950-73 growth rates of per-capita GDP in Eastern Europe were roughly as high as those prevailing in the West (see also Ofer (1992)), while the Iberian economies under Salazar and Franco were among the fastest growers in the West itself, as Table 1.4 shows. A useful, if impressionistic, perception of the convergence processes may perhaps be drawn from Tables 1.3 and 1.4. We observe both a strong tendency for initially poorer countries to grow faster than richer ones and a decline in the cross-sectional scatter of per-capita output. As a result, in the early 1970s, the dispersion in levels of per-capita income among European countries was much less pronounced than it had been a quarter of a century earlier. Moreover, the formation of a core of economies characterized by roughly similar standards of living was, by then, well under way. Several features of the experience of Western Europe stand out: • The ratio between the highest and the lowest per-capita incomes fell very fast between 1950 and 1973, and more than halved in the period of 1950-92. • By the early 1990s there had emerged a 'core' group of nine countries with a very similar per-capita income all within ± 8.0 per cent of the median, whereas in 1950 only two were.
6
Nicholas Crafts and Gianni Toniolo
Table 1.4. GDP per person in 1990 international dollars
1. Switzerland 2. UK 3. Sweden 4. Denmark 5. Netherlands 6. Belgium 7. France 8. Norway 9. W. Germany 10. Finland 11. Austria 12. Ireland 13. Italy 14. Spain 15. Portugal 16. Greece
1. Switzerland 2. Sweden 3. Denmark 4. W. Germany 5. France 6. Netherlands 7. UK 8. Belgium 9. Austria 10. Finland 11. Italy 12. Norway 13. Spain 14. Portugal 15. Ireland 16. Greece
1950
1973
1950-73 growth % (rank)
8939 6847 6738 6683 5850 5346 5221 4969 4281 4131 3731 3518 3425 2397 2132 1558
17953 11932 13494 13416 12763 11905 12940 10229 13152 10768 11308 7023 10409 8739 7568 6229
3.1 (12) 2.4 (16) 3.1 (12) 3.1 (12) 3.4 (10) 3.5 (9) 4.0 (8) 3.2(11) 5.0 (4) 4.2 (7) 4.9 (5) 3.1 (12) 4.9 (5) 5.8 (2) 5.6 (3) 6.2 (1)
1973
1992
1973-92 growth % (rank)
17953 13494 13416 13152 12940 12763 11932 11905 11308 10768 10409 10229 8739 7568 7023 6229
21036 16927 18293 19351 17959 16898 15738 17165 17160 14646 16229 17543 12500 11130 10711 8238
0.9 (16) 1.2(15) 1.7 (9) 2.1 (5) 1.7 (9) 1.5 (12) 1.5(12) 1.9 (7) 2.2 (3) 1.6(11) 2.4 (2) 2.9 (1) 1.9 (7) 2.1 (5) 2.2 (3) 1-5(12)
Source: Maddison (1995). • Both Switzerland at the top and Greece at the bottom remained in the same positions. • There is an inverse rank correlation between initial level of income and growth, which is particularly marked in the Golden Age. A similar inverse correlation for the Golden Age between initial income level and
Postwar growth: an overview
7
Table 1.5. Unemployment rates
Twelve countries (benchmark years)
Four countries (benchmark years)
1929 1938 1950 1973 1989
1920-9° 1930-8 1932* 1950-9 1960-73 1962 1974-81 1982-9 1990-3
4.0 5.0 4.5 2.9 7.4
4.4 7.1 12.8 4.2 2.5 1.9 5.2 8.8 9.2
a
Three countries only (excludes Italy): France three years only. Three countries only (excludes France). Sources: Maddison (1991), OECD (1993).
b
subsequent growth is noticeable in Eastern Europe in Table 1.3. Here, however, performance is somewhat less impressive in the years 1950-73 and, obviously, very disappointing since 1973. Comparison of Tables 1.3 and 1.4 indicates that, normalizing for initial income level, East European countries' growth was less than for comparable Western European countries in 1950-73. For example, Czechoslovakia does about as well as Ireland, which is an unusually poorly performing country in the West, while the rapidly rising difference between Austrian and Czech income levels as the postwar period unfolds is a stark testimony to Communist failure. Finally, while labour market statistics are among the least comparable over time and cross-country, there can be little doubt that European unemployment reached its lowest secular level during the so-called Golden Age. Some evidence is provided in Table 1.5, showing benchmark-year unemployment rates for twelve European countries10 and period averages for the four largest ones.11 Data for years up to 1913 are not available on a roughly comparable basis, but scholars agree that it is unlikely that unemployment reached the low levels of the 1960s. Moreover, it is believed that data for the interwar period tend to underestimate underemployment and, therefore, unemployment when measured according to postwar statistical definitions. 2.3
The post-1973 slowdown
The end of the Golden Age is often conventionally dated as 1973, the year of the first oil shock. In accepting this benchmark year, we imply neither that a sudden discontinuity in productivity growth can be detected in this particular year, nor that its main cause is changing terms of trade energy input. It is true, however, that the early 1970s saw a major departure from the main trends of the previous two decades, heralded by the end of the so-called Bretton Woods system in 1971 and characterized by severe shocks both to aggregate supply and aggregate demand. As mentioned earlier, production and productivity growth levelled off around the
8
Nicholas Crafts and Gianni Toniolo
secular trend (Table 1.1). The slowdown affected market and socialist economies alike (Table 1.3). Unemployment rates almost trebled in the 1980s relative to the 1960s (Table 1.5). Cyclical volatility increased with several European countries experiencing negative growth rates in 1975, some time between 1980 and 1983, and again in the early 1990s. Prima-facie evidence of ^-convergence appears to be much weaker after 1973. In the aggregate of the European twelve and America, GDP per person grew roughly at the same rate despite Europe's lower 'starting' point. Intra-European association of growth rates with initial GDP levels is less clear-cut than it appeared to be in the previous (1950-73) period (Table 1.5). Nevertheless, by 1992, not only was the dispersion in per-capita income levels among the sixteen countries in Table 1.4 much less pronounced than it had been forty years earlier, but there had emerged a distinctive core of nine countries with a very similar per-capita income, all within ± 8.0 per cent of 16900 1990 international dollars. 3
Previous interpretations of European postwar growth
Between the early 1960s and the early 1980s, an earlier vintage of growth theory produced a large research effort on the European experience. The focal point of this writing was the rise and fall of the Golden Age. There are interesting themes to be considered from both traditional neoclassical and Keynesian writers. This section offers a brief and somewhat critical survey of this literature. 3.1
Growth accounting
The use of growth accounting to investigate postwar growth was pioneered by Denison (1967) in a deservedly famous study. Perhaps the best-known recent research in this tradition is that of Maddison (1987), subsequently updated and revised (1991). The central approach embraces a traditional neoclassical (Solow) view, based on diminishing returns to factor accumulation and exogenous total factor productivity (TFP) growth. The methodology relies on assumptions rather than estimation and uses factor shares to weight the contributions of capital and labour. The main proximate source of the acceleration in growth during the Golden Age and subsequent slowdown emerges as TFP growth. A large role is assigned to residual efficiency both in speed-up and slowdown and also in the growth rate gap between the UK and, on the other hand, France and Germany in 1950-73 (see Table 1.6). A similar pattern of TFP growth shows up in Table 1.7, which takes in a wider group of countries. Rapid TFP growth reflected, in part at least, technology transfer, postwar reconstruction and structural change. Growth accounting also sees a strong supporting role for increased physical capital formation in the faster growth of the Golden Age compared with before and since, although this is given relatively little weight in explaining the UK's slower growth. Surprisingly, human capital formation is seen as playing only a very minor role in accounting for variations in growth rates. The reliability of conclusions drawn from growth accounting has, of course, often been questioned and is potentially undermined by the advent of new growth theory which stresses the importance of human capital and externalities to investment. The general thrust of
Postwar growth: an overview 9 Table 1.6. Differences in the sources of growth (% per year)
1913-50 Labour input Education Capital input Total factor productivity Backwardness Other specific Residual efficiency 1950-73 Labour input Education Capital input Total factor productivity Backwardness Other specific Residual efficiency 1973-87 Labour input Education Capital input Total factor productivity Backwardness Other specific Residual efficiency
France
Germany
UK
-0.17 0.36 0.65 0.67 0.04 0.06 0.57
0.38 0.24 0.62 0.28 0.20 -0.09 0.17
0.12 0.33 0.82 0.35 -0.04 0.05 0.34
0.18 0.39 1.84 3.02 0.71 0.43 1.88
0.15 0.19 2.27 3.50 0.70 0.54 2.26
0.01 0.20 1.75 1.27 0.17 0.35 0.75
-0.24 0.56 1.48 0.92 0.11 0.19 0.62
-0.49 0.05 1.28 1.01 0.28 0.22 0.51
-0.19 0.41 1.12 0.82 -0.25 0.33 0.74
Source: Derived from Maddison (1991) as in Crafts (1992a).
early empirical work of this kind is to regard growth accounting's treatment of physical capital as basically acceptable, but to argue that the approach underestimates the role of human capital by ignoring training and externalities (Oulton and O'Mahony, 1994). In placing his findings in a historical context, Maddison stressed the special and unrepeatable characteristics of the Golden Age, including liberalization of international trade, successful macroeconomic management and a backlog of opportunities for catch-up growth (1991: 168). Growth accounting has proved fruitful in suggesting working hypotheses like these, but cannot in itself test them. The growth accounting approach has also been used to consider the experience of Eastern Europe, which, as Table 1.3 demonstrated, shared in the growth of the Golden Age. Here the data difficulties are potentially particularly daunting, but research is now becoming available which seems to correct more successfully for the problems like hidden inflation that bedevil this area. The adjusted figures in Table 1.7 show that the pattern of acceleration and decline in TFP growth also appears in Eastern Europe, but at lower rates, again suggesting that catch-up was important but less successfully achieved.
10 Nicholas Crafts and Gianni Toniolo
Table 1.7. TFP growth in different periods (% per year)
Belgium Denmark France West Germany Italy Netherlands UK
Czechoslovakia Official Adjusted East Germany Official Adjusted
1950-62
1960-73
1973-9
1.9 1.8 3.5 4.5 4.3 2.6 1.3
3.9 2.8 4.0 2.7 4.6 3.1 2.3
1.5 1.2 1.7 1.8 2.2 1.5 0.6
1.4 1.3 1.7 0.8 1.3 0.9 1.6
n.a. n.a.
2.3 1.3
2.1 0.3
-1.5
4.0 2.9
2.4 1.8
1.9 1.0
-0.5
1979-90
0.2
1.2
Note: Data for 1950-62 and East Europe relate to GDP, others to business sector. Sources: Denison (1967), OECD (1991a), Klacek et al. (1993) and Ritschl (1996).
3.2
Eurosclerosis
The notion of sclerotic tendencies in economic growth performance achieved widespread attention in the 1980s. The seminal work was that of Olson (1982). The approach is based on public choice economics and argues that democracies tend through time to accumulate interest groups which inhibit growth unless these turn out to be encompassing organizations. Thus, 'distributional coalitions slow down a society's capacity to adopt new technologies and to reallocate resources in response to changing conditions, and thereby reduce the rate of economic growth' (Olson, 1982: 65). Olson developed his argument both to explain the general speeding up of TFP growth in the Golden Age and the subsequent slowdown, as well as to account for the different experiences of countries like Sweden, West Germany and the UK. He argued that trade liberalization, and particularly the formation of new arrangements such as the European Community, broke down interest group barriers to growth, although eventually these reappeared in the new context. Fast growth in the early postwar period in countries like Germany compared with the UK would be expected since distributional coalitions were temporarily destroyed by initially totalitarian government and then more fully by foreign occupation. Better growth in Sweden than the UK could be explained by the highly encompassing nature of Swedish but not British special interest groups, as reflected, for example, by the different nature of their trade unions. Further reflections on these themes can be found in Olson (1996). Although political economy considerations surely do play a serious part both in the success of postwar reconstruction and in relative success and failure in postwar growth, the original formulation proposed by Olson has been quite strongly criticized in subsequent research. Two points have been well made in recent papers.
Postwar growth: an overview
11
Table 1.8. Demandfluctuationsand growth (% per year) Effect of standard deviation of Y Austria Belgium Denmark Finland France Germany Ireland Italy Netherlands Norway Portugal Spain Sweden Switzerland UK
1.3 1.1 1.3 1.6 0.9 1.5 1.2 1.4 1.4 0.8 1.6 1.5 0.9 1.8 0.9
Effect of standard deviation of RM -1.1 -0.5 -1.1 -1.5 -0.6 -0.8 -0.8 -0.8 -0.9 -0.5 -1.1 -0.9 -1.0 -0.9 -0.9
Sum 0.2 0.6 0.2 0.1 0.3 0.7 0.4 0.6 0.5 0.3 0.5 0.6
-0.1 0.9 0.0
Source: Derived from Kormendi and Meguire (1985: 149, equation (4)). First, the suggestion that Germany was purged of sclerotic interest groups in the 1940s is hotly contested by Paque (1994). Second, more generally, Unger and van Waarden (1994) suggest that the overall impact of interest groups on growth is much less adverse than Olson claims, and that the differing experiences of the 1930s and 1940s had little impact on the subsequent power of distributional coalitions. Econometric analysis of international cross-sections has also tended to reject the sclerosis hypothesis. Crafts (1992b) found that adding various measures of corporatism as a way of looking at the degree of encompassing in organizations added no explanatory power to a conventional model of catch-up growth. Similarly, Castles and Dowrick (1990) found that, having allowed for capital and labour growth and an initial income gap, a variable measuring time since the last interruption to domestic democracy was insignificant. 3.3
Demand stabilization
Keynesian economists have always stressed the absence of economic crises in the Golden Age, and have argued that the postwar commitment to demand management and the stability achieved under the Bretton Woods system were more conducive to investment and growth than the shocks of the 1920s and 1930s. Similarly, the stagflation of the 1970s is blamed for the abrupt ending of the Golden Age. Arguments of this kind also feature prominently in Maddison (1991). Perhaps the best-known exposition of these views is by Boltho, who argues that 'If cheap technology and abundant labour and raw materials were not novel features of industrial Europe, something else must still explain why in the 1950s and 1960s they led to "supergrowth"... Such an explanation would seem to have to come from the demand side'(1982: 15).
12 Nicholas Crafts and Gianni Toniolo
The best-known attempt to test this hypothesis was that of Kormendi and Meguire (1985), some of whose results, based on international cross-section regressions for 1950-77, are reported in Table 1.8. They found that a higher variance of output was actually positively rather than negatively related to growth, but that greater variation in unanticipated money (RM) did, as expected, reduce growth. Comparison of equations with and without investment as an additional explanatory variable led Kormendi and Meguire to conclude that the impact came partly through greater investment, but more through higher returns to investment. These results provide rather mixed support for Boltho's position. They do not suggest that fine tuning via demand management explains relative success and failure in growth, but they do indicate that unpredictable policy is bad for growth. No comparison is available with other periods, but it does seem probable that monetary shocks were larger in the interwar period, and the regressions show that this might have had a serious adverse effect on growth. While this may be true, Boltho's emphasis on demand as the key difference between the interwar and postwar periods seems to be overstated. In particular, it underplays the much greater postwar ease of technology transfer relative to the previous decades, together with the reduced importance of natural resources and domestic market size in making it easier for Europe to catch up with the USA in the postwar period (Nelson and Wright, 1992). 3.4
Export-led growth
An alternative Keynesian view of postwar growth was put forward by Thirlwall (1979) following in the footsteps of Beckerman (1962) and Lamfalussy (1963). Here the central hypothesis is of a balance of payments constraint on demand and thus on growth. The full story then involves virtuous (vicious) circles in which better (worse) export demand growth prospects promote higher (lower) investment, which enhances (retards) productivity and competitiveness, and further reinforces differential export growth prospects. Theflavourof these arguments is well reflected in the following: 'Rates of growth of capacity, and hence output, in the long run are determined by expectations about future demand prospects' and cthe first place to start seeking an explanation of widely divergent growth rates is in the field of the conspicuous role played by buoyant exports and the foreign balance in determining relative growth rates in Europe in the last decade' (Beckerman, 1962:916,918). Shortly afterwards, estimates of income elasticities of demand for exports by Houthakker and Magee (1969) seemed to show a starting point for these virtuous and vicious circles; their results, reported in Table 1.9, showed a huge apparent advantage for Japan over the UK. However, subsequent research has not been very kind to these export-led growth hypotheses. Three points in particular should be noted. • Houthakker and Magee's estimates are now generally regarded as unreliable. Balassa calculated constant market share (ex-ante) elasticities based on a detailed disaggregation of trade by commodity and market. As Table 1.9 shows, these indicate quite minor differences in initial positions and thus no real basis from which the vicious/virtuous circles could develop.
Postwar growth: an overview
13
Table 1.9. Income elasticities of demand for exports
Austria Belgium Denmark France Germany Italy Japan Netherlands Norway Sweden UK
Houthakker and Magee (1951-66)
Balassa (1953-71)
1.59 1.87 1.69 1.53 2.08 2.95 3.55 1.88 1.59 1.76 0.86
2.04 1.98 1.82 2.04 2.27 2.07 2.00 1.91 1.82 1.93 2.20
Sources: Houthakker and Magee (1969) and Balassa (1979).
• Indeed, it seems likely that Houthakker and Magee's estimates reflect differences in export supply capabilities. Krugman (1989) set out a plausible model based on intra-industry trade, in which countries with a faster rate of growth of productive potential exhibit a faster rate of increase in the variety of their exports, gain world market share and appear to have faced higher income elasticities of demand for exports. In Krugman's reformulation, all the differences originate on the supply side. • The Beckerman/Thirlwall model depends on an absence of conventional neoclassical price effects, which would correct balance of payments disequilibria, and, at bottom, supposes a world of real wage rigidity. This is widely rejected empirically as an acceptable long-run proposition (Layard et al, 1991: 210).
3.5
Labour supply
A number of writers have seen an elastic labour supply as a key factor underwriting the supergrowth of the Golden Age and another reason why, in the 1950s especially, German growth couid outstrip that of the UK. Several variants on this theme were proposed during the 1960s. As was mentioned earlier, Denison (1967) found reallocation of labour from low-productivity agriculture to be a major (transitional) addition to T F P growth. A much more controversial hypothesis was advanced by Kaldor (1966). Kaldor argued that the manufacturing sector was central to postwar growth and that Verdoorn's Law operated in this sector, i.e. that labour productivity growth was a positive function of the growth of employment. This might be seen as a kind of dynamic economies of scale. Countries in which there were reserves of agricultural labour or substantial immigration could benefit from Verdoorn's Law and would experience faster growth. Relatively slow growth would be the outcome in the UK. A number of subsequent articles challenged these claims. Crucially, Kaldor's
14 Nicholas Crafts and Gianni Toniolo approach and that of his disciples, Cripps and Tarling (1973), assumed away catch-up growth based on reducing the technology and capital/labour ratio gaps between Europe and the USA. The econometric evidence in favour of Verdoorn's Law as a general proposition was readily shown to be very weak (Chatterji and Wickens, 1983; Gomulka, 1971; Rowthorn, 1975). The most intriguing version of the labour supply hypothesis was proposed by Kindleberger (1967). He was careful not to claim too much: 'the availability of excess labour does not determine the levels at which growth will be maintained . . . An elastic labour supply supports high rates of investment... But the labour supply... is a permissive rather than an initiating or even a determining factor' (1967: 4). Kindleberger invoked the famous Lewis model of development with unlimited supplies of labour, and argued that an elastic labour supply sustained the share of profits. This prediction is not, however, one which would be derived from the models of growth and unemployment normally applied to European economies today. In the standard bargaining model framework of the NAIRU model, profits share is independent of the bargained wage and of union power, while at the macro level unemployment adjusts to bring the target real wage into line with the feasible real wage (Layard et ai, 1991: 100-9). In a growth model employing intertemporal maximization, accumulation depends on the return to investment and the perceived opportunity cost of investment, not the share of profits. Nevertheless, Kindleberger's suggestion that wage moderation may have been important to the launching of the Golden Age is still one that deserves to be taken seriously, and it is explored in considerable detail by Eichengreen (1996). While at the macro level unemployment can sustain profits, at the micro level, where investment decisions are made, the expected rate of return depends on workers' ability to commit themselves to restrain wage claims, despite the fact that they are too small to influence the NAIRU. Postwar circumstances may indeed have been unusually favourable to the creation of institutions which would facilitate wage moderation by workers in return for high investment by firms. Overall, this earlier body of work suggests that the Golden Age was an unusual period and provides some useful insights both into the achievement of high growth and into the relative success and failure of individual countries. Nevertheless, at the very least it must be recognized that even the most plausible of these accounts leaves a large number of loose ends. By comparison with more recent theorizing on growth, several points appear rather unsatisfactory. In particular, there is very little emphasis placed on human capital, demand seems to be given too much prominence, and questions of catch-up and convergence are treated in a rather casual fashion. Above all, the large role left for 'residual efficiency' changes in growth accounting suggests a need for further research. 4
Recent developments in growth theory
As we have seen, until quite recently mainstream economics analysed economic growth in terms of the Solow model and its empirical counterpart, the growth accounting pioneered by Abramovitz, Denison and Kendrick and continued most notably by Maddison (1987). Nevertheless, economists were uneasy about this. As
Postwar growth: an overview
15
van de Klundert and Smulders put it, 'The neoclassical model can explain the stylized facts as listed by Kaldor ... However, the way in which the model explains these facts is unsatisfactory. All long-run growth stems from factors outside the model . . . An intuitively appealing model of growth would allow for permanent effects of savings and taxation by endogenizing technological change' (1992:180-1). Three aspects of recent attempts to improve upon traditional growth economics are of particular interest to postwar European experience. First, accumulation of capital plays a bigger part in growth outcomes in the new growth theory. The concept of capital which is embraced is typically 'broad capital', which subsumes both physical and human components, the latter comprising skills and/or a stock of knowledge. A common feature of balanced growth solutions involves a constant steady-state ratio of physical to human capital, and an implication of this approach, at a minimum, is that diminishing returns to the accumulation of reproducible factors of production are much less severe than in a traditional Solow model. The well-known Mankiw, Romer and Weil AugmentedSolow model (1992) has an exponent on broad capital in a Cobb-Douglas framework of 2/3, double that on capital in the Solow model. In general, there is a strong expectation that with a broad concept of capital there may be important externalities to investment. Some models, including, for example, the famous Rebelo (1991) Y = AK formulation, assume constant returns to broad capital formation, in which case growth can be regarded as fully endogenous. It should be noted that radical revision to growth accounting methods may be required if models of this kind are empirically verified. Second, savings/investment decisions are now explicitly placed in terms of optimizing behaviour rather than exogenous as in the Solow formulation. Similarly, substantial attention has been given to introducing imperfect competition into growth models to provide an appropriate incentive structure for investment in innovative activity (Romer, 1994). An important implication of these developments is to allow economic policy and institutional structures a much greater potential impact on growth outcomes through the investment processes on which they impinge. A third important feature of recent work is research into convergence. The traditional neoclassical and Augmented-Solow growth models embody convergence to a steady-state growth rate. If countries had identical technology and the same savings and population growth rates, then in the long run convergence would in all cases be to the same steady-state growth rate and per-capita income level. The speed of convergence depends in particular on the severity of diminishing returns to capital. (No such mechanism operates in the Rebelo model.) As this convergence process takes place, we would expect in the cross-section of countries to observe both a negative relationship between the initial level of income per head and subsequent growth, and also a reduction in the coefficient of variation of per-capita income, which would eventually tend to zero. Barro and Sala-i-Martin (1991) called these /?- and cr-convergence respectively. Actually, the process which may be observed is probably more likely to be one of 'conditional convergence', as growth accounting implicitly assumes, as economic historians have always tended to believe, and as new growth theory emphasizes. Here allowance would be made for additional variables such as 'social capability', technology gaps and accumulation strategies, and these would be controlled for in
16 Nicholas Crafts and Gianni Toniolo Table 1.10. Estimated trend rates of growth of output per person in different periods (% per year)
Belgium Denmark France Germany Italy Sweden UK USA
1860-1914
1920-39
1951-73
1974-89
0.90 1.77 0.96 1.47 1.47 1.52 1.04 1.70
1.01 1.58 0.78 2.91 0.21 3.03 1.56 0.86
3.90 3.46 4.92 5.11 5.31 3.42 2.24 1.54
2.09 1.59 1.42 1.26 2.05 1.62 1.83 1.89
Note: Dataset taken from Maddison (1991); start date for Italy = 1861, USA = 1869. Source: Crafts and Mills (1996), which includes full details of estimation methods, statistical significance and testing for unit roots. regressions of income on initial income level. Conditional convergence might in principle be based on initial disequilibria in factor proportions, or on technology transfer in a world where obstacles are reduced. Indeed, 'technological catching-up' might be observed in contexts which produce local rather than global convergence (Durlauf and Johnson, 1992). It should be recognized that, while the new growth economics has produced a substantial volume of empirical research, in several important respects theory has run ahead of measurement, especially with regard to examining differences within the advanced countries rather than between the First and Third Worlds. This is particularly true with respect to measuring human capital, which is unfortunate given the new weight attached to it by theory, and in the area of quantifying influences on the ex-ante decision to invest. 5
Europe's postwar growth in a long-run perspective: some quantitative aspects
Section 2 established the broad quantitative elements of the unique experience of the Golden Age and of the subsequent slowdown. In this section we scrutinize the data more closely, using modern analytic tools to review key ideas both from economic history and the new growth economics. Initial insights into growth in the Golden Age can be obtained using time series econometrics, as in Crafts and Mills (1996). Some of their results are summarized in Table 1.10. Crafts and Mills' approach takes seriously Maddisons's suggestion of epochs in growth, and estimates a segmented trend with breaks at 1914,1919,1939, 1950 and 1973 for countries with long-run GDP data. Tests are also given of the Janossy (1969) hypothesis, well known to students of West German growth, that the Golden Age was an era of reconstruction and rapid recovery back to a long-run trend path of growth as in the Solow model, based on a constant exogenous rate of Harrod-neutral technological change and, thus, labour productivity growth.
Postwar growth: an overview
17
6.50-
6.00-
5.50-
' Reverse Janossy' extrapolated trend
5.00-
4.50-
Fitted trend
V
4.00
1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 Figure 1.1 Trend growth of real output per person in Italy, 1870-1990
The estimates in Table 1.10 give strong support to the hypothesis that for European countries the trend rate of growth was exceptionally rapid during the Golden Age. The technical details of the estimation also establish that the hypothesis of a unit root in GDP per person can be rejected in favour of the segmented trend stationary alternative in every case, a result which goes against the idea of fully endogenous growth with constant returns to accumulation. Further testing revealed that, when growth slowed down again, in all but one case (Denmark) European countries were enjoying a higher output per person to that which might have been expected simply on the basis of extrapolating their pre-1914 experience. However, although the pure Janossy hypothesis is generally rejected, in a number of cases backward extrapolation of the post-Golden Age trend does go back to join the actual series at a pre-disturbance date, if the final break of trend is dated by an endogenous search procedure. Italy is a case in point, as is shown in Figure 1.1, where the contrast with the USA stands out. Thus, the Janossy (1969) stress on reconstruction as an important ingredient in Golden Age growth in many European countries is still useful, even though his traditional constant-trend neoclassical growth model appears too rigid. Levine and Renelt (1992) conducted an extensive analysis of the findings of empirical work on growth in the 1960-89 period, based on international cross-section datasets. Their work is set in a catching-up framework and provides estimated equations to account for growth, while at the same time investigating the robustness of alternative specifications. Their results offer an alternative to Maddison-style growth accounting and do not impose assumed elasticities of output with respect to physical and human capital formation. In Table 1.11 we use one of Levine and Renelt's equations to consider what new
18 Nicholas Crafts and Gianni Toniolo Table 1.11. Using the Levine-Renelt model to account for changes in European growth of output per head
Constant Initial GDP/head Investment/GDP Secondary enrolment Primary enrolment Government/GDP Forecast Actual
1923-38
1950-73
1973-89
2.01 -2.43 1.42 0.16 1.90 -0.62 2.44 2.12
2.01 -2.49 2.22 0.68 1.99 -0.87 3.54 3.84
2.01 -3.55 2.06 0.79 1.79 -1.27 1.83 2.14
Sources: Estimates are for unweighted averages of European countries in Maddison (1991), excluding Belgium and Switzerland, and use Levine and Renelt (1992) equation (ii), with population growth and irrelevant dummies ignored. The initial income variable was expressed as a percentage of the US level in each year, and this figure was then multiplied by 1950 US income per person as measured by Levine and Renelt. Basic sources of data were Maddison (1991), Mitchell (1988) and OECD (1991b), augmented for interwar investment by Maddison (1992) and for interwar government consumption expenditure by den Bakker et al. (1990), Feinstein (1972), Hansen (1974), Hjerppe (1989), Krantz and Nilsson (1975), Rossi et al (1992), Sommariva and Tullio (1986) and Villa (1993). growth theory might suggest accounts for the speeding up and slowing down of European growth in the postwar years. This equation predicts growth of output/person using measures of the initial productivity gap, investment and government consumption shares and school enrolment rates as independent variables. Although Levine and Renelt doubt the robustness of the government consumption variable, we include it here given the strong evidence that there is a robust relationship in OECD countries, if not across the whole world (Dowrick, 1992). Taken at face value, the results in Table 1.11 indicate the following: • The acceleration in growth in the Golden Age came from greater investment in both physical and human capital, rather than a larger initial income gap than in the interwar period. • The speed-up in growth in the Golden Age has a significant unexplained component (of about one-third); the Golden Age saw improved growth which may reflect an enhanced catch-up capability in the postwar world and/or adverse effects of protectionism and demand shocks in the interwar period. • The slowdown in growth after 1973 is quite well explained and reflects lower scope for catch-up, as captured by a lower initial income gap and a rising share of government consumption. Compared with the results obtained by traditional growth accounting, this approach attributes more of the acceleration in growth in the Golden Age to human capital and less to TFP growth. Thus 1.41 percentage points of the 1.72 increase in growth between 1923-38 and 1950-73 is attributed to broad capital accumulation. However, the slowdown after 1973 is attributed much more to erosion of catch-up
Postwar growth: an overview
19
opportunities than in Maddison's work (cf. Table 1.6) - in Table 1.11 this explains about three-fifths of the change in growth rate pre- and post-1973. Such mechanical exercises must, however, be treated with some scepticism, as the failure fully to account for the change in growth before and after the Second World War underlines. A central difficulty is the crudeness of the proxy for human capital formation, which, as with traditional growth accounting, does not encompass on-the-job and vocational training. Also, although the results hint at an enhanced 'social capability' for growth in postwar Europe, they cannot identify its source and in that respect offer no advance on earlier growth accounting. The suggestion that there was something different about the post-1945 growth process is strengthened by explicit consideration of convergence. As has been widely recognized, the evidence is against /?- and cr-convergence prior to 1950 in Maddison's (1991: 53) dataset for the advanced countries. Since 1950 /?- and crconvergence in GDP/hour worked have been strongly present. Conditional convergence since 1960 in the OECD has been extensively analysed by Andres et al. (1996). They broadly support the Mankiw et al. (1992) analysis in that they find a convergence coefficient of about 2 per cent per year and slowly diminishing returns to broad capital, which has an exponent of about 0.8. These results are subject, however, to a number of important qualifications. In particular, Andres et al. find that convergence characterizes the period before 1975 rather than after, and that the model estimates are not stable over subsamples, suggesting differences in the underlying production functions across countries. Indeed, according to Broadberry (1996), at least as far as manufacturing is concerned, the evidence suggests that there is a European 'convergence club', but that North America still has a large productivity lead after the Golden Age. This may prove hard to reduce further as market size and natural resource endowments still play a role, albeit less significant than before the Second World War. In other words, North America may belong to a different 'convergence club', and this suggests that there were limits to the profitability of imitating American methods. Broadberry's long-run data are shown in Table 1.12. This table tends to suggest that catching up in manufacturing was incomplete during the Golden Age, and that other elements such as structural change and reduced productivity gaps elsewhere in the economy also played a part in overall catch-up. An important element in the catch-up growth of the Golden Age was clearly a reduction in technology gaps, which had previously been sustained by protectionism and limited transferability of American know-how (Nelson and Wright, 1992). Verspagen (1995), in an investigation of cr-convergence based on analysis of the effects of patenting on growth, concluded that catching-up effects were related to genuine convergence in technological competence, but that nevertheless serious differences in technological capabilities remain, i.e. that the technology assumptions underlying pure neoclassical growth models did not obtain. These results are in many ways similar to those of Helliwell (1992), whose econometric study found strong convergence effects in technical progress in the OECD during 1963-89, with evidence that these were intensified by increasing openness to international trade. Ben-David (1993) filled out this picture by establishing also that cr-convergence among European countries was particularly marked in the immediate aftermath of episodes of trade liberalization.
20
Nicholas Crafts and Gianni Toniolo
Overall, what picture has emerged from this long-run perspective? The following are the key points. • The central aspect is the highly unusual nature of growth in the Golden Age, 1950-73, which has been revealed by each of the approaches we have considered. The period was one of unprecedentedly rapid catch-up by Europe of the USA and, among the European countries, one which is characterized by both /?- and aconvergence in Barro and Sala-i-Martin's terminology. • The empirical implementation of new growth theory considered in Table 1.11 offers further support for the argument in section 3 that traditional accounts of postwar growth give too little weight to human capital. At the same time, the improvement in growth relative to the interwar period is not fully accounted for by current methods of quantification. • While reconstruction played a part in the Golden Age, catching up involved far more than this. Technology gaps were reduced and investment reached unprecedently high levels. As Abramovitz put it (1986: 395), this was 'the period when - exceptionally - the three elements required for rapid growth by catching up came together... large technological gaps; enlarged social competence... and conditions favouring rapid realization of potential'. • The European countries seem to have been in an Augmented-Solow rather than a Rebelo-type world - in which case the ending of the Golden Age should not have been such a surprise. This seems to be a reasonable inference based on the results of Andres et al. (1996) and Mankiw et al. (1992), together with the general absence of a unit root in GDP. In the longrun, though perhaps not straightaway, weakly diminishing social returns to the postwar investment boom could be expected, and scope for catch-up would be diminished. 6
The years of high growth
As we have seen in the previous sections, the so-called Golden Age (roughly dated 1950-73) is unique in the history of the European economy in terms of both growth rates and cyclical stability. So far we have examined a number of quantitative explanations of the causes of the Golden Age. Some of them are more satisfactory than others. The recent stress on conditional convergence, human capital and endogenous technical change has undoubtedly advanced understanding for this fast-growth episode. Nevertheless, much remains to be explained. In order to fill out the picture, we now consider some key historical and institutional factors that make the 1950s and 1960s a unique period in the economic history of Europe. 6.1
The legacy of the 1930s and of the war
During the interwar years, the aggregate European economy grew at a slower pace than the American, in terms both of output and of productivity. The performance of Europe was also below its pre-1913 record (Maddison, 1991). The features of the European slowdown during the 1930s were likely to give way to a process of catch-up according to a neoclassical view, if we consider the latter as a function of an increasing gap between steady-state and actual output per worker (Barro and
Postwar growth: an overview
21
Table 1.12. Comparative levels of labour productivity in manufacturing (UK output per employee = 100)
UK USA Canada Australia Germany Netherlands Norway Sweden Denmark France Italy Japan
1870
1913
1929
1938
1950
1973
1989
100 204 132
100 213 230 138 119
100 250 256 102 105 102 109 94 115 82 59 32
100 192 218 101 107 117 95 100 98 76 49 42
100 263 227 96 96 88 103 118 88 84 68 20
100 215 229 86 119 133 104 128 89 114 96 95
100 177 185 81 105 128 85 121 93 115 111 143
100
90 102 79 59 24
Source: Broadberry (1996). Sala-i-Martin, 1991). The latter resulted from the slowdown in the accumulation of physical capital not being matched by an equal deceleration in human capital formation. Even during the war, the depletion rate of the two kinds of capital was biased against physical assets, particularly those in the social overhead stock. In addition, diffusion of technology speeded up and American know-how was more fully deployed in Europe. In effect, the new postwar world offered greater opportunities for effective technology transfer (Nelson and Wright, 1992). A typical situation for swift catch-up growth was thus created: once trade and intellectual barriers were removed, the transfer of technology - both embodied and disembodied - was met by an adequate stock of human capital in the recipient countries that made possible a rapid narrowing of the gap between steady-state and actual product per worker. In a more general Abramovitz (1986) framework, one may say that Europe's overall 'social capability' for growth had been hibernating, but had not been destroyed. 6.2
Reconstruction and postwar settlement
Wartime loss of output was considerable. In the largest continental countries (Germany, France and Italy), the productive effort of more than an entire generation was lost as 1945 per-capita incomes stood at around the 1900 levels, as Table 1.2 reported. The speed of'reconstruction' (by which we mean the recovery of the 1938 levels of production per head) took most observers by surprise. Infiveor six years, at most, the losses in output and capital stock were made good. The process of reconstruction does not need to concern the present discussion of longer-run causes of postwar growth, except for noting (1) that it provides yet another proof of the existence of an exceptional 'social capability' for growth almost everywhere in Europe and (2) that it is likely to have set in motion the exploitation of scale economies and thus longer-run increasing returns.
22
Nicholas Crafts and Gianni Toniolo
The social, legal and political institutions, both international and national, created soon after the war are of paramount importance in explaining the extraordinary growth of the 1950s and 1960s. The establishment of such institutions, or the 'postwar settlement' as it has come to be known in the literature (Eichengreen, 1993), is plausibly seen (Maier, 1987) as the terminal point in a process aimed at reversing the crisis that afflicted European liberal capitalism at the end of the nineteenth century. In this view, two related problems needed to be solved during the twentieth century by the elites superintending Western European societies: those of legitimation and those of production. The leaders of the 1920s 'rallied with persuasive justifications of capitalist entrepreneurship', but 'they did not solve the economic dilemma of continuous production and high employment... only by the 1950s were afflictions that undermined capitalist stability effectively overcome as a whole. The cumulative achievement required the institutional flux that was left in the wake not of one but of two wartime upheavals' (Maier, 1987: 162). Rather than simply contrasting the unwise outcome of the Versailles Peace Conference with the prudence of Bretton Woods and of the Marshall Plan, this perspective to twentieth-century history helps in understanding the roots of the second 'postwar settlement' and, thus, its pervasive and lasting impact on European societies. In employing its newly found leadership to 'ensure the primacy of economics over politics and to de-ideologise issues of political economy into questions of output and efficiency', the United States simply encouraged an existing trend, did not impose it (Maier, 1987:146-9). It is then possible to speak of a 'consensual American hegemony', in the sense that a social basis for productivity-enhancing politics had acquired solid foundations in Europe. The 'postwar settlement' consisted of two mutually reinforcing parts: international and domestic. The high investment ratios that characterized the Golden Age rested on more or less explicit social pacts aimed at increasing productivity, whereby workers exercised wage moderation on the understanding that capitalists would plough back their profits into the productive process. The relative degree of economic success of individual countries depended to a fairly large extent on the ability to make the 'pact' appealing and credible in the long run. The international setting was crucial in both respects. On the one hand, it made the pact appealing by guaranteeing a minimum standard of living during the early postwar years through the Marshall Plan and the European Payments Union.12 On the other hand, it created an environment (characterized by stability in domestic prices and exchange rates, and by growing international trade) which made agents confident about the real value of their incomes as well as about their future increases. A somewhat similar pact was reached in Eastern Europe, albeit with a higher degree of coercion. There too production goals acquired primary social relevance. The socialist state, with the crucial political mediation of the party, took upon itself the role of guaranteeing a high rate of investment and full employment in exchange for low wages and social discipline. 6.3
Causes of growth during the Golden Age
We have seen that catch-up was an important element of postwar European growth, particularly during the so-called Golden Age. Nevertheless, we maintain that
Postwar growth: an overview
23
catch-up is by no means an automatic phenomenon. The peculiar economic, institutional and political circumstances created by the Great Depression, autarky, war and reconstruction were particularly favourable to postwar European growth in general and to catch-up in particular. In what follows, we briefly discuss the interplay of favourable factors that made this unique episode in the economic history of Europe possible. Most explanations are unsatisfactory. This is true both of those advanced in the 1960s and of the most recent ones proposed by the 'new growth theory'. A stress on the pure mechanics of 'catch-up' is likewise inadequate. In order to account for the exceptional speed and stability of European economic growth in the 1950s and 1960s, one must take into account the equally exceptional circumstances whereby a number of factors affecting the production system and the macroeconomic structure were allowed to interact with singularly apt institutional arrangements in the international economy and in the individual countries. The main 'virtue' of backwardness lies in the fact that agents can appreciate the enormous gains potentially involved in borrowing technology and capital from more advanced countries, once the 'social capability' for growth is created in the recipient country. The introduction of the Taylorist system looked particularly promising in economies that, by the early 1950s, enjoyed a level of per-capita income (not too far off the US level in the early 1920s) characterized by high income elasticity of demand for consumer durables. Some of these countries had begun experimenting with mass production just before the war (the UK, Germany). The appreciation of the existing potential for growth is witnessed by the number of productivity missions sent to, or invited from, the United States to see 'how American prosperity could be emulated' (Glyn et al.y 1990: 56). In order to take full advantage of the adaptation of American technologies to European conditions, business and trade union practices had to be adjusted accordingly. The speed and the lasting impact of adaptation varied from country to country, but the spread of the new productivity ideology, stressed by Maier, was universal. The 'new international order' was crucial in inducing both entrepreneurs and workers to adopt the necessary changes in their respective behaviours. In particular, it provided both a short-term bridge to the new growth path and a guarantee of its payoff in the long run. It was this extraordinary mix, seldom to be found in other historical circumstances, that triggered the 'virtuous circle' of the Golden Age. It was often the case that governments were faster than either workers or business organizations in appreciating the potential long-run benefits to be drawn from the new situation; almost everywhere they created a social and institutional environment consistent with long-run sustained growth. For imported technology to produce high productivity growth, an educated and disciplined workforce was a necessary precondition. As we have seen, one of the great advantages of Europe over other areas lay precisely in the 'previous accumulation' of good-quality human capital. Some wage moderation, however, was necessary for productivity gains to generate investment ratios high enough to yield economies of scale and further productivity gains. Wage moderation depended (1) on a fairly elastic supply of labour (as described in Kindleberger (1967)), and (2) on the expectation of long-term improvement in the workers' standard of living. The latter condition could be created only by a credible commitment on the part of
24
Nicholas Crafts and Gianni Toniolo
employers to plough back retained profits into investment. A stable international business environment, expectations about a steady decrease of barriers (both tariff and non-tariff) to trade and about the stability of the foreign exchange at a slightly devalued rate made the commitment both possible and credible (Eichengreen, 1996). On the domestic front, conditions for a pledge to high rates of investment were reinforced by a monetary policy which, for being aimed at keeping a fixed parity with the dollar, promised low inflation rates as well. Controls on international capital movements - a feature of the Bretton Woods system - allowed some governments to pursue a policy of'artificially low' interest rates, thereby providing a further stimulus to investment in manufacturing, housing and infrastructure. A commitment to welfare state provisions was also instrumental in moderating wage demands. In terms of new growth theory, these institutional arrangements played a role similar to that of low direct taxation, and indeed operated to counteract the record levels of direct taxation which were a further aspect of the new postwar European scene. Parallel growth of investment, productivity and real wages was a key element in the unique episode in the economic history of mankind that has come to be known as the Golden Age of capitalism. Income distribution was crucial to the continuation of this process. As long as it was consistent with a balanced growth of consumption and investment, high rates of growth in output and productivity could continue in an environment of extraordinary macroeconomic equilibrium, which came to be built into expectations and, therefore, to be an important reinforcing element of the growth process itself. In section 8 below we briefly discuss some of the country-specific features of the process of rapid growth of 1950-73. Here, however, we would like to stress that many of the causes that are likely to explain such process are common to all the countries that shared in it. In particular, the overwhelming importance of the international environment in generating and sustaining growth is witnessed by the fact that, while Spain and Portugal did not experience in 1945 a real break with the past and were throughout ruled by authoritarian regimes, their rates of growth accelerated in response to the decision to link the economic future of both countries to that of the West. Thus, in 1953, with the so-called Pact of Madrid, Spain allowed the United States to open airforce and navy bases in return for technological aid and, more generally, for better prospects of long-term integration in the 'Atlantic economy' (see, for instance, Preston (1991)). Portugal almost doubled its average rate of growth, relative to the previous decade, after joining the Bretton Woods system in I960.13 The Irish case can, perhaps, be taken as a proof a contrario. The Anglo-Irish agreement of 1948 was too narrow in international scope to result in a positive shock to entrepreneurial expectations, while the best performing years of the Irish economy are those that followed the 1958 commitment to trade liberalization (6 Grada and O'Rourke, 1996).14 Relatively rapid productivity growth also surfaced in Eastern Europe during the Golden Age, although here a high rate of capital accumulation enforced by Communist governments loomed large. Indeed, growth in these countries relied heavily on the mobilization of factors of production into the industrial sector, in which central planning tended to promote inefficient use of resources together with inferior innovation performance (Bleaney, 1988: ch. 4). Given these characteristics of the
Postwar growth: an overview
25
Communist bloc, marked tendencies to slowdown towards the end of the Golden Age were to be expected as capital stock growth and capital productivity growth declined and transfers from the backward sector dried up (Ehrlich, 1993: 312-19). 7
Slowdown and instability
After the early 1970s, the pace of expansion of European production and productivity fell back to its secular trend, close to Kuznets' modern economic growth averages. In particular: • The average annual growth rate of GDP fell to 2.6 per cent in 1973-9 and 2.0 per cent in 1979-90; the performance in the latter period was slightly worse than that of the USA (Glyn, 1993: table 1). • Investment rates equally declined by about 3 percentage points for the whole of the business sector and 1.3 per cent for manufacturing in 1979-89 (Armstrong et ai, 1991: tables A5 and A6). • Both GDP per worker and manufacturing output per hour worked growth halved relative to the Golden Age. If seen in a broad secular perspective, the explanandum in postwar European economic history appears to be not so much the slowdown of the 1970s as the growth spurt of the previous two decades. Nevertheless, a large body of literature exists trying to explain the reasons for and the timing of the slowdown as well as why the hopes for a revival of the Golden Age after two oil shocks have proven short-lived.* 5 The reasons for the longer-run unsustainability of the high growth rates of 1950-73 are to be found in the very success of the Golden Age,16 and in the fact that it rested on the simultaneous - and improbable - occurrence of a number of favourable factors. As we have seen, high growth was made possible by the gains deriving from the transfer of the (Taylorist) mass production technology in a receptive (socially capable) environment stabilized by a strong American leadership. This made possible an efficient co-ordination in the bargaining process for income distribution, leading to high and balanced demand for investment and consumption and, therefore, to macroeconomic stability, which in turn reinforced the process in a sort of virtuous circle. As the individual building blocks of this construction interacted in a mutually reinforcing process, so the cracking and eventual crumbling of one or a few of them undermined the solidity of the entire building. Triffin (1957) was the first to point to the self-defeating instability of a system of international payments which postulated both growing American current account deficits and the willingness of the Europeans to hold dollar-denominated reserves rather than gold. It was first of all rapid growth itself - more than American domestic monetary policies - that stressed the stability of the Bretton Woods system and eventually brought it to an end (Bordo, 1993). In a broader perspective, one can say that it was American leadership itself that rested on feet of clay. Soon after the war, its legitimacy hinged on the United States providing economic assistance in return for policy compliance; US hegemony lost some of its legitimacy in European eyes when, rightly or wrongly, they perceived the Bretton Woods system as being utilized 'to exact tribute through seignorage (accumulating dollar liabilities abroad and buying European assets with an overvalued currency)'(Maier, 1987:149). At the
26
Nicholas Crafts and Gianni Toniolo
same time, the pursuit of international leadership slowly lost consensus at home as 'American hegemons... could not demand renunciation on the part of the American working classes for the sake of providing liquidity to the West' (Maier, 1987: 150). International leadership could be exercised only as long as its domestic costs were relatively negligible, and for this very reason could not be maintained in an expanding world. At the same time, returns deriving from high investment rates and from the transfer of technology eventually began to diminish. And, since the (more or less implicit) social pact governing class relations in most European countries rested on the ability of the economy to deliver rapidly expanding standards of living to the workers, it became less and less feasible to maintain the particular income distribution that had made possible both high investment rates and macroeconomic stability. As the latter began to flounder and profit shares to fall, the postwar social pact was virtually over in a number of countries, and a chain reaction of scrambling for income distribution was set in motion. In countries such as France and Italy, where both the economy and social cohesion were weaker, tensions surfaced in the late 1960s: they set in motion a decline in profit rates incompatible with the long-run continuation of the postwar boom, while at the same time tax rates were now rising quickly. If the above factors explain the deterioration of the delicate international and domestic equilibria of the Golden Age, other more specifically economic reasons may be found behind the long-term decline in productivity in Europe, as well as in Japan and the United States. Among these, one might mention: (1) a slowdown in the potential for exploitation of economies of scale, (2) changes in production technology and (3) changes in demand patterns. Taylorism and chain production technology were particularly apt at producing economies of scale internal to thefirm;at the same time they were singularly fit to profit from the possibility of drawing relatively unskilled labour from agriculture to manufacturing. This technology slowly gave way to new productive methods (highlighted, for instance, by the Japanese lean production in the automotive sector), the introduction of which was partly a response to a new consumer sophistication (emphasis on quality, individual specification of products and the like) resulting from growing incomes and, therefore, from Europe's economic success itself. However, the shift to new production processes required long and costly adjustments not only on the shop floor but, much more importantly, in society at large. In particular, a moreflexibleuse of resources (labour) in response to demand and supply shocks was postulated. Such changes took a long time to come about and are not yet fully embedded in the European work culture and organization. At the same time, growing per-capita incomes entailed a demand shift to the service sector. Studies on the latter's productivity do not yield unequivocal results, and show a wide cross-country dispersion both in the levels and in the rates of growth of productivity. They seem, however, to point in the direction of slower productivity growth relative to the manufacturing sector (e.g. Elfring, 1988). From the point of view of catch-up and convergence, post-1973 European growth experience was much less clear-cut than during the Golden Age. In particular: (1) relative to the USA, the productivity catch-up proceeded at a slower rate during the 1970s and stopped, if not reverted, in the following decade (underperformance
Postwar growth: an overview
27
vis-a-vis Japan continued throughout), and (2) evidence of intra-European catch-up is weaker than in the past (Table 1.4), but the previous convergence process within Europe continued. As mentioned earlier, partly as a result of the latter process, around the Netherlands and Sweden's median per-capita income (Table 1.4) a cluster of nine countries now exists, with incomes falling within an interval of ± 8.0 per cent. If we take it, with Blanchard (1991), that capital and labour mobility are likely to produce convergence, the clustering of so many European countries in the same per-capita income range may be seen partly as the result of increased factor mobility within the community, as well as the result of its enlargement from the mid-1970s onward. The standard subdivision of the most recent period in Europe's economic history - 1973-79 and 1979-92 - is still the most useful when focusing on policy making and on institutions (the latter including the rise and fall of the European Monetary System). The 1973-9 period was substantially one of transition and adjustment. The changes in industrial relations, highlighted by the wage explosion and strikes which accompanied the end of the Golden Age, account for most of the double-digit inflation that characterized ten of the sixteen countries in our sample. However, money supply accommodation of distributional conflicts was only a transitory solution. The response to the second oil shock, therefore, showed that an overall change in policy stance was under way in most European countries. Floating exchange rates had created a vicious circle of inflation in the weaker economies, and had added uncertainty to the business environment everywhere due to the much increased level of speculative capital flows. The creation of the European Monetary System (EMS) in 1978 was seen as a response to this situation, in that it required member countries partly to renounce their monetary sovereignty. Similarly, after the dollar crisis of 1978-9, the Federal Reserve inaugurated a period of tight money. This overall change in policy stance opened the way to Industrial restructuring': in some countries (Germany) it went almost unnoticed, being the continuation of a process of mergers and reorganization that had started much earlier; in others (UK, Italy) it was the result of a deliberate policy aimed at increasing productivity. In the latter case, it was made possible by the intervening weakening of the trade unions17 due to high unemployment in the 1970s as well as to broader changes in the composition of the labour force. However, except in the case of the UK where industrial restructuring resulted in higher output per worker, labour productivity growth was on average lower in the 1980s than in the previous decade. 8
Factors in the performance of individual countries
In earlier sections we have emphasized common aspects of postwar European experience. Now we switch the focus to an analysis of'winners' and 'losers', and to briefly considering the contribution made by new growth theory in explaining relative success and failure. Table 1.13 gives a first view of success or failure in the postwar convergence process and takes account of different initial starting points which imply, in an augmented neoclassical model, differential scope for 'catch-up growth'. The estimates are based on the well-known Barro and Sala-i-Martin (1991) model of
28
Nicholas Crafts and Gianni Toniolo
Table 1.13. Actual and forecast convergence, 1950-8617 (GDP I hour) (US'A =100)
Australia Austria Belgium Canada Denmark Finland France Germany Italy Japan Netherlands Norway Sweden Switzerland UK
Actual
Forecast
78 74 86 92 68 67 94 80 79 61 92 90 82 68 80
81 43 60 86 61 48 58 47 48 27 64 61 67 72 73
Argentina Brazil Chile Columbia India Korea Mexico Peru Philippines Taiwan
Actual
Forecast
28 25 33 28 4 21 27 20 11 20
50 29 53 38 8 18 35 34 20 15
Source: Crafts (1992a) based on applying the Barro and Sala-i-Martin (1991) model to data supplied by Maddison. convergence, with its suggestion that on average catching up reduces the productivity gap by 2 per cent per year. Two points stand out in the table. • European countries tend generally to exceed predicted rates of convergence, while the opposite is the case for South America. Perhaps Europe and South America belong to different convergence clubs and have different steady-state income levels; perhaps this table also emphasizes the unusual conjuncture of conditions favourable to catch-up in postwar Europe. • While the model's predictions are unbiased, nevertheless there is a wide variation of performance around the forecast convergence rate. Some European economies did distinctly better than others on this normalized basis: for example, France and the Netherlands outperformed Denmark and the UK. The estimation of catching-up growth models might offer some clues to this experience, and Table 1.14 returns to the Levine and Renelt equation discussed in Table 1.11. This allows comparisons of predictions of growth for each country with the average, based on the factors listed in the table. The model works surprisingly well, given that it was estimated for a much wider sample, and also predicts quite a wide range of outcomes within Western Europe, partly because it embodies quite a strong catching-up effect. To reduce distortions from reconstruction, the Golden Age is viewed in terms of the post-1960 period. Table 1.14 suggests that during the Golden Age differences in initial income level had the largest single effect on relative growth performance, given the large initial dispersion of income levels. As with traditional growth accounting, relatively little difference between countries shows up in education, as enrolment trends were very similar across Europe, although France and Germany on the positive side and Spain on the negative side are noticeable in the 1960-73 period. The UK's weak
Postwar growth: an overview
29
Table 1.14. Growth performance and its sources relative to the European average: an analysis based on Levine-Renelt (% per year)
1960-73 Austria Belgium Denmark Finland France W.Germany Ireland Italy Netherlands Norway Portugal Spain Sweden Switzerland UK 1973-39 Austria Belgium Denmark Finland France W. Germany Ireland Italy Netherlands Norway Portugal Spain Sweden Switzerland UK
Scope for catch-up
Investment Education Public sector
Unexplained
Total
0.24 -0.08 -0.85 0.13 -0.15 -0.66 1.18 0.71 -0.30 -0.51 1.95 1.27 -0.61 -1.78 -0.49
0.24 -0.24 -0.03 0.21 -0.03 0.07 -0.32 0.03 0.08 0.38 -0.02 0.01 -0.12 0.36 -0.55
-0.11 0.19 0.04 0.05 0.55 0.43 -0.20 -0.19 -0.01 0.16 -0.01 -0.34 -0.17 -0.17 -0.08
0.02 0.06 -0.21 0.02 -0.02 -0.08 0.04 0.05 -0.03 -0.09 0.10 0.34 -0.33 0.25 -0.19
-0.32 0.24 0.42 -0.14 -0.28 -0.59 -1.23 -0.23 -0.37 -0.67 0.85 1.89 0.40 0.11 -0.32
0.07 0.17 -0.63 0.27 0.07 -0.83 -0.53 0.37 -0.63 -0.73 2.87 3.17 -0.83 -1.23 -1.63
0.33 -0.20 -0.35 0.17 -0.28 -0.20 1.30 0.36 -0.26 0.08 1.13 0.77 -0.63 -2.05 -0.18
0.22 -0.28 -0.30 0.29 -0.07 -0.15 0.02 -0.03 -0.18 0.47 0.38 0.04 -0.26 0.17 -0.39
-0.02 0.11 0.08 -0.05 0.24 -0.07 0.05 -0.04 -0.01 0.12 -0.18 0.13 -0.03 -0.07 -0.01
0.02 0.14 -0.44 -0.07 0.02 -0.07 0.05 0.18 0.13 -0.09 0.21 0.31 -0.55 0.35 -0.14
-0.20 0.24 0.86 -0.59 -0.06 0.74 -1.07 J 0.08 -0.03 0.47 -1.29 -1.20 0.82 0.65 0.37
0.35 0.05 -0.15 -0.25 -0.15 0.25 0.35 0.55 -0.35 1.05 0.25 0.05 -0.65 -0.95 -0.35
Sources: As for Table 1.11. record in physical investment is highlighted as a problem, as is the relatively large public sector in Denmark and Sweden. In terms of 'unexplained' performance, which presumably reflects aspects of 'social capability' not captured in the regressions, Ireland stands out as a disappointment, but a notable feature is the lack of correlation in this component between the two periods. Once again, it is important to treat the results of this accounting exercise with caution. Clearly, at best this approach cannot deal with issues of endogeneity, nor can it explain why investment strategies differed. Moreover, the proxies for human capital formation (enrolment ratios) are crude and, in particular, ignore training.
30 Nicholas Crafts and Gianni Toniolo
Also the results are somewhat at odds with those of other studies, such as Dowrick and Nguyen (1989), suggesting that in comparisons of the advanced countries the precise attribution of sources of growth is quite sensitive to small changes in specification of the model. Nevertheless, the exercise confirms that normalizing for catch-up potential still leaves something to be explained. We can think of two rather different, although potentially complementary, ways of using new growth theory. The first would be essentially microeconomic in focus and would consider the roles of Marshallian externalities, learning effects and specific technological expertise in fostering differences in comparative advantage and growth potential, as in Lucas (1993). The prospect of serious research along these lines is exciting and may ultimately prove fruitful. At present, however, the second line of research is further advanced and, as set out above, offers the main existing set of empirical evidence. This research looks at the overall accumulation strategies of different countries in an effort to explain differences in growth performance through both the extent and the effectiveness of investment. If catch-up is not automatic and 'social capability' for growth varied within postwar Europe, this may be part of the reason for the results in Tables 1.13 and 1.14. Although a macro perspective will undoubtedly miss part of the story, there are good reasons, in principle, for expecting it to be useful. In particular, it should be noted that relatively weak productivity performance frequently appears to have been a national rather than a sectoral phenomenon (Dosi et ai, 1990). Quite a number of variables have been suggested by researchers working in the new growth economics as additional explanatory factors in winning and losing in the growth league. Data on some of these are reported in Table 1.15. Among these European countries, the differences are generally not very large, and such variables may be much better at separating third world 'failures' from OECD 'successes' than at explaining high investment or 'social capability' for growth in a European context. There are, however, more important reasons for doubting the value of standard new growth variables in providing a full account of postwar European growth based on new growth economics. These appear to be particularly important in the areas of incomplete catching up and incentives to investment. They relate to the roles of institutions and policy in economic growth. Abramovitz, in stressing that catch-up growth is not automatic, noted that it can be blocked by 'vested interests, established positions, and customary relations amongfirmsand between employers and employees' (1986:389). As noted in section 3, attempts to operationalize this proposition in regressions following up sclerosis in an Olsonian framework have not been successful. Such influences may in fact not be amenable to quantification and may, in the end, be unduly neglected by analysis of this kind. In the case of incentives to invest, the position is more complicated and less well understood. The potential value of new growth theory is to place investment and the growth to which it gives rise in an optimizing framework, in which success and failure are the outcome of rational behaviour rather than the laundry list of scapegoats provided by traditional economic history. In such an approach, institutions and government policy will surely matter, but may not be easy to model. For example, empirical investigations of the effects of taxation on growth have
Postwar growth: an overview
31
Table 1.15. Some new growth variables LLY (1964/5) Austria Belgium Denmark Finland France Germany Italy Netherlands Norway Sweden Switzerland UK
0.48 0.46 0.46 0.39 0.42 0.65 0.59 0.52 0.47 1.01 0.35
Td/Y(%) RD/Y(%) (1960)
(1970)
19.7 16.9 15.9 14.9 21.5 21.6 20.4 21.7 17.9 19.0 10.6 19.1
0.61 1.31 0.96 0.78 1.91 2.06 0.88 2.01 1.10 2.25 2.18
Sources: LL Y is ratio of liquid liabilities to GDP as denned by King and Levine (1993), Td/Y is direct taxes/GDP from OECD (1981), RD/Yis research and development spending/GDP from OECD (1991c). provided much less decisive results than one would expect from the a priori reasoning of new growth theory (Easterly and Rebelo, 1993). Part of the reason for this may be that readily available measures are of average rather than marginal tax rates. Also, ex-ante capital taxation rates depend crucially on expectations of inflation, as the work of King and Fullerton (1984) emphasizes. Bean and Crafts (1996) and Eichengreen (1996) both explore a related point: namely, the possibility of firms and their workers making credible commitments to a high investment/wage moderation equilibrium. They recognize that government may have an important role in brokering a 'social contract' and ensuring appropriate industrial relations. Such 'deals' in postwar Europe were likely to involve higher taxes and welfare spending as part of the package, which, in context, were investment enhancing rather than the opposite. In general, recent work on investment has stressed the importance of irreversibilities and the sensitivity of required rates of return to the degree of uncertainty in the economic environment (Pindyck and Solimano, 1993). All these considerations will not be captured by a conventional 'tax rate' measure, but should, in principle, be subsumed in the ex-ante hurdle rates imposed on investment decisions in different countries in different time periods. Such data would capture the full flavour of recent theorizing, but unfortunately do not presently exist. 9
Concluding comments
Recent developments in growth economics have changed the way we think about the postwar period, but they have by no means yet provided us with a complete
32
Nicholas Crafts and Gianni Toniolo
understanding of the Golden Age and subsequent slowdown. Nevertheless, current models and empirical techniques allow more plausible insights into the experience than was possible with the older theories reviewed in section 3. The main findings of this overview essay are the following: • It is useful to think in terms of epochs of growth. In that context, the years c. 1950-73 stand out as a time when European growth was exceptionally rapid and the economic environment favoured rapid catch-up and convergence. Postwar reconstruction of the international economy delivered a large positive shock to the European economy. • New growth theory seeks to place more emphasis on factor accumulation and less on unexplained TFP growth in explaining long-run growth. Empirical investigation using the Levine and Renelt (1992) approach supports this interpretation of the acceleration of growth postwar and, in particular, suggests that human capital formation deserves much greater weight than it was accorded by earlier writers. Obtaining better measures of human capital formation is a high priority because it is in this area that empirical work suggests that externalities may be especially important, whereas it seems that for physical capital these are trivial (Oulton and O'Mahony, 1994). • Nevertheless, a significant fraction of the acceleration in growth during the Golden Age does not seem to be readily accounted for by new growth models. This would appear to result from changes facilitating technology transfer and raising returns to investment. There may be a role here for some of the policy changes stressed by early postwar accounts, including trade liberalization and demand stabilization, in eliminating obstacles to the faster growth which had prevailed between the wars. • The slowdown after 1973 appears less surprising when viewed in a long-run context and against the evidence, which suggests that in the end there are diminishing returns to investment in broad capital and that the notion of reverting to earlier trend rates of growth still has some validity. • A full understanding of Europe's Golden Age of economic growth requires a subtle appreciation of the impact of policy and institutions on incentives to invest and obstacles to complete catching up. Recent work in growth theory provides new insights which will require detailed historical research before their full implications can be brought out. NOTES This paper is produced as part of a CEPR Research Programme on Comparative Experience of Economic Growth in Postwar Europe, supported by a grant from the
Commission of the European Communities under its SPES Programme (no. SPESCT 910072). We received very helpful comments from Steve Broadberry, Mark Harrison, Brendan Walsh and members of seminars at University College Dublin, Strathclyde, North Carolina State at Chapel Hill and Warwick Universities as well as at the CEPR Conference, 'Europe's Postwar Growth' at Oxford, December 1993. We are to blame for any errors. 1 Western Europe is by and large the 'unit of observation'of this book. In this context it comprises the twelve countries mentioned in Table 1.1. Further on we
Postwar growth: an overview
2
3 4 5
6 7
8 9
33
shall refer to Europe as a unit including sixteen countries (the above-mentioned twelve plus Spain, Portugal, Ireland and Greece).Were data for these countries available and reliable for the whole 1870-1993 period, their inclusion would not have greatly altered the figures in Table 1.1, their demographic weight being roughly 15 percent of the total. See, for instance, Kuznets (1965:48-58). He argues that modern economic growth necessarily produces a shift in the disparities in economic and social conditions among nations due to 'rapid growth of new units or relatively rapid decay of others. [Such shifts] often produce elements of aggression by which the new leaders attempt to claim the prerequisites of economic power, the old and surviving leaders attempt to deal with the new weakness that may have arisen' (1965:50). Given the wide scope for measurement errors and sensitivity of growth rates to the selection of beginning and end years, the difference between the long-run performance and that of the post-1973 years is hardly significant. Between 1870 and 1890 the rate of growth of per-capita output in the twelve countries included in Table 1.1 was only 0.7 per annum (data from Maddison (1991)). Reference here is made to the highest peacetime income, since wartime production (which peaked in 1942 or 1943) is at best of dubious significance in terms of long-run growth in income and welfare, particularly as far as defeated countries are concerned. Discussion of business cycles is beyond the scope of this paper: suffice it to say that never during the period did any European country suffer a decline in GDP. The average inflation rate in the four largest European countries (France, Germany, Italy and United Kingdom, totalling about threequarters of the population of the twelve countries in Table 1.1) between 1950 and 1973 was 4.0 per cent when measured by consumer price indices. It was 8.4 between 1973 and 1989 (10.0 if Germany is excluded). The 1950-73 record compared unfavourably with the virtual absence of inflation in the two decades preceding the First World War, when consumer prices rose on average by 0.6 per cent per annum in the four countries during 1890-1913. (Data from Maddison (1991: 296-307). Japan's growth in GDP per capita between 1950 and 1973 was 8.0 percent per annum, two and a half times faster than Japan's own secular trend (1900-92), tentatively put at around 3.1 per annum (Maddison, 1995). Rates of growth in real GDP per capita Country
1900-92
Australia Canada
1.5 2.1 1.8
USA
1950-73 2.4 2.9 2.4
1973-92 1.4 1.7 1.4
Source: Maddison (1995). All three countries show (1) a secular rate of growth within Kuznets' 'norm' and not far from the European average, and (2) much less pronounced acceleration and deceleration during 1950-73 and 1973-92 than was the case with the European economy as a whole. 10 The same twelve as in Table 1.1. 11 France, Germany, Italy and the United Kingdom, totalling between 75 and 80 per cent of the aggregate labour force for the twelve countries in Table 1.1. Given the rather similar size of the three labour markets, arithmetic averages provide a
34 Nicholas Crafts and Gianni Toniolo fairly accurate assessment of the unemployment rate for the four-country aggregate. 12 The latter mattered in as much as it allowed the avoidance of a choice between devaluation and domestic deflation, had full convertibility been chosen early in the game (Eichengreen, 1993: 86-9). 13 Portugal's average growth rate in GDP per capita was 3.87 percent in 1950-60 and 6.45 percent in 1960-73. 14 The commitment was taken in the Government's Programme for economic expansion in 1958, later known as the First Programme (6 Grada and O'Rourke, 1996). 15 See, for instance, the excellent bibliography in Marglin and Schor (1990). 16 The catch-up process by its own workings undermined the bases of the growth boom to which it gave rise' (Abramovitz, 1990:9). 17 Judging from the number of days' strikes per 100 employees, this was certainly the case in the UK, France and Italy, where they declined sharply in the 1980s relative to the 1970s. They remained stable in Germany and the Netherlands and rose in the Scandinavian countries. (Data on strikes from Glyn (1993)). REFERENCES Abramovitz, M. (1986) 'Catching up, forging ahead and falling behind', Journal of Economic History, 46, pp. 385-406. (1990) The catch up factor in postwar economic growth', Economic Inquiry, 38, pp. 1-18. Andres, J., R. Domenech and C. Molinas (1996) 'Growth, convergence and macroeconomic performance in OECD countries: a closer look', in B. van Ark and N.F.R. Crafts (eds.), Quantitative Aspects of Postwar European Economic Growth, Cambridge: Cambridge University Press. Armstrong, J., A. Glyn and J. Harrison (1991) Capitalism since 1945, Oxford: Blackwell. Balassa, B. (1979) 'Export composition and export performance in the industrial countries, 1953-71', Review of Economics and Statistics, 61, pp. 604-7. Barro, R. and X. Sala-i-Martin (1991) 'Convergence across states and regions', Brookings Papers on Economic Activity, pp. 107-82. Bean, C. and N.F.R. Crafts (1996) 'British economic growth since 1945: relative economic decline... and renaissance?', in N.F.R. Crafts and G. Toniolo (eds.), Economic Growth in Europe since 1945, Cambridge: Cambridge University Press. Beckerman, W. (1962) 'Projecting Europe's growth', Economic Journal, 72, pp. 912—25. Ben-David, D. (1993) 'Equalizing exchange: trade liberalization and income convergence', Quarterly Journal of Economics, 108, pp. 653-79. Blanchard, O. (1991) 'Comment', on Barro and Sala-i-Martin (1991). Bleaney, M. (1988) Do Socialist Economies Work?, Oxford: Blackwell. Boltho, A. (ed.) (1982) The European Economy: Growth and Crisis, Oxford: Oxford University Press. Bordo, M. (1993) The Bretton Woods international monetary system: an historical overview', in M. Bordo and B. Eichengreen (eds.), A Retrospective on the Bretton Woods System, Chicago, IL: Chicago University Press. Broadberry, S.N. (1996) 'Convergence: what the historical record shows', in B. van Ark and N.F.R. Crafts (eds.), Quantitative Aspects of Postwar European Economic Growth, Cambridge: Cambridge University Press. Castles, F.G. and S. Dowrick (1990) The impact of government spending levels on medium-term economic growth in the OECD, 1960-85', Journal of Theoretical Politics, 2, pp. 173-204.
Postwar growth: an overview
35
Chatterji, M. and M. Wickens (1983) 'Verdoorn's Law and Kaldor's Law: a revisionist interpretation', Journal of Post-Keynesian Economics, 5, pp. 397-413. Crafts, N.F.R. (1992a) 'Productivity growth reconsidered', Economic Policy, 15, pp. 387^26. (1992b) Institutions and economic growth: recent British experience in an international context', West European Politics, 15 (4), pp. 16-38. Crafts, N.F.R. and T.C. Mills (1996) 'Europe's Golden Age: an econometric investigation of changing trend rates of growth', in B. van Ark and N.F.R. Crafts (eds.), Quantitative Aspects of Postwar European Economic Growth, Cambridge: Cambridge University Press. Cripps, T.F. and R.J. Tarling (1973) Growth in Advanced Capitalist Economies, 1950-70, Cambridge: Cambridge University Press, den Bakker, G.P., T.A. Huitker and C.A. van Bochove (1990) T h e Dutch economy 1921-39: revised macroeconomicdata for the interwar period', Review of Income and Wealth, 36, pp. 187-206. Denison, E.F. (1967) Why Growth Rates Differ, Washington, DC: Brookings Institution. Dosi, G., K. Pavitt and L. Soete (1990) The Economics of Technical Change and International Trade, London: Harvester Wheatsheaf. Dowrick, S. (1992) 'Estimating the impact of government consumption on growth', Australian National University Discussion Paper No. 258. Dowrick, S. and D.T. Nguyen (1989) 'OECD comparative economic growth 1950-85: catch-up and convergence', American Economic Review, 79, pp. 1010-30. Durlauf, S.N. and P.A. Johnson (1992), 'Local versus global convergence across national economies', NBER Working Paper No. 3996. Easterly, W. and S. Rebelo (1993) 'Fiscal policy and economic growth: an empirical investigation', Journal of Monetary Economics, 32, pp. 417-58. Ehrlich, E. (1993) 'Economic growth in Eastern Central Europe after World War IF, in A. Szirmai, B. van Ark and D. Pilat (eds.), Explaining Economic Growth, Amsterdam: North-Holland. Eichengreen, B. (1993) Reconstructing Europe's Trade and Payments, Manchester: Manchester University Press. (1996) 'Institutions and economic growth: Europe after World War IF, in N.F.R. Crafts and G. Toniolo (eds.), Economic Growth in Europe since 1945, Cambridge: Cambridge University Press. Elfring, T. (1988) Service Sector Employment in Advanced Economies: A Comparative Analysis of its Implications for Economic Growth, Aldershot: Gower. Feinstein, C.H. (1972) National Income, Expenditure and Output of the UK 1855-1965, Cambridge: Cambridge University Press. Glyn, A. (1993) 'Stability, inegalitarianism and stagnation: an overview of the advanced capitalist countries in the 1980s', mimeo. Glyn, A., A. Hughes, A. Lipietz and A. Singh (1990) 'The rise and fall of the Golden Age', in S. Marglin and J. Schor (eds.), The Golden Age of Capitalism, Oxford: Clarendon Press. Gomulka, S. (1971) Inventive Activity, Diffusion and the Stages of Economic Growth, Aarhus: Institute of Economics, Aarhus University. Hansen, S.A. (1974) 0konomisk voekst i Danmark, Copenhagen: Institute of Economic History. Helliwell, J.F. (1992) 'Trade and technical progress', NBER Working Paper No. 4226. Hjerppe, R. (1989) The Finnish Economy, 1860-1985, Helsinki: Bank of Finland. Houthakker, H.S. and S.P. Magee (1969) 'Income and price elasticities in world trade', Review of Economics and Statistics, 51, pp. 111-25.
36
Nicholas Crafts and Gianni Toniolo
Janossy, F. (1969) The End of the Economic Miracle, White Plains, NY: ISAP. Kaldor, N. (1966) Causes of the Slow Rate of Growth of the United Kingdom, Cambridge: Cambridge University Press. Kindleberger, C.P. (1967) European Postwar Growth: The Role of Labor Supply, Cambridge, MA: Harvard University Press. King, M.A. and D. Fullerton (1984) The Taxation of Income from Capital, Chicago, IL: Chicago University Press. King, R.G. and R. Levine (1993) 'Finance and growth: Schumpeter might be right', Quarterly Journal of Economics, 108, pp. 717-37. Klacek, J., M. Hayek and V. Nachtigal (1993) 'Long term economic growth in Czechoslovakia', mimeo., Prague. Kormendi, R.C. and P.G. Meguire (1985)'Macroeconomic determinants of growth', Journal of Monetary Economics, 16, pp. 141-63. Krantz, O. and C.A. Nilsson (1975) Swedish National Product, 1861-1970, Lund: CWK Gleerup. Krugman, P. (1989) 'Differences in income elasticities and trends in real exchange rates', European Economic Review, 33, pp. 1031-54. Kuznets, S. (1965) Economic Growth and Structure, London: Heinemann. Lamfalussy, A. (1963) The United Kingdom and the Six, London: Macmillan. Layard, R., S. Nickell and R. Jackman (1991) Unemployment, Oxford: Oxford University Press. Levine, R and D. Renelt (1992) 'A sensitivity analysis of cross-country growth regressions', American Economic Review, 82, pp. 942-63. Lucas, R.E. (1993) 'Making a miracle', Econometrica, 61, pp. 251-72. Maddison, A. (1987) 'Growth and slowdown in advanced capitalist economies', Journal of Economic Literature, 25, pp. 649-98. (1991) Dynamic Forces in Capitalist Development, Oxford: Oxford University Press. (1992) 'A long run perspective on saving', Scandinavian Journal of Economics, 94, pp. 181-96. (1995) Monitoring the World Economy, Paris: OECD. Maier, C. (1987) In Search of Stability: Explorations in Historical Political Economy, Cambridge: Cambridge University Press. Mankiw, N.G., D. Romer and D.N. Weil (1992) 'A contribution to the empirics of economic growth', Quarterly Journal of Economics, 107, pp. 407-37. Marglin, S. and J. Schor (1990) The Golden Age of Capitalism, Oxford: Clarendon Press. Mitchell, B.R. (1988) European Historical Statistics, London: Macmillan. Nelson, R.R. and G. Wright (1992) T h e rise and fall of American technological leadership', Journal of Economic Literature, 30, pp. 1931-64. OECD (1981) Long Run Trends in Tax Revenues of OECD Member Countries, Paris. (1991a) Economic Outlook, Paris. (1991b) Historical Statistics, 1960-90, Paris. (1991c) Main Science and Technology Indicators, Paris. (1993) Economic Outlook, No. 54, December, Paris. Ofer, G. (1992) 'Productivity, competitiveness and the socialist system', in B.G. Hickman (ed.), International Productivity and Competitiveness, New York: Oxford University Press. 6 Grada, C. and K. O'Rourke (1996) 'Irish economic growth, 1945-88', in N.F.R. Crafts and G. Toniolo (eds.), Economic Growth in Europe since 1945, Cambridge: Cambridge University Press. Olson, M. (1982) The Rise and Decline of Nations, New Haven, CT: Yale University Press.
Postwar growth: an overview
37
(1996) The varieties of Eurosclerosis', in N.F.R. Crafts and G. Toniolo (eds.), Economic Growth in Europe since 1945, Cambridge: Cambridge University Press. Oulton, N. and M. O'Mahony (1994) Productivity and Growth, Cambridge: Cambridge University Press. Paque, K.-H. (1994) 'Why the 1950s and not the 1920s?', CEPR Discussion Paper No. 981. Pindyck, R.S. and A. Solimano (1993) 'Economic instability and aggregate investment', NBER Macroeconomics Annual, 7, pp. 259-303. Prais, S. (1993) 'Economic performance and education: the nature of Britain's deficiencies', National Institute Discussion Paper No. 52. Preston, P. (1991) 'Spain', in A. Graham and A. Seldon (eds.), Government and Economies in the Post-war World, London: Routledge.
Rebelo, S. (1991) 'Long run policy analysis and long run economic growth', Journal of Political Economy, 99, pp. 500-21.
Ritschl, A. (1996) 'An exercise in futility: East German economic growth and decline, 1945-89', in N.F.R. Crafts and G. Toniolo (eds.), Economic Growth in Europe since 1945, Cambridge: Cambridge University Press. Romer, P.M. (1994) The origins of endogenous growth', Journal of Economic Perspectives, 8(1), pp. 3-22. Rossi, N., A. Sorgato and G. Toniolo (1992) 'Italian historical statistics: 1880-1990', mimeo., University of Venice. Rowthorn, R.E. (1975) 'What remains of Kaldor's Law?', Economic Journal, 85, pp. 10-19. Sommariva, A. and G. Tullio (1986) German Macroeconomic History, 1880-1979,
London: Macmillan. Thirlwall, A.P. (1979) The balance of payments constraint as an explanation of international growth rate differences' Banca Nazionale del Lavoro Quarterly Review, 128, pp. 44—53. Triffin, R. (1957) Europe and the Money Muddle, New Haven, CT: Yale University Press. Unger, B. and F. van Waarden (1994), 'Interest associations and economic growth: a critque of Mancur Olson's Rise and Decline of Nations', CEPR Discussion Paper No. 894. van de Klundert, T. and S. Smulders (1992) 'Reconstructing growth theory: a survey', De Economist, 140, pp. 177-202. Verspagen, B. (1996) Technology indicators and economic growth in the European area: some empirical evidence', in B. van Ark and N.F.R. Crafts (eds.), Quantitative Aspects of Postwar European Economic Growth, Cambridge:
Cambridge University Press.
Villa, P. (1993) Une Analyse Macroeconomique de la France an XXe Siecle, CNRS
Editions.
2
Institutions and economic growth: Europe after World War II BARRY EICHENGREEN
1
Introduction
The quarter century that ended around 1973 was for Western Europe a Golden Age of economic growth. Real GDP rose nearly twice as rapidly as over any comparable period before or since.l Understanding the sources of this admirable performance would shed important light on the causes of the slowdown through which Europe has suffered subsequently. Part of the explanation is surely 'catch-up', as Abramovitz (1986) emphasized. The gaps that had opened up vis-d-vis both the United States and Europe's own prewar trend as a result of nearly two decades of depression and war offered exceptional scope for growth after 1945. Figures 2.1 and 2.2 summarize the standard evidence on the operation of these effects.2 But cross-section regressions relating growth rates to per-capita GDP differentials show that 'catch-up' and 'spring-back' explain only part of the postwar acceleration: purged of their effects, growth from 1950 to 1973 was still more than 50 per cent faster than it became subsequently.3 And even in so far as these factors provide the explanation, understanding what enabled post-World War II Western Europe so effectively to exploit the opportunity for closing the gap can have important implications for countries in Eastern Europe and the developing world currently seeking to join the 'convergence club'. Aside from catch-up, the proximate cause of postwar Europe's growth miracle was high investment. Net investment rates in Europe were nearly twice as high in the 1950s and 1960s as before or since.4 Scatter plots like those of Figure 2.3 suggest that increasing the gross investment share of GDP from 20 to 30 per cent increased the growth rate by as much as 2 percentage points. Countries with different investment rates were not identical in other respects, of course; multivariate regressions for this period, using Maddison's sample of sixteen advanced countries, which attempt to control for some of these differences, suggest that an extra 10 points on the investment rate translated into a more modest half a point on the growth rate.5 Still, together with catch-up this gets us a long way towards 'explaining' in an accounting sense postwar Europe's rapid growth. 38
Institutions and economic growth JUU
-
250-
• Germany • Austria • Italy \
\ • France 200- Finland \ ,
O
B e l g i u m ^ Netherlands Denmark* • \. Norway y * \^ 150" Sweden \^ Switzerland •
\. • UK
100"
\ • USA
\ \
50100
200
300 Initial
400
500
Source: Maddison (1982: 212). Figure 2.1 Initial (1950) GDP per man-hour and growth, 1950-73 J\J\J
\^
• Germany N.
• Austria
• Italy
250Finland ^ y 2002 O 150"
• France \
J^Netherlands Belgium • Denmark Norway N. •• Sweden Switzerland •
> • UK
100-
USA^V 50-25
\
i
25 Shortfall
50
75
Source: Maddison (1982: 212). Figure 2.2 Wartime and postwar (1938-50) change in GDP per man-hour and growth, 1950-73
39
40
Barry Eichengreen o.io0.090.08-
0 0.07-
1 0.06|
0.05 0.040.03 0.02 0.15
0.20
0.25 Investment share
0.30
0.35
Regression /-statistic 3.00. Finland from 51, Japan 52. (a) 1950-73
0.045 0.0400.035 Q
0.030 0.025 0.020 0.015 0.010 0.15
0.20
0.25 Investment share
0.30
0.35
Regression /-statistic 3.31. (b) 1973-89 Figure 2.3 Investment share and growth rates of GDP: 16 countries, 1950-89
Institutions and economic growth
41
Two things then remain to be understood: what made this high investment possible, and what made it so productive? This directs our attention to the other elements of the postwar growth recipe: wage moderation and export growth. Wage moderation stimulated both the supply of and the demand for investment - demand by making investment profitable, supply by making available the profits to finance it.6 The openness of European economies and the growth of their exports, whose volume expanded in the 1950s and 1960s by more than 8 per cent a year, allowed investment to be allocated to the sectors where its contribution to productivity growth was greatest. Nations could exploit their comparative advantage without being constrained by the composition of domestic demand. (Figure 2.4 displays the export-GDP growth nexus, indicating, suggestively, that it collapsed after 1973.) Having stripped another layer off the onion, what next must be explained is wage moderation and the growth of trade. Both were achievements by the standards of the interwar period, when exports stagnated and wage pressure was intense.7 Part of the explanation is surely that postwar policy-makers and market participants had learned from the disasters of the interwar years and determined not to repeat them. But the desire for a better outcome may not suffice; mechanisms are needed to achieve it. The argument of this chapter is that postwar growth benefited from the presence of institutions singularly well suited to reconstruction and growth. Those institutions solved commitment and coordination problems in whose presence neither wage moderation nor the expansion of international trade could have taken place. On the domestic side, social and economic institutions disseminated information and monitored the compliance of economic interest groups with the terms of their agreement to moderate wage claims and boost investment.8 They helped to lock in the bargain by creating bonds that would be lost in the event that someone reneged. They worked to coordinate the terms of the agreement across sectors of the economy. On the international side, institutions were created to coordinate national programmes of economic restructuring along export-oriented lines, and to lend credibility to European governments' commitment to openness. This encouraged countries to restructure their economies and to exploit more fully their comparative advantages, enhancing the productivity and profitability of investment. Institutions were not equally well adapted to the needs of growth in all European countries. Some, notably the UK and Ireland, failed to develop the relevant domestic institutions. Others, such as France and Italy, managed to do so only with delay. Some countries, like France in the 1950s, failed to align domestic arrangements with the evolving international framework and to exploit the opportunities offered by the external sector. These different institutional responses go a fair way towards accounting for variations across countries and over time in European growth performance. To complete the story it is necessary to account for these different responses. The argument of this chapter is that the relevant socioeconomic institutions necessarily displayed considerable inertia. Their function, in part, being to serve as coordinating mechanisms, their very nature created coordination problems for altering them. Institutions function as standards, giving rise to network externalities that tend to lock in their operation. The exceptional circumstances of war and reconstruction provided singular opportunities for coordinating wholesale adjustments in institutional
42
Barry Eichengreen 0.10-j 0.090.08-
§ 0.07 -I 3 0.062 |
0.05 0.040.030.02 0
0.05
0.10 0.15 Growth rate of exports
0.25
0.20
Regression ^-statistic 5.35. (a) 1950-73
0.0450.040 0.035 -
i
0.030 0.025 0.020 0.015 0.010 0.035
0.040
0.045
0.050 0.055 0.060 Growth rate of exports
0.065
0.070
Regression f-statistic 1.51. (b) 1973-89 Figure 2.4 Growth rates of GDP and exports: 16 countries, 1950-89
0.075
Institutions and economic growth
43
arrangements. Even under these extraordinary conditions, however, radical changes in coordinating institutions were necessarily difficult to organize. Inevitably, the important institutional changes of the postwar period were only marginal adaptations. They were feasible only where considerable progress had already been made in developing the institutional structures required for growth after World War II. Because of the evolving structure of the institutions analysed in this chapter, it is important to be precise about the period under consideration. I concentrate here on the institutional arrangements of the recovery period (1945-9) and thefirstpostwar decade of sustained growth (1950-9), and on the critical departures from pre-1945 institutional arrangements that made possible the inauguration of Europe's postwar golden age. Subsequently, one sees in many countries the institutional structures of the postwar years weaken and break down. I examine institutional responses to these difficulties in the 1960s, which typically involved heightening the centralization of wage bargaining and expanding the role of government in concertizing sectoral negotiations. I describe the problem to which these initiatives were the response - how the end of 'catch-up' growth made cooperative agreements more difficult to sustain - and explain why the response ultimately proved inadequate. My goal is to show that the perspective developed here is capable of explaining both why postwar institutions were initially so conducive to growth and why they ultimately broke down.
2
The model
2.1
Domestic institutions
Lancaster (1973), Grout (1984) and van der Ploeg (1987) model a dynamic game between capital and labour with a common general structure.9 Welfare is maximized when capitalists and workers both agree to defer current compensation in return for future gains. Workers moderate their wage claims in order to make profits available to enterprise and to make profitable their investment in capacity modernization and expansion. Capitalists restrain dividend payout in order to reinvest. Investment stimulates growth, raising the future incomes of both capitalists and workers. In the cooperative equilibrium in which both workers and capitalists exercise restraint, the costs of forgoing current consumption are dominated by the benefits of the future increase in incomes accruing to both. This cooperative equilibrium may be impossible to sustain, however, for the sequencing of events renders it time inconsistent. Consider the problem for labour created by uncertainty concerning subsequent investment. If investment requires liquidity, and liquidity requires profits, then workers must restrain wages now in order to make profits available to capitalists for investment later. Once the wage restraint has occurred, however, capitalists are best off if they renege on their agreement to invest, paying out profits as dividends instead. Since investment is no higher than if workers had failed to moderate their wage demands, labour has no incentive to exercise restraint. In this non-cooperative equilibrium, workers pursue wage increases, management pays out profits as dividends, and investment and growth are depressed. Even if workers can be assured of capital's willingness to invest, this is not a
44
Barry Eichengreen
complete solution to the problem. Unions may be able to recontract after investment has taken place. Once the investment has occurred, workers can renege on their agreement to restrain wages, seeking to capture the surplus instead. Since profits are not higher than if management had failed to invest, management has no incentive to plough profits into investment. Again, in the non-cooperative equilibrium, workers pursue wage increases, management pays out profits, and investment and growth lag. A contract that binds capitalists to invest profits and workers to exercise wage restraint overcomes the problem of dynamic inconsistency, rendering both groups better off. The social and economic institutions developed in Europe after World War II can be thought of as mechanisms to enforce this agreement.10 Institutions were developed to monitor the compliance of capitalists with their deferred contribution to the bargain and to disseminate evidence of non-cooperation; by reducing the likelihood that shirking went undetected, these mechanisms reduced the returns to doing so. Institutions were used to create bonds that would be lost in the event of reneging, effectively increasing the stakes and providing a further deterrent to shirking. By credibly committing capital to invest the profits made available by wage moderation, they provided labour with the incentive to be moderate. Institutions also deterred workers from appropriating the profits of firms that undertook investment Long-term contracts, social pacts between labour, management and government, and statutory wage and price controls are three mechanisms that could be used to precommit unions to wage moderation and thereby to induce management to invest. Unemployment, health and retirement programmes - the institutions of the welfare state, in other words - served as bonds that would be jeopardized if labour reneged. Centralization and concertation of sectoral wage negotiations further encouraged moderation on the part of unions. In so far as onefirm'searnings could pass through the capital market and finance another firm's investment, the benefits of wage moderation by one group of workers then accrued to other workers. Since the level of wages affected economy-wide determinants of investment like the interest rate, there was a need to coordinate wage demands across sectors to render a bargain to moderate wage claims attractive to each party to negotiations.l x Hence the need for institutions to centralize or concertize sectoral bargaining. On the employer side, any one firm contemplating investment had reason to worry that its decision to invest would encourage its workers to raise their wage demands in order to appropriate the profits thereby generated. But if wages were determined in economy-wide rather than at enterprise-level negotiations, an individual firm's investment decision no longer affected the wages it had to pay. In these circumstances, centralized wage negotiations led to a higher level of investment and, in so far as labour productivity was raised, to higher wages in equilibrium.12 2.2
International institutions
For deferring consumption to be worthwhile, investment has to be productive. For investment to stimulate growth, in other words, there has \o be a market for the goods produced by domestic industries whose capacity is augmented and whose efficiency is enhanced. Here the postwar expansion of trade was key. International
Institutions and economic growth
45
trade, and intra-European trade in particular, allowed countries to specialize in the production of goods in which they had a comparative advantage without regard to limits on the demand for those products existing at home. It allowed them to rely on cheap foreign supplies of raw materials and intermediate inputs that were costly to produce domestically. But the expansion of trade created further coordination and commitment problems. Restructuring along export-oriented lines was costly. Sinking the costs of reallocating resources along lines of comparative advantage could turn out to be an expensive mistake if one's trading partners reneged on their commitment to openness. Encouraging the expansion of steel production on the assumption that coal and iron ore could be imported from abroad could be a costly error if foreign supplies were not forthcoming. Augmenting the capacity of such industries would not pay if other countries ultimately refused to draw down their import tariffs on final goods. Before encouraging the rationalization of domestic production along lines of comparative advantage, governments had to be convinced that their partners' commitment to openness was permanent. Here again institutions solved these commitment and coordination problems. The European Coal and Steel Community (ECSC) created monitoring and surveillance technologies that guaranteed the French steel industry access to German coal, and the German industry access to French iron ore. A Joint High Authority monitored the compliance of participating countries to the terms of the agreement. The European Payments Union (EPU) coordinated the simultaneous move of European countries to currency convertibility for intra-European current account transactions and precommitted the participants to a sequence of trade liberalization measures. An EPU Managing Board was created to monitor the policies of member countries and to discourage them from reneging on their commitments. The participants contributed hard currency and credit to the EPU's central fund; access to these resources was contingent on their adherence to the terms of the EPU agreement, which served as an effective bond. Compared to unilateral convertibility, then, the payments union was a more credible commitment mechanism.13
3
The structure of domestic institutions
3.1
The nature of the bargain
That wage restraint and high investment were the dual cornerstones of the postwar settlement is evident in statements of the time. The Dutch case is prototypical. The Netherlands perceived itself as suffering from overpopulation (partly because of the return of expatriates from Indonesia), the solution for which was wage moderation and investment to enhance international competitiveness and raise living standards (Shonfield, 1965: 212). Between 1947 and 1954, consumer prices and the gross weekly earnings of adult male workers rose in lock step - by 32 and 33 per cent, respectively - this despite the fact that productivity increased by nearly 50 per cent between 1950 and 1954 alone, very considerably raising capital's share of national income (Flanagan et a/., 1983: 107).14 In Germany, trade unions observed 'significant wage restraint' throughout the 1950s in conjunction with the adoption
46
Barry Eichengreen
of investment-friendly policies designed to stimulate growth (Maier, 1984). Norwegian unions followed policies of wage moderation into the 1950s, in return for government policies attaching an overriding priority to capital formation. Even in the UK, hardly a paragon of labour/management harmony, contemporary observers were impressed by 'the wisdom, the moderation and sense of responsibility of the great bulk of the trade union movement' (to quote Sir Walter Monckton, Churchill's Minister of Labour, in 1954).15 This wage moderation was part of a bargain in which labour agreed to restraint now in return for more investment, faster growth and higher incomes later. In the Netherlands, the unions conceded that the fruits of all productivity increases in the first half of the 1950s should be used tofinanceinvestment.16 They allowed wages to lag behind productivity in the 1950s 'so that industry could earn profits which would pay for expansion and modernization of the productive apparatus' (from a union publication, cited in Windmuller, 1969:350-1). A clearer statement of a wage moderation-for-investment trade-off is hard to imagine. The movement extended beyond this one case, however. In Belgium, wage restraint was part of a Social Pact (Projet d'Accord de Solidarity Sociale) initialled in 1944, in which unions and employers agreed to seek means of enhancing the competitiveness and productivity of enterprise.17 Workers agreed to moderate their wage demands and to refrain from striking except as a last resort, in return for a management commitment to boost productivity. In the mid-1950s unions and employers concluded a series of'Productivity Agreements' in which workers agreed to link their wage demands to productivity increases in return for a management pledge to stimulate the latter.18 In the 1960s, as part of the programmation sociale, labour agreed to renewed wage moderation in return for agreement by industry to a programme of capacity expansion and the provision by government of investment subsidies and guarantees. In Norway, unions agreed to restrain wages in return for promises from government and management to limit distributed profits to acceptable levels (Esping-Andersen, 1985). The Labour government mobilized the populace around the 'Work for Everybody' programme, under which unions agreed to restrain wages in return for a pledge that profits would be ploughed back into capital formation.19 Price, profit and dividend controls, maintained through 1953-4, provided a guarantee that capital would keep to its part of the bargain. The government used profit taxes and other income sources to underwrite the highest investment rates in Europe. Labour embraced the bargain because it was promised higher future incomes and full employment. Industry acquiesced because it was promised cheap credit and wage restraint (Esping-Andersen, 1985: 216-19). In Austria, interest groups came together in a 'social partnership' whose goal, from the end of World War II to the late 1960s, was 'the creation of a favourable climate for growth by preventing an outbreak of inflation from uncoordinated and mutually inconsistent income claims by the various economic interest groups (Flanagan et aL, 1983: 63). Wage growth was moderate, fuelling the economy's relatively high investment rates. In Germany in the 1960s, as part of a programme of concerted action, 'the unions were urged to restrain their wage demands, not to puncture the entrepreneurs' improved profit expectations. Unions were promised a later settlement, which would include these delayed wage increases and restore
Institutions and economic growth
47
"social symmetry"' (Bergmann and Muller-Jentsch, 1975: 259). In other countries, attempts to cut such a bargain proved a failure. In the UK, Ireland, France and Italy, wage pressure was intense, squeezing capital's share of the national income and limiting the availability of domestic investment finance. Through the 1950s, investment ratios and economic growth rates were disappointing by the standards of northern European countries, where cooperative solutions were reached, and even by the standards of countries like France and Italy in the 1960s. 3.2
Commitment mechanisms
The immediate problem for postwar Europe was how to mobilize the funds to finance capital formation. International lending remained depressed; while it was possible to tap direct foreign investment flows from the United States, access to foreign capital was limited by the legacy of interwar default and the persistence of wartime controls. For growth it was therefore necessary that wages be moderated so that profits could be generated domestically to finance investment. Hence, workers had first to restrain wages in order to make profits available to capitalists for investment later. But once the wage restraint had occurred, capitalists could renege on their commitment to invest the profits, paying them out as dividends instead, leaving workers no reason to exercise restraint. And capitalists had no incentive to keep their part of the bargain in the absence of reassurance that workers would continue to exercise restraint in the future, appropriating the entire surplus. Workers and management thus confronted a prisoner's dilemma which could be resolved only through the introduction of institutional restraints on opportunistic behaviour. These institutions restrained such behaviour by disseminating information, monitoring compliance, and offering bonds. 3.2.1 Monitoring compliance and disseminating information One set of institutions monitored compliance and disseminated evidence of non-cooperation; by reducing the likelihood that shirking would go undetected, this reduced the returns to doing so. Workers were allowed to monitor management decisions relating to investment. Unions and employers' associations were encouraged to exchange information on wage and investment decisions through governmentsanctioned peak associations. The representation of labour unions on advisory and administrative committees of industry and government was made compulsory. German workplace co-determination, giving labour input into firms' investment strategies, provides a clear illustration of the operation of this mechanism.20 The 1951 Co-Determination Law mandated the appointment to the executive boards of all corporations of significant size a representative of labour, with the same rights as other members. By the late 1950s some 100firmshad labour representatives on their boards (Dahrendorf, 1959: 264-6). Streeck (1984: 147) argues that granting labour institutionalized influence on the management of the enterprise was a way of reassuring them that wage restraint would actually translate into investment. Works councils (required in all workshops with five or more employees) played a similar information-disseminating role in small German firms not covered by the Co-Determination Law. Further examples abound. In Austria, German-style wage moderation and high
48
Barry Eichengreen
investment were secured by regular consultation between representatives of labour, management and government (Katzenstein, 1984). In Holland, representatives of government, firms and workers sat together on PBOs (Publiek Rechtelijke Bedrijfsorganisatie) to develop common employment and investment policies for Dutch firms.21 In Norway, the Social Democrats introduced a system of Bransjerad (planning councils) and Produktionsutvalg (production committees) to 'promote democratic participation in matters of investment, planning and industrial organization'.22 In the Netherlands, the government used leverage acquired through the administration of price controls to inspect the books of individual companies, mainly regarding prices and costs, but on other matters as well. In Belgium, works councils were installed in every company with more than 100 employees under the provisions of the 1948 law on the 'Organization of Industry', and union and industry representatives sat together for the first time on the boards of public and semi-public industrial and financial institutions. Even France, which otherwise failed to put the institutional prerequisites for growth in place, made some progress in this one respect. Labour-management plant committees (comites d'enterprise) were required by law for all enterprises employing fifty or more workers, starting in 1945. Two members were entitled to attend the meeting of the board of directors of limited liability companies. Though the comites have been dismissed as ineffectual in shaping managers' production and investment decisions, to do so is to misunderstand their role. Never intended to give workers veto power over investment decisions, they simply provided a conduit for information regarding investment behaviour. An Act of 1946 required management to inform the comite and receive its opinion before finalizing investment decisions (ILO, 1950: 206). Limited liability companies had to inform the committee of the profits made by the undertaking, permit it to audit its accounts, and allow it to examine the record of discussions of the board of directors. Information relevant to the investment bargain was thereby disseminated.23 3.2.2 Bonding A second set of institutions locked in the bargain by creating 'bonds' that would be lost in the event of reneging.24 A variety of devices were utilized to achieve these ends. In Austria, for example,firmswere guaranteed raw materials and semi-finished products at submarket prices from public enterprises in return for their willingness to adhere to the terms of the bargain. Private business was given representatives on the boards of the state holding company to lend credibility to this arrangement (Kurzer, 1993: 35-6). Government subsidies for firms similarly functioned as bonds that would be forfeited in the event that capitalists failed to plough profits back into investment. In Germany, for example, subsidies for the steel and the aircraft industry were seen as part and parcel of the social market economy that provided a social safety net for workers.25 The tax and transfer system could also be used. The Austrian government threatened to levy tax penalties onfirmsthat failed to keep their part of the bargain. That labour possessed a powerful voice in postwar coalition governments right-wing parties having been discredited during the final years of the First Republic and the Nazi occupation - rendered this threat credible. Similarly, in the
Institutions and economic growth
49
late 1940s the Swedish government introduced regulations limiting the payments of dividends by public companies and taxing away nearly half of supernormal profits. An investment reserve scheme, established before the war but greatly expanded in the 1950s, 'invited' corporations to place as much as 40 per cent of their annual pre-tax profits into a closed public account, to be released with government approval. Business received substantial tax savings on such deposits. A clearer example of bonding is hard to imagine. The institutional initiatives enumerated above were all designed to bond capital, since in an environment of scarce liquidity thefirstproblem was to get management credibly to commit to invest, so that labour had the incentive to restrain wages and make requisite profits available. But these bonding devices were likely to suffice only if supplemented by mechanisms also locking in future as well as current wage restraint; otherwise management, fearful that an aggressive labour movement would appropriate future profits, would be tempted to disregard its forgone bonds and renege on its commitment to invest. A variety of postwar institutional initiatives can be understood as devices to bond labour. In Belgium, the first postwar government pushed through a scheme for social security in return for labour's adherence to the 1944 Social Pact (Dancet, 1988: 98). The welfare state was greatly extended between 1964 and 1969 (to entail expanded provision of health insurance, child and unemployment benefits, and centralized pensions), coincident with labour and capital's agreement to a new and more ambitious cooperative agreement (the so-called programmation sociale). In Norway, the government in the mid-1950s offered legislation improving work conditions, including vacations and the length of the working week, in implicit compensation for wage restraint (Schwerin, 1980: 84). In Sweden, the government's role in supporting the Rehn model of wage concessions for firms that succeeded in raising productivity was to protect workers against redundancies by providing active manpower policies entailing retraining, education, job placement services, and public job creation.26 The Austrian government agreed to extend tax and social insurance concessions to labour in return for its agreement to moderate wages.27 Governments surmounted problems of intergenerational equity by establishing pension schemes. Wage moderation now which translated into higher incomes later might benefit future generations at the expense of present ones, a fact which might cause those currently working to hesitate to defer their gratification. Governments therefore indexed pensions not just to changes in the cost of living, but to increases in the living standards of the currently employed. In Germany, state pension schemes were revised in the 1950s to 'ensure that persons retiring on a pension of a given money value should have the right to an adjustment in line with the increase in the general prosperity of the country, which was confidently expected to result from the rising productivity of those at work' (Shonfield, 1965: 92). 3.2.3 Coordinating mechanisms A third set of institutions coordinated bargains acrossfirmsand sectors, internalizing the externalities from wage moderation in one sector for investment and hence future wage growth in others.28 Either bargaining was centralized in the hands of a trade union federation and a national employers' association, or the government took an active role in the concertation of bargains reached by individual unions and
50 Barry Eichengreen
employers. Under these 'corporatist arrangements', the major labour and management associations enjoyed quasi-public status, and negotiations in different industries were coordinated under state auspices (see Bruno and Sachs, 1985; Newell and Symons, 1987). Bonding was important here as well: wage compression or 'solidarity' enticed sectoral unions to accept the loss of autonomy implied by centralized bargaining, while promises to provide private companies with inputs produced by public enterprises at submarket prices elicited the cooperation of employers. Thus, the Rehn model in Sweden made wage compression an explicit part of the bargain. In Holland the compression of wage differentials was a goal of the Board of Mediators from the immediate aftermath of World War II until the end of the 1950s; a skilled/unskilled differential of 10 per cent for weekly wages and 20 per cent for hourly earnings was maintained virtually unchanged to 1959. In Belgium, the progranimation sociale of the 1960s entailed a narrowing of regional and intrasectoral wage differentials and of disparities between men and women (Windmuller, 1969: 340; Dancet, 1988: 103-4). It is important to highlight two respects in which this framework differs from the recent literature on labour markets and macroeconomic outcomes. First, the present framework departs from the 'hump-shaped' model (Calmfors and Driffill, 1988), in which it is argued that both highly centralized and highly decentralized labour markets produce efficient outcomes. The argument here is that, while both highly centralized and highly decentralized structures may have facilitated adjustment to changing supply and demand conditions, competitive decentralized wage setting was ill-suited for internalizing the externalities for investment that spilled across sectors and over time. To the extent that such externalities prevailed, coordinated wage bargaining was required for efficiency. In the presence of externalities, competitive, decentralized markets did not suffice. Second, the term 'corporatism' is used differently here than in the recent literatures in both macroeconomics (e.g. Bruno and Sachs, 1985; Crouch, 1985) and comparative politics (e.g. Schmitter, 1981).29 Here it refers to the concertation of negotiations between labour and management by government. Those who follow the above-mentioned authors use it instead to refer to countries where labour is centralized and where there exists low shop floor autonomy. Here I refer to these factors as 'centralization'.30 This term as well is used differently than in the recent literature, where in practice it refers to the centralization of labour representation.3 * Here it refers instead to centralization on both the labour and employers' sides. 4
The evolution of domestic institutions
Authors like Olson (1982) emphasize the cleansing role of World War II in dissolving the quilt of interest groups and institutional restraints that had grown up in previous years and stifled growth. Their removal, in this view, freed the market to operate efficiently. The perspective I develop here is different. I emphasize continuity between pre- and postwar institutional arrangements. Postwar arrangements, rather than being created from scratch to fill a temporary institutional vacuum, were lineal descendants of earlier institutional structures. As I argue in more detail below, coordinating institutions are resistant of change. While the
Institutions and economic growth
51
disruptions of war and reconstruction afford more scope for change than do normal peacetime periods, this does not alter the fact that the scope of such change is necessarily incremental, implying continuity with the past.32 4.1
An overview of the historical dynamics
As Panitch (1979) and Maier (1984) remind us, efforts to institutionalize labourmanagement cooperation under state aegis can be traced to the late-nineteenth century. Protectionist responses to the crisis of 1873 in many European countries constituted a part of a political bargain between nascent interest organizations, the alliance of iron and rye in countries like Germany and France affording the most prominent example. With the rise of unionization, labour was introduced into discussions between producers' organizations. World War I was a watershed: it created excess demands for labour and enhanced the bargaining power of unions, conscious of the special vulnerability of governments and societies to work stoppages. Unions and management were brought together by government in an effort to conclude economy-wide wage negotiations that might head off labour unrest. Public officials used the authority of ministries of munitions to encourage this process without resorting to legislation (see Maier, 1984: 43). Labour's leverage diminished with the termination of hostilities and the re-emergence of unemployment. In some countries, new initiatives- the Stinnes-Legien Agreement in Germany, the NIRA in the USA, Popular Front labour policies in France, the Saltsjobaden agreement in Sweden - ensured continuity. Such frameworks were concessions to the labour movement by governments seeking to head off more radical alternatives. Frequently they were negotiated with the political arm of the labour movement, which had acquired a parliamentary presence. Maier suggests that these arrangements are more accurately characterized as 'supervised pluralism' (separate, decentralized negotiations conducted under broad procedural guidelines set down by government) than as economy-wide concertation.33 Centralized negotiations and encompassing organizations - 'state corporatism' to use Schmitter's (1979) phrase - were established only in Mussolini's Italy and Hitler's Germany.34 To minimize work stoppages and mobilize resources for the pursuit of war, between 1940 and 1945 many countries centralized bargaining between labour and capital under the heavy hand of government. Where postwar reconstruction was particularly difficult, these arrangements were allowed to persist into the postwar years. By the 1950s, however, reaction had set in. Governments withdrew, allowing labour and capital, often still federated into peak associations, to concentrate on bilateral bargaining. Prominent exceptions notwithstanding, intervention was limited to granting recognition and a representational monopoly to the leading interest organizations, using moral suasion to facilitate their cooperation, and posting bonds that would be lost in the event that any party reneged on the terms of its long-term commitment. In the 1960s, as bilateral cooperation between capital and labour became increasingly problematic, for reasons discussed below, governments intervened directly with statutory sanctions and controls.35 Maier (1984) argues that corporatist structures were especially precocious in small, open economies whose vulnerability to external disturbances encouraged the development of cooperative response mechanisms. Where wartime damage had
52
Barry Eichengreen
been extensive, as in Norway, the case for government involvement in the economy was particularly compelling. Germany and Austria, as subjects of occupation, felt vulnerable for different reasons, but their response was similar. In such countries, sectoral unions were more willing to delegate authority to peak associations, while labour and management were willing to tolerate government guidance of negotiations (Esping-Andersen, 1985: 217). It might be added that the large-numbers problem that bedevils attempts to arrange cooperative solutions was less debilitating in smaller than larger European countries.36 4.2
Overall trajectories
Most countries experienced increases in centralization over the two postwar decades, as employers' associations were rebuilt and union membership expanded. The growing centralization of wage negotiations can be seen as a mechanism for preventing individual groups of workers from responding to additional investment by their particular employer, by escalating their wage demands in order to appropriate the surplus generated by that investment. That this danger had mounted with time reflected the decline of unemployment to extremely low levels by the mid-1960s and fading memories of high unemployment in the 1930s, which had worked to subdue labour militancy; hence the institutional response. In all countries but Portugal and the Netherlands, arrangements moved in more corporatist directions over time. In the cases of Norway and Austria this movement was slight, since the highly developed state of corporatism in the 1950s left relatively little scope for further articulation. In Portugal, the authoritarian Salazar regime applied a strong corporatist hand throughout the period, while the Dutch government, so heavily involved in bargaining at the beginning of the period, reduced its role over time. The growing involvement of government in the economy-wide concertation of sectoral negotiations can r5e seen as a way of supplementing efforts at centralization, with the goal of preventing individual groups of workers from appropriating any surplus generated by decisions to invest. A third direction taken by institutional developments in the 1960s was the expansion of the welfare state. In many countries, health, unemployment, retirement and training programmes were elaborated. Governments more aggressively pursued their commitment to policies of full employment. This can be seen as an attempt to provide new bonds for labour to contain the resurgence of labour militancy. 4.3
Recapitulation
The preceding sections have traced the evolution of institutions developed in Western Europe after World War II to coordinate, monitor, bond and enforce the compliance of economic interest groups with the terms of their bargain to moderate wage claims and boost investment. For reasons rooted in national histories and politics, European countries differed in the articulation and subsequent trajectory of such institutions. France, Italy, Ireland and the UK possessed neither the centralization nor the corporatization required to coordinate, monitor and enforce such bargains, though they moved in more corporatist directions over time. Danish and Belgian arrangements started out more centralized, though there too the initial, extent of
Institutions and economic growth
53
corporatism was modest; they moved more quickly in the corporatist direction subsequently. The Netherlands, another country characterized by an intermediate degree of centralization, was initially highly corporatist and became less so over time. The role of government in supporting corporatist arrangements increased significantly in the more centralized German and Swedish economies, and modestly in the already highly centralized and corporatized Norwegian and Austrian ones.
5
The structure of international institutions
5.7
The nature of the bargain
For deferring consumption to be worthwhile, investment had to be productive. To put it another way, for investment to stimulate growth, there had to be a market for the goods produced by domestic industries whose capacity was augmented and whose efficiency was enhanced. Otherwise the terms of the domestic bargain would be rendered unattractive; if investment failed significantly to boost productivity and future incomes, workers and managers would be unwilling to defer current consumption for future gains. International trade, and intra-European trade in particular, allowed countries to specialize in the production of goods in which they had a comparative advantage without regard to limits on the demand for those products existing at home. By enhancing efficiency, this improved the intertemporal terms of trade between current and future incomes, encouraging wage restraint and investment. But efforts to expand trade had to surmount a further set of coordination and commitment problems. Restructuring a national economy along export-oriented lines was costly. Sinking the costs of reallocating resources along lines of comparative advantage could be an expensive mistake if one's trading partners reneged on their commitment to openness. Having incurred the costs of shifting resources into the production of exportable goods, countries mightfindtheir access to export markets blockaded, rendering their investments uneconomical. Before reorienting policy in this direction, governments therefore had to be convinced that their partners' commitment to openness was credible. This problem of collective action, though relevant to all European countries, was particularly pressing in the German case. Other European countries were particularly sceptical of its commitment to openness, given memories of the Schachtian policies of the 1930s and the Second World War (Berger and Ritschl, 1995). Germany had been Europe's dominant supplier of capital goods and the single largest demander of raw materials and consumer goods produced by other European countries. If it could not be relied on to supply the capital goods needed for the expansion and modernization of industry in other European countries, and to purchase the consumer goods and other merchandise which would be produced with that capacity, sacrificing efficiency for self-sufficiency would be the sensible strategy. Institutions which rendered credible Germany's commitment to intra-European trade could thus go a long way towards reconstituting the traditional pattern of comparative advantage and towards curing the dollar shortage (the balance of payments deficits of European countries vis-d-vis the USA, due mainly to their excess demand for capital goods).
54
Barry Eichengreen
For Germany, recovery and growth required the consent of the occupying powers to the removal of controls on industrial production. Postwar policy-makers required in return that Germany be integrated into the European economy and that barriers to exit be erected to prevent that commitment from being reversed. German policy-makers required guarantees of their country's access to imported raw materials, industrial intermediates and foodstuffs in light of the prominence the Nazis had lent to Germany's dependence on foreign supplies. Institutional solutions to these commitment and coordination problems were provided by the European Payments Union and the European Coal and Steel Community. The ECSC provided monitoring and enforcement technologies guaranteeing national steel industries access to one another's coal and iron ore. The EPU entailed the establishment of institutions to monitor member country trade policies and the posting of bonds increasing the cost for any European country of reneging on its commitment. As an internationally coordinated initiative, it solved the 'after you Alphonse' problem discouraging unilateral steps towards convertibility. 5.2
Locking in intra-European trade
The effort to rebuild Europe's trade proved a remarkable success. The volume of Western European exports rose by nearly 9 per cent annum in the 1950s and 1960s. The exports of all Western European countries but the UK grew at rates in excess of 5 per cent, with those of West Germany and Italy expanding at a double-digit pace. The ability of European countries to exploit their comparative advantage heightened the efficiency of investment, contributing to the postwar growth miracle. In this development the European Payments Union played an important role. Participating countries were required to agree to a schedule of intra-European trade liberalization. A Code of Liberalization formalized their commitment. By February 1951, less than a year after the EPU went into effect, all existing trade measures were to be applied equally to imports from all member countries. Participants were required to reduce trade barriers by one-half initially, and then by 60 and 75 per cent. The share of quota-free intra-European trade was to rise to 90 per cent by the beginning of 1955. Countries failing to comply with this schedule, or employing policies to manipulate the terms or volume of trade in undesirable ways, could expect to be denied access to EPU credits. Operating the EPU required creating a set of institutions (the Organization for European Economic Cooperation (OEEC), which worked in tandem with the Bank for International Settlements) to monitor compliance and impose sanctions. Drawings on the system were embedded in a mechanism minimizing the likelihood that a country could use EPU credits to exploit its partners by remaining in persistent deficit. No conditions were attached to a country's drawings on its quota of 15 per cent of its intra-EPU trade. But additional credits could be obtained only if a country agreed to conditions set down by the EPU's Managing Board. Discussions were often initiated well before a country's quota was exhausted, and it was made clear that the provision of exceptional assistance was contingent on the country's early adoption of policies of adjustment. Officials of governments receiving exceptional credits were required to appear at meetings of the Board for questioning and to submit memoranda regarding their progress for its review.
Institutions and economic growth
55
Moreover, US Marshall Plan administrators supported the EPU by providing $350 million of working capital to finance its operation. (In fact, the Economic Cooperation Administration and the State Department had to overcome the opposition of the Treasury and the US Executive Director to the IMF, who feared that the EPU would allow Europe to discriminate against US exports.) The fact that Europe and the EPU depended on Marshall aid reduced the likelihood that a debtor would renege on its agreement with the Managing Board and fail to take corrective action to eliminate its deficit. The provisions of Marshall aid and Europe's financial dependence on the USA allowed the United States to convince a reluctant UK to participate in the system rather than proceeding toward convertibility unilaterally (see Dickhaus, 1993).37 Indeed, for those concerned to construct a commitment technology, the EPU was preferable to unilateral current account convertibility, the other basis on which postwar Europe's trade might have been rebuilt. Convertibility was not technically infeasible, but, as a unilateral policy, it was too easy to reverse (Eichengreen, 1993a). It lacked the multilateral surveillance and conditionality that rendered the EPU an effective institutional barrier to exit. 5.5
Industrial self-sufficiency and national security
The ECSC further enhanced the credibility of Germany's commitment to openness by ensuring the French steel industry access to the Ruhr coal that was indispensable to its survival, and by providing German steel producers with guaranteed access to the iron ore of Lorraine. Coal and steel were viewed, rightly or wrongly, as essential to national security and to the rehabilitation of Europe's industrial base. As Pollard (1981: 86) put it, 'it had been precisely these industries which had become a focus of international hostility and national armaments and war-mongering'. By 1950 the inevitability that Allied control of German heavy industry would soon be terminated could not be denied. The question was whether Germany would use its industrial capacity benignly and allow other European nations free access to its products, or whether the rest of Europe would have to build up its self-sufficiency. In response to these dangers, the Schuman Plan of May 1950 proposed to create a common market in coal and steel. Its goal was to abolish protection in trade in coal, iron and steel among the six member states. The discussions culminating in the ECSC banned price discrimination between domestic and foreign customers. A Joint High Authority was created to monitor compliance with the terms of the agreement. Its decisions, within circumscribed limits, were binding on the national governments and private citizens of the partner countries. To render the High Authority accountable to its constituency and thereby lend it legitimacy, a Consultative Committee of industrialists, trade unionists, consumers and dealers was created, and a body of delegates from the parliaments of the member countries (the Common Assembly) was established with the power to vote 'no confidence' in the High Authority and thereby force its members to resign. Finally, a Community Court was established to adjudicate disputes between member governments, individual enterprises and the High Authority. Thus, by joining the Community, member states committed themselves to renouncing certain unilateral initiatives. As Gillingham (1995) puts it, the ECSC
56
Barry Eichengreen
'was based on a new idea, supranationality. Membership required transference of sovereign powers to a new European authority.' It is hard to imagine a more effective barrier to exit.38
5.4 Extra-European institutions The EPU and the ECSC were just two of the institutional arrangements committing countries to openness and international trade. They were tailored to Europe's special economic and security needs, and they spoke to the historically specific fears aroused by the continent's experience in the period after World War I, when the commitment to openness had proved ephemeral. At the same time, they were buttressed by the global framework in which they were embedded, especially the General Agreement on Tariffs and Trade (GATT) and the Bretton Woods system. In contrast to the ad hoc tariff truce conferences of the 1920s, GATT was quickly recognized as an ongoing process. Even before completing one round of negotiations, signatories agreed to another. The repeated-game nature of their interaction mitigated against the temptation to engage in non-cooperative behaviour. The many GATT rounds, aimed at slashing tariffs, proved effective in dealing with the ever-present protectionist pressures' (Bhagwati, 1988: 41). Another important institutional innovation was to transfer tariff-setting authority from the US Congress to the Executive Branch, which was less susceptible to narrow constituency pressures.39 As epitomized by the Smoot-Hawley Tariff, US commercial policy had been dominated by Congressmen who proposed measures that afforded protection to industries in their districts without taking into account the negative externalities imposed on other districts. The 1934 Reciprocal Trade Agreements Act delegated authority to an executive with a national constituency more likely to take into account the external effects of district-specific trade policy measures. The executive was authorized to negotiate and implement pacts with other nations in which each agreed to cut tariffs on items of interest to the other by up to 50 per cent without further recourse to Congress. New centres of expertise like the State Department were thereby created which diluted the protectionist agenda with broader international economic and political considerations. Given US dominance of the postwar trading order and its leverage over multilateral trade negotiations, this lent impetus to the GATT process. Perhaps most importantly from the present point of view, GATT avoided conflicts between regional and global trade liberalization, conferring institutional blessings on discrimination in trade when pursued with the goal of establishing a regional customs union. Countries were allowed to offer one another intraregional trade preferences so long as in the course of doing so they did not raise barriers against imports from the rest of the world. This provision worked to the special advantage of Europe, where for historical reasons the scope for regional liberalization initiatives was greatest. The role in the early postwar years of the International Monetary Fund and the Bretton Woods system of pegged exchange rates should not be exaggerated. Efforts to have members establish par values and declare their currencies convertible produced little of more than symbolic value. Initially, foreign exchange rationing
Institutions and economic growth
57
remained widespread. France experimented with multiple exchange rates. Devaluations were undertaken in 1949 and on other occasions without the prior consultation written into the IMF Articles of Agreement. Starting in 1950, Europe's international monetary relations were shaped by the EPU, not the Bretton Woods institutions. The strength of the Bretton Woods Agreement was its ability to prove compatible with specially tailored arrangements like the EPU. After 1958 and the termination of the EPU, the Bretton Woods system came into its own. For nearly a decade and a half, the international monetary system operated without a major crisis (Bordo, 1993). The UK's 1967 devaluation and revaluations by Germany and the Netherlands perturbed but did not disrupt the system. Exchange rate stability encouraged the expansion of trade. Large-scale private foreign lending resumed. Inflation, though accelerating in the 1960s, remained moderate by the standards of prior and subsequent decades. The import of the Bretton Woods system,firstas a structure into which the EPU happily fitted and then as the framework for Europe's international monetary affairs, was that it provided a nominal anchor that stabilized price expectations. So long as the USA remained committed to pegging the dollar to gold at $35 an ounce, and European countries remained committed to pegging their currencies to the dollar at declared par values, clear limits were placed on inflationary tendencies. Persistent high inflation in the USA would drive the American balance of payments into deficit and exhaust the country's gold reserves. Persistent high inflation in Europe would drive a European country's balance of payments into deficit and threaten the viability of its par value. Initially, ample US gold reserves and the dollar's role as a reserve currency relaxed the first of these constraints, while the maintenance of capital controls placed limits on the operation of the second. But while such factors relaxed the constraint on policy, they did not remove it. Hence, governments' investments in the Bretton Woods system of pegged exchange rates lent credibility to their stated commitments to the maintenance of price stability. The stability of price expectations under Bretton Woods in turn enhanced the effectiveness of macroeconomic stabilization policy, averting serious recessions and preventing cyclical declines in investment expenditure.40 So long as the pegged exchange rates of the Bretton Woods system remained credible, workers agreeing to a sequence of wage bargains did not have to worry that their nominal value would be inflated away. Any acceleration of inflation could be regarded as temporary, rendering compensating wage increases less urgent. Thus, when governments used Keynesian demand stimulus to counter a recession, the pressure of demand was less likely to translate into wage inflation and more likely to encourage production. Demand management policy proved effective. The consequent absence of serious recessions sustained investment at high levels.41 Indeed, the main difference in investment behaviour between the 1950s and 1960s and the interwar period was not that postwar investment rates were higher during expansions, but that governments succeeded in preventing investment from collapsing in recessions. This encouraged firms to contemplate sequences of related investment projects which would yield high returns if not interrupted by recessions. And the high returns on investment improved the terms of the trade-off between current and future consumption for workers and capitalists contemplating policies of moderation.
58
5.5
Barry Eichengreen
Implications
Institutions like the EPU and the ECSC, buttressed by extra-European arrangements like the Bretton Woods system and GATT, solved commitment and coordination problems that had to be surmounted for intra-European transactions to be rebuilt and rapid growth to be sustained. Still, not all countries took equal advantage of the opportunities afforded by this new institutional basis for intra-European trade. The best counterexample is France, whose governments pursued policies that cloistered industry from international economic pressures until the 1960s. The OEECs Code of Liberalization anticipated removing three-quarters of quotas on private imports among the member countries and their dependencies by February 1951. While France temporarily reached this level in April of that year, the subsequent deterioration of its balance of payments led the government to reimpose quotas on all intra-European imports in February 1952. The OEECs 'minimum goal' of 75 per cent was only reached in February 1955 (Table 2.1), and even then a special 10-15 per cent 'provisional compensatory tax' was applied to most of the freed commodities to insulate French producers from import competition (Baum, 1958: 102-3).42 In manufacturing, imports accounted for only 8 per cent of domestic consumption as late as 1959. For half of all manufacturing industries thefigurewas less than 5 per cent. In thirteen of thirty-two industries, imports were essentially nil (Adams, 1989: 155). These low levels of import penetration reflected the relatively restrictive non-tariff barriers to trade maintained by France even during the EPU years. Only following the establishment of the European Common Market did import exposure rise, from 8 per cent of domestic consumption of manufactures in 1959 to fully 25 per cent in 1980, with most of the change occurring in the 1960s. The impact on domestic market conditions was profound: all French industries experiencing an increase in import penetration due to the establishment of the European Economic Community (EEC) saw their product prices fall in thefirstfive Common Market years. All industries with an increasing propensity to export saw their product prices rise (Adams, 1989: 172). A clearer example of restructuring along the lines of comparative advantage is hard to imagine. With this restructuring came a dramatic increase in the rate of French economic growth, plausibly reflecting higher returns on investment due to a greater tendency to invest along lines of comparative advantage, which in turn enhanced the incentive and hence the tendency to invest. 6
The decline of the postwar settlement
Eventually the Golden Age drew to a close. Output and productivity growth decelerated. Wages exploded, investment rates slumped, and the growth of trade no longer outstripped the growth of output to the same extent as in the 1950s and 1960s. The domestic and international settlements that had provided the basis for the postwar growth miracle ceased to function. A satisfactory explanation for the rapid growth of the Golden Age must also account for this subsequent deceleration. Afirstset of arguments is based on the idea that the postwar growth process contained the seeds of its own destruction. On the
Table 2.1. Intra-European trade: percentage liberalized at selected dates0 Country
30 June 30 June 30 June 30 June 30 June 30 June 30 June 30 June 30 June 30 June 30 June 30 June 1954 1952 1957 1950 1953 1955 1956 1959 1951 1958 1960 1961
Austria
53 56 55 53 58 47 — — 64 54 39 53 — 53 81 42 57
— 75 66 50 75 — — 41 75 76 51 83 — 75 85 63 90
— 75 75 68 — 77 — 41 75 100 75 100 — 75 88 63 46
36 87 92 76 — 90 (90) — 75 100 75 93 — 91 92 — 58
76 87 93 76 51 90 (90) 29 77 100 76 93 — 91 92 — 80
82 88 93 76 75 90 (95) 29 77 99 75 93 — 91 92 — 84
90 91 — 86 82 92 (95) 29 90 99 78 94 — 93 92 — 94
90 96 — 86 — 93 (95) 29 90 99 81 94 — 93 91 — 94
90 96 — 86 — 94 (95) 29 90 98 81 94 — 93 91 — 94
90 96 — 96 91 92 (76) 29 90 98 81 94 — 93 91 — 95
90 97 — 95 (97) 92 (76) 85 90 98 81 94 61 93 91 — 97
90 97 — 95 (99) 93 (72) 85 90 98 85 85 70 98 91 (52) 97
56
65
66
71
81
84
89
83
83
91
91
(94)
BLEU Netherlands Denmark France Germany (FR) Greece Iceland Ireland Italy Norway Portugal Spain Sweden Switzerland Turkey United Kingdom M e m b e r countries combined a
Percentage of private trade freed from quantitative restrictions. Base year is 1948, except for: Germany (FR) (1949), Spain (1950), Austria, Greece (1952), Turkey (1959), Iceland after 1960 (1958), and Benelux after 1957 (1953). In the case of Benelux, the percentages shown for 1959 and 1957 reflect the change in the base year rather than any effective change in liberalization. Source: Eleventh Annual Economic Review, Europe and the World Economy, Paris: OEEC, 1960, p. 185.
60
Barry Eichengreen
domestic side, wage moderation was supported not just by centralized bargaining and corporatist arrangements, but by relatively elastic supplies of underemployed labour and memories of high unemployment in the 1920s and 1930s.43 The passage of time allowed memories of unemployment a generation before to fade, taking with them their restraining influence. At first, ample supplies of labour flooding into Germany from its East and into Holland from Indonesia encouraged docility on the part of their unions. Underemployed labour in the agricultural sector had similar effects in countries like Italy and France. But a decade and a half of rapid growth transformed Germany and the other countries of Western Europe from high- to low-unemployment economies, encouraging the unions to flex their muscles. A problem with this explanation is that it underestimates the market power of unions in many European countries. As early as the 1950s, supplies of skilled workers were inelastic to industries. Accounts are unanimous in ascribing market power to unions, aside from a few countries, such as Italy and France prior to 1960. Indeed, had elastic labour supplies stripped workers of market power, there would have been no need for institutions to lock in wage moderation once the growth process got under way, since wages would have been determined by competitive labour market conditions; there would have been no danger of wage rises designed to appropriate future profits to discourage investment. Many readily observed domestic institutions would have been without a clear rationale. On the international side, the success of GATT in stripping away tariff barriers may have encouraged governments to cultivate less transparent and tractable non-tariff barriers to trade with which the GATT process was less able to cope. The growing number of GATT signatories created a large-numbers problem hindering negotiations. The scope for the United States to encourage other countries to liberalize by allowing them to discriminate against it diminished hand in hand with declining US industrial predominance and Europe's catching up. While this explanation has appeal when applied to the 1980s, it is anachronistic to attribute mounting tensions on the international side in the late 1960s to the breakdown of the GATT process. The most important voluntary export restraints came later. US economic leverage, while in decline, was still considerable. The large-numbers problem cited by commentators on the Uruguay Round was not yet pressing. Although there may still be a sense in which the very success of the postwar growth miracle contributed ultimately to its downfall, the framework developed above, with its emphasis on institutions, points to five additional factors which together helped to ring down the curtain on the Golden Age. It directs attention to one of those five factors as possessing special importance. 6.1
Capture
An obvious place to start is Olsonian capture (Olson, 1982). Olson's argument is that the passage of time enables special interest groups to gain leverage over the operation of institutions and to turn them to their selfish ends. The new-installed union leaders of the postwar period, who initially identified with the workforce as a whole, may have come to favour senior workers at the expense of their junior counterparts as they themselves gained seniority, creating the insider-outsider
Institutions and economic growth
61
problems that undermined the smooth operation of the labour market (Lindbeck and Snower, 1988). Governments initially able to use public programmes to bond unions and firms may have fallen prey to capture by the very agents whose behaviour they sought to influence, as the latter acquired knowledge of the workings of government bureaucracy. While this is an appealing story, the general formulation does not explain the timing of events - why the Golden Age of rapid growth drew to a close in the late 1960s or early 1970s - or their abruptness. Perhaps public institutions are especially susceptible to capture; if so, then the increasingly prominent role played by government in the operation of domestic institutions starting in the 1960s, as it sought to paper over cracks in the postwar settlement through the use of wage and price controls and direct intervention in labour-management negotiations, could have opened up avenues for capture where these had not existed before, thereby helping to explain the timing of events. Unfortunately, like other variants of the Olson thesis, this one does not easily lend itself to formal statistical tests. 6.2
The oil shocks
A second explanation for the end of the Golden Age emphasizes the oil shocks of the 1970s. A large literature addresses the question of whether these have the capacity to explain the productivity slowdown, generally reaching agnostic to negative conclusions.44 An institutional framework suggests a different perspective. By reducing even modestly the returns on investment (by raising the relative price of energy, or even just by making that price more variable and therefore requiring firms to diversify away the associated risks by investing in a wider variety of different kinds of equipment), the oil shock worsened the intertemporal terms of trade for workers offering wage concessions now in return for higher investment, productivity and incomes later. Thus, even a modest impact of the oil shock on the productivity of new investment could have tipped the balance for workers deciding whether to offer wage moderation now in return for higher incomes later, and thus had major implications for the level of investment and rate of growth. The problem with this explanation is in the timing. Thefirstoil shock came along in 1973, but the wage explosions that mark the breakdown of the postwar settlement were already evident at the end of the 1960s. The oil shocks might help to explain why the postwar settlement, once shattered, proved so hard to put back together, but they are less obviously a factor in explaining its downfall. 6.3
The breakdown of Bret ton Woods
A third potential explanation is the collapse of the Bretton Woods system of pegged but adjustable exchange rates. With the breakdown of Bretton Woods, not just inflation but the business cycle grew increasingly volatile.45 The elimination of par values removed a nominal anchor that had stabilized price expectations. With the commitment to par values removed, agents had no reason to regard an acceleration of inflation as temporary. When governments stimulated demand in the effort to offset a recession, this provoked compensating wage increases; aggregate demand policies therefore elicited inflation rather than stabilizing output. Governments
62 Barry Eichengreen
having lost the capacity to counteract incipient recessions, the business cycle and investment grew more volatile.46 The institutional perspective developed above points to channels through which this increased volatility could have depressed the rate of economic growth. Workers were willing to trade wage moderation now for higher incomes later, so long as they had confidence that sacrifices now would deliver investment, growth and higher incomes later. Recessions were an obvious threat to this confidence. Wage moderation might not succeed in stimulating investment if in the interim the level of activity, profits and the incentive to invest collapsed into a recession. As noted above, investment rates were higher in the 1950s and 1960s than they had been in the 1920s and 1930s not because those rates were significantly higher during expansion phases of the business cycle, but because serious recessions in which investment rates collapsed were successfully averted after World War II.47 So long as serious recessions were absent, the trade-off between wage moderation now and investment later remained attractive to European labour. Once the Bretton Woods nominal anchor was removed and Keynesian stabilization policy lost its power, allowing recessions to re-emerge, labour lost the incentive to keep its part of the bargain. The timing of events is suggestive: although the breakdown of the Bretton Woods system only took place in 1971-3, those events were widely anticipated, and the system was already showing signs of instability in the late 1960s.48 The limitations of this explanation are two. There is reverse causality: labour militancy developing for other reasons could have increased the responsiveness of wages to prices and the pressure of demand, implying that the change in labour market behaviour was a cause rather than a consequence of the breakdown of Bretton Woods. In addition, the view emphasizing the effectiveness of countercyclical stabilization under Bretton Woods may exaggerate the prevalence of such policies in Europe prior to the 1970s. In the 1950s Keynesian demand management policies were little used outside the UK. In Germany, to take one example, they were quite inconsistent with the prevailing ideology of ordoliberalism. In the 1960s they may have gained popularity in countries like France, but over wide portions of Europe they were still scarcely used. 6.4
Capital mobility
A fourth explanation for the end of the Golden Age is rising capital mobility. This is the factor whose particular importance is highlighted by the institutional framework of previous sections.49 Section 2 suggested that the postwar growth process had to surmount two problems of dynamic inconsistency. Thefirstwas that capitalists had to be precommitted to invest their profits in the future, so that workers would restrain their wage demands now and the liquidity would be made available to finance investment. The second was that workers had to be precommitted to moderate their wage demands later, so that capitalists would invest now in the expectation of reaping future profits. Institutions undertaking monitoring, informationdisseminating and bonding functions were developed with both of these problems in mind. It can be argued that the structure of the relevant institutions was predicated on limited international capital mobility. In a situation of limited capital mobility, it
Institutions and economic growth
63
was possible, for example, to threaten the imposition of taxes on distributed profits at rates that would not be feasible when capital could evade them by moving abroad. As capital mobility began to increase in the 1960s, the bonding mechanisms of the 1950s therefore lost much of their effectiveness. Where wage moderation might once have guaranteed investment, international capital mobility now gave management the option of investing abroad rather than at home. Domestic workers no longer assured that wage moderation would translate subsequently into higher productivity and labour incomes had less incentive to be moderate. The problem with this explanation, advanced by Thelan (1991) and others, is that there exist equally compelling arguments working in the other direction. Capital's possession of an exit option, by providing investors with a credible threat, should have worked to encourage the acquiescence of labour. The knowledge that labour militancy might cause capital to move abroad was precisely the kind of information that should have induced labour to exercise restraint in order to prevent capital from exercising its exit option. Indeed, if capital was mobile inward as well as outward, wage restraint could induce foreign capital to flow in, augmenting the plant and equipment with which labour could work and eventually rendering it better off. Thus, it is still possible to use the capital mobility argument to explain why dividend payout rates might have increased and investment rates slumped in the 1970s. Capital was able to indulge its preference for consuming a higher share of profits without provoking wage demands sufficient to prevent the accrual of those profits, labour's militancy having been restrained by capital's exit option. What this hypothesis cannot explain, given that capital mobility should have exercised a restraining influence on labour militancy, is the simultaneous explosion of wage demands. 6.5
The end of catch-up
In the introduction to this paper, catch-up - the scope for exceptionally rapid expansion as European economies recovered prewar levels of output and closed the productivity gap vis-a-vis the United States - was offered as an alternative to institutional explanations for postwar Europe's growth. A different view is that the scope that existed for rapid growth during the catch-up period reinforced the institutions erected to solve commitment and coordination problems, and that once the catch-up phase passed, those institutional arrangements lost much of their effectiveness. Recall that at the centre of the institutional argument is a time-inconsistency problem. Workers and capitalists are both best off if they can agree to defer current consumption in return for future gains, which take the form of additional investment that results in higher productivity and income for all concerned, but neither is willing to agree to defer without an assurance that the other will do the same. Assume now a decline in the rate of return on deferring current consumption because, for example, the return on investment and the underlying rate of productivity growth both fall. The incentive for the parties to the agreement to resist the temptation to renege on its terms is correspondingly reduced. Institutions which
64
Barry Eichengreen
were adequate to contain this temptation previously may no longer suffice, since the intertemporal terms of trade have changed. The end of the catch-up phase of postwar Europe's growth could have played precisely this role, by reducing the return on investment and weakening the incentive for capital and labour to adhere to their agreement to defer current consumption in return for now more meagre future gains. From this perspective, the scope for catch-up and the institutional arrangements developed to solve time-inconsistency problems following the war should be viewed as interacting positively with one another, rather than as competing explanations. The question then becomes why, if institutions were developed after World War II to solve the commitment and coordination problems that confronted the European economy at that time, it did not prove possible to reinforce them to meet these new circumstances. In fact, there was some institutional adaptation in the 1960s. Welfare state programmes to bond labour expanded greatly in the 1960s. Government acquired an increasingly prominent role in the operation of labour market institutions in many countries over the course of this decade. Growing use of statutory wage controls, concertation of sectoral negotiations, and direct intervention in wage bargaining are all characteristics of government policy in the 1960s that can be seen as a response to the need to deal with the second dynamic-inconsistency problem. But the increasingly prominent role of government was superimposed on existing institutions rather than superseding them. The underlying structures - craft versus trade unions, cohesive or fragmented trade associations, the degree of centralization of wage bargaining - changed only slowly. There are good reasons to expect socioeconomic institutions like those described in this paper to change only gradually. Institutional arrangements coordinate the actions of a large number of individual parties. It may not be in the interest of any one of those parties to modify existing arrangements unless the others can be made to go along, which poses formidable coordination problems in a decentralized setting. Even in a reasonably centralized market like Germany, where there are sixteen large sectoral unions and a matching number of sectoral employers' associations, it may be hard to get the social partners in any one sector independently to alter the timing and structure of their bargaining arrangements, if the cost is forgoing the coordination benefits provided by an established tradition of intersectoral pattern setting. As David puts it, 'Institutions generally turn out to be considerably less "plastic" than is technology.' (See David, 1993; Zysman, 1993.) In this light, the exceptional opportunity that is afforded by an extraordinary event like World War II is clearly evident. A wartime disruption which suspends the operation of normal peacetime institutions reduces the opportunity costs of coordinating a shift to new arrangements. That the most dramatic changes in domestic arrangements took place in countries like Norway, the Netherlands and Belgium, many of whose labour, management and government leaders were exiled, and who took the interlude as an opportunity to plan a wide-ranging package of institutional changes, is much more than a coincidence from this point of view. Yet what is remarkable given the extent of wartime dislocation is how durable early institutional arrangements remained. As section 3 shows, there was very considerable continuity between Europe's postwar institutions and the institutional developments
Institutions and economic growth
65
of the interwar years and even the pre-1913 period. The exceptional circumstances of World War II and its aftermath may have provided an opportunity to adapt Europe's institutions to the special needs of postwar economic growth, but this could be done only where the departure from prewar arrangements was not too great.50 It is not surprising that, when in the 1960s circumstances again changed but without an event like World War II to interrupt the coordination functions of existing institutions, innovation was slow to come. 7
Conclusion
European economic growth in the quarter of a century that ended in 1973 outstripped growth in any period of comparable length before or since. The elements of Europe's growth miracle - wage moderation, high investment and rapid export growth - were delivered by a tailor-made set of domestic and international arrangements - on the domestic side, the social market economy; on the external side, international agreements and supranational institutions - that solved problems of commitment and cooperation that would have otherwise hindered the resumption of growth. Why then did growth slow after 1971? These exists a wealth of explanations, including the capture of socioeconomic institutions by special interest groups, the breakdown of the Bretton Woods system, the oil shocks of the 1970s and the rise in international capital mobility. While it is not possible to rule out these explanations, each of them has serious limitations. I therefore single out the role played by another factor: the interaction of postwar 'catch-up' with the sustainability of cooperative arrangements. The faster was growth, the greater was the willingness for workers and capitalists to defer current compensation in return for future gains. The period of rapid catch-up following World War II was therefore ideal for sustaining cooperative behaviour. As the scope for catch-up was spent, the incentive to renege on cooperative agreements was heightened, and agreements to cooperate broke down. Wage pressure intensified, investment slumped and the rate of growth was further depressed. Thus, whereas the beginning of the introduction to this chapter presented catch-up and institutional arrangements as two alternative explanations for postwar growth, my conclusion is that it would be more accurate to see them as complementary factors that reinforced one another. NOTES Prepared for thefinalconference of the Commission of the European Communities' Programme on 'Comparative Experience of Economic Growth in Postwar Europe', All Souls College, Oxford, December 1993. This chapter is part of a project on postwar recovery and growth in Western Europe, which receivesfinancialsupport from the National Science Foundation and the Center for German and European Studies of the University of California. It is an elaboration and extension of a short paper presented to the European Economic Association's Annual Meetings in August 1993, and published in the European Economic Review (June 1994). I thank Brian A'Hearn, Lisa Ortiz and Pablo Vasquez for research assistance, and Lars
66
Barry Eichengreen
Calmfors, Paul David, Assar Lindbeck, Torsten Persson, Bob Powell, Gilles St Paul, Solomus Solomou, Rick van der Ploeg and Charles Wyplosz for helpful comments. 1 The unweighted average of the annualized growth rate of GDP per hour worked for eight European countries was 4.4. per cent in 1950-73 but only 2.4 per cent in 1922-37 and 2.3 percent in 1979-88. Calculated from Crafts (1992: table 1) and Boltho (1982: table 1.1). The contrast with recent years would be even more dramatic if the comparison were with 1973-88. 2 Note that the slopes of the regression lines in both of these figures differ significantly from zero at the 99 per cent level of confidence. Interestingly, when both effects are included in a single growth equation, the initial (1950) level of GDP per man-hour remains statistically significant, but not so the wartime and immediate postwar shortfall. 3 Crafts (1992) presents calculations of the growth bonus due to catch-up vis-a-vis the USA and spring-back to prewar levels for the same eight European countries, finding that purged of catch-up and spring-back, growth rates decelerated from 3.1 per cent in 1950-73 to 1.9 per cent in 1979-88. 4 The estimates of Maddison (1976) show the investment rate in Western Europe rising from 9.6 per cent in 1920-38 to 16.8 per cent in 1950-70. 5 See, for example, Crafts (1992: table 2). 6 As Shonfield (1965: 6) put it, The success of the modern capitalist society in reversing the pressures making for high consumption at the expense of investment is one of its outstanding achievements.' In the discussion that follows, I suggest that postwar policy-makers were especially concerned with the second, or supply-side, influence of wage restraint on investment. Moderate and slowly growing real wages can been seen as theflipside of the coin of the low and slowly rising real exchange rate that some authors (e.g. Boltho, 1982, 1993) emphasize as an element of Europe's postwar growth miracle. In simple open-economy macroeconomic models, in which domestic and foreign economies are completely specialized in the production of different goods that are imperfect substitutes for one another, there is a one-to-one correspondence between the real exchange rate and the relative real wage. 7 For an overview of the interwar situation in various European countries, see Maier (1975). A detailed and controversial analysis of the German case is Borchardt (1991). A few attempts to contrast the interwar and postwar situation are cited below. For example, Broad berry (1994) shows for the UK that wage pressure was more intense before than after World War II. 8 This emphasis on the postwar settlement and the institutions used to support it is not original with this chapter. Przeworski and Wallerstein (1982: 218) refer to such settlements as 'class compromise' and stress the need for capitalists to 'consent to institutions that would make it reasonably certain that wages would increase as a function of profits according to some rule'. Boyer (1988) uses the term 'Fordism' to refer to the cooperative structure of industrial relations and equitable division of productivity gains, in whose operation government played a prominent role, that existed in Europe after World War II. 9 Indeed, one might argue that the structure of the game can be traced back to Marx, if not earlier. 10 The notion that institutions can be used to create a credible commitment is prominent in the work of North and Weingast, among others. See North (1993) and North and Weingast (1989). 11 Otherwise a prisoner's dilemma situation may arise, in which any one sectoral
Institutions and economic growth 67
12 13
14 15 16 17
18 19
20 21 22 23 24 25 26 27
bargaining unit wishes to moderate its demands only if it expects others to do the same, but in the absence of an agreement to harmonize demands, no one unit has an incentive to be moderate. These possibilities are modelled by Hoel (1990). A more comprehensive treatment would analyse also the coming of the European Economic Community in the late 1950s, GATT and the Bretton Woods international monetary system. I leave these aside, given what is already an overlong paper. Starting in 1955, significant increases in real wages were granted; this can be seen as labour's reward for having kept its bargain in the immediate postwar years. Cited in Flanagan et al. (1983: 378-9). For details, see Abert (1969). In fact, wages jumped up following the armistice, but their growth was moderated subsequently. It was viewed as desirable to raise wages with the goal of stimulating demand, rather than lowering them to boost supply, because Belgian industry had survived the war largely intact, making capacity in the war's immediate aftermath much larger than in other European countries, which in turn had few resources with which to purchase Belgian exports. Cited in Beaupain (1983: 148). Starting in the 1960s, the peak associations of 'social partners' negotiated a series of 'social programming agreements' coordinating changes in wages and working conditions (Lorwin, 1954: 255). As Gruchy (1966:325) put it, 'Organized labour in Norway has accepted the goal of a high-investment, low-consumption economy with all that means for wage and other economic policies because organized labor approves the general economic objectives of the Norwegian Labor Party and the Norwegian Labor government. After the high-investment economy has been in operation for a number of years and the Norwegian economy has been converted from a relatively underdeveloped to a mature industrial economy, organized labor would then expect the government to shift back to the goal of a high-consumption, low-investment economy in which a high and expanding standard of living would be widely shared.' McCain (1989) provides a model of co-determination as a solution to a game between labour and management, where cooperation leads to higher investment. Ultimately, the need for political compromise scaled down the extent of the PBOs' power (Appels, 1986: 175). In the end, these failed to retain the full decision-making power envisaged by their architects (Esping-Andersen, 1985: 218). The committee has therefore the most complete information and is directly informed as to the management of the undertaking' (ILO, 1950: 211). On bonding, see Schelling (1960). For elaboration of this argument, see Appels (1986). The USA and its allies acceded to these policies, of course, because they supported American Cold War aims. The model was named after Gosta Rehn, the Swedish labour movement economist. For details on the model, see Esping-Andersen (1985: 229-31). The UK is another case where it might be argued that the extension of the welfare state can be seen as an attempt to provide this bonding function. In his wartime pamphlet, How to Pay for the War, Keynes had urged the government to adopt not only a compulsory savings scheme, but also a more generous national minimum, to be extended through the provision of family allowances. Though Keynes's scheme proved too ambitious for the Treasury to swallow, the
68
28
29
30
31 32 33 34 35
36
37 38 39
Barry Eichengreen extension of family allowances following the war can be seen as a quid pro quo for wage restraint. Exactly this point can be found in Elder's (1982) article on Sweden. Describing the situation from the mid-1950s through the 1960s, he writes, 'in general the nature of the collective-bargaining arrangements in Sweden made it simpler than in many other countries to reach well-founded investment decisions. Highly centralised and self-disciplined organisations on either side of the labour market helped ensure that commitments once entered into would not be reneged upon' (1982: 66). Similarly, authors differ in referring to structural corporatism (the organization of conflicting interests into a limited number of functionally differentiated organizations controlled or recognized by the state) and functional corporatism (the institutionalized participation of private interest organizations in public policy), and to sectoral and macroeconomic (economy-wide) corporatism. As will become apparent, my emphasis here is on structural macroeconomic corporatism. In line with this distinction, Gennard and Wright (1978) contrast countries like Sweden where a relatively high degree of centralization allowed intersectoral harmonization to proceed with only indirect guidance from government, with countries like France and the Netherlands, where the more decentralized nature of the government required a more forceful government hand. This point is argued at length by Thelan (1994), who points to some notable exceptions. As van Waarden (1993) notes, such a perspective implies that war, rather than destroying existing interest organizations, may actually heighten the concentration of influence in them. The Mond-Turner talks in the UK might be cited as an exception, but it is questionable whether they had significant effects of an enduring legacy. Other examples that might be cited include Vichy France, Portugal under Salazar and Spain under Franco. This generalization also neglects the importance of government arbitration of private negotiations in countries like Germany. Though such public support initially sustained the operation of corporatist institutions, it also widened the scope for capture, ultimately undermining the efficacy of corporatist arrangements. I develop this last point below. In the exceptional cases, such as the Netherlands, government intervention had been extensive from the start. Hence, there was little scope in the 1960s for supporting corporatist arrangements by broadening the role of government. It is no surprise that Dutch corporatism displayed widening cracks from the early 1960s. Thus, Ingham (1974) argues that centralization is easiest in 'relatively small-scale societies which export a large part of their manufactured goods', since these tend to have a low level of 'industrial differentiation' and a high level of industrial concentration (see p. 42 and passim). Roberts (1958: 119 and passim) makes this same argument for Holland. The UK objected to endowing the EPU with authority over the blocking of accumulated sterling balances and resisted the creation of a powerful Managing Board. The same terminology is used by Meade et al. (1962). The contribution of the ECSC to the genesis of the European Economic Community remains a debated issue. I do not take up the origins and role of the EEC in this paper. These arguments are most clearly elaborated in Pastor (1980) and Lohmann and O'Halloran (1993).
Institutions and economic growth 69 40 The argument that follows is drawn from Eichengreen (1993b). 41 For evidence, see Alogoskoufis and Smith (1991). 42 As Sheahan (1963: 233-4) put it, 'Practically every OEEC meeting on trade liberalization up to the end of the 1950s featured a good deal of special criticism of French import restrictions. It was not merely that tariffs on imports of manufactured goods averaged unusually high, or that the percentage of trade allowed to enter without quantitative controls was unusually low, but that the small areas of competition permitted were restricted to those few industries in which French firms were able to hold their own anyway.' 43 On the elastic labour supply argument, see Postan (1964), Kaldor (1966) and Kindleberger (1967). 44 See Griliches (1988) for a survey of the relevant literature. 45 The contrast is especially pronounced if one compares the period of Bretton Woods convertibility (1959-70) with the post-Bretton Woods years. See Bordo (1993). 46 Alogoskoufis and Smith (1991) provide evidence of differences in the time-series process generating inflation before and after Bretton Woods. Eichengreen (1993b) shows that the same shift in time-series process is evident in the behaviour of inflationary expectations, and develops the argument of this paragraph concerning the effectiveness of countercyclical stabilization policy. 47 Boltho (1982) notes how there was no year between 1945 and 1971 when the real GNP of Europe as a whole fell by as much as 1 per cent. 48 The persistence of inflationary impulses and inflationary expectations was already trending upwards prior to the breakdown of the system. 49 Kurzer (1993) also emphasizes the implications of capital mobility for the changing operation of Western European institutions, but the perspective she adopts is different from the one utilized here. Although explaining why capital mobility increased in the 1960s is beyond the scope of this of this paper, I would mention four factors: first, the inevitable decontrol of the European economies following the exceptional period of heavy government intervention during and immediately following World War II; second, the restoration of current account convertibility at the beginning of 1959, which increased the scope for capital mobility through leads and lags; third, technological change which rendered controls on capital movements increasingly difficult to enforce; and fourth, the rise of competingfinancialcentres, which created a free-rider problem in which national policy-makers had an incentive to cheat on international understandings to maintain controls in the hope of acquiring increasing amounts of financial business from abroad. 50 Thus, by way of counterexample, in the UK, a country which had never possessed strong industrial unions or cohesive employers' associations, the attempt to adapt domestic institutions to the new imperatives of the postwar period proved a failure. REFERENCES Abert, J.G. (1969) Economic Policy and Planning in the Netherlands, New Haven,
CT: Yale University Press. Abramovitz, Moses (1986) 'Catching up, forging ahead, and falling behind', Journal of Economic History, 46, pp. 385-406.
,
Adams, James (1989) Restructuring the French Economy, Washington, DC: The Brookings Institution.
70
Barry Eichengreen
Alogoskoufis, George and Ron Smith (1991) T h e Phillips curve, the persistence of inflation, and the Lucas critique: evidence from exchange rate regimes', American Economic Review, 81, pp. 1254-75. Appels, A. (1986) Political Economy and Enterprise Subsidies, Tilburg: Tilburg University Press. Barnett, Corelli (1986) The Audit of War, London: Macmillan. Baum, Warren C. (1958) The French Economy and the State, Princeton, NJ: Princeton University Press. Beaupain, Therese (1983) 'Belgium: collective bargaining and concertation system inhibited by economic crisis and government', in Solomon Barkin (ed.), Worker Militancy and its Consequences, New York: Praeger. Berger, Helge and Albrecht Ritschl (1995) 'Germany and the political economy of the Marshall Plan, 1947-1952', in Barry Eichengreen (ed.), Europe's Postwar Recovery, Cambridge: Cambridge University Press. Bergmann, Joachim and Walther Muller-Jentsch (1975) T h e Federal Republic of Germany: cooperative unionism and dual bargaining system challenged', in Solomon Barkin (ed.), Worker Militancy and its Consequences, 1965-15, New York: Praeger. Bhagwati, Jagdish (1988) Protectionism, Cambridge, MA: MIT Press. Boltho, Andrea (1982) 'Growth', in Andrea Boltho (ed.), The European Economy: Growth and Crisis, Oxford: Oxford University Press. (1996) 'Convergence, competitiveness and the exchange rate', in N.F.R. Crafts and G. Toniolo (eds.), Economic Growth in Europe since 1945, Cambridge: Cambridge University Press. Borchardt, ICnut (1991) Perspectives on Modern German History and Policy, Cambridge: Cambridge University Press. Bordo, Michael D. (1993) T h e Bretton Woods international monetary system: an overview', in Michael D. Bordo and Barry Eichengreen (eds.), A Retrospective on the Bretton Woods System, Chicago, IL: University of Chicago Press. Boyer, Robert (1988) 'Wage labour relations, growth and crisis: a hidden dialectic', in Robert Boyer (ed.), The Search for Labour Market Flexibility: The European Economies in Transition, Oxford: Clarendon Press. Broadberry, Stephen (1994) 'Employment and unemployment', in Donald McCloskey and Roderick Floud (eds.), The Economic History of Britain Since 1700,2nd edn, Cambridge: Cambridge University Press. Bruno, Michael and Jeffrey Sachs (1987) The Economics of Worldwide Stagflation, Cambridge, MA: Harvard University Press. Calmfors, Lars and John Driffill (1988) 'Bargaining structure, corporatism and macroeconomic performance', Economic Policy, 6, pp. 14-61. Crafts, N.F.R. (1992) 'Institutions and economic growth: recent British experience in an international context', Western European Politics, 15, pp. 16-38. Crouch, Colin (1985) 'The conditions for trade unions' wage restraint', in Leon Lindberg and Charles Maier (eds.), The Politics of Inflation and Economic Stagnation, Washington, DC: The Brookings Institution. Dahrendorf, Ralf (1959) Class and Class Conflict in Industrial Society, London: Routledge and Kegan Paul. Dancet, Geert (1988) 'From a workable social compromise to conflict: the case of Belgium', in Robert Boyer (ed.), The Search for Labour Market Flexibility: The European Economies in Transition, Oxford: Clarendon Press. David, Paul A. (1993) 'Why are institutions the "carriers of history"?' unpublished manuscript, Stanford University, CA. Dickhaus, Monika (1993) T h e European Payments Union', forthcoming in
Institutions and economic growth
71
Richard Griffiths (ed.), Explorations in OEEC History.
Eichengreen, Barry (1993a) 'A payments mechanism for the former Soviet Union: is the EPU a relevant precedent?', Economic Policy, 17, pp. 310-53. (1993b) 'Epilogue: three perspectives on the Bretton Woods system', in Michael Bordo and Barry Eichengreen (eds.), A Retrospective on the Bretton Woods System, Chicago, IL: University of Chicago Press. Elder, Neil (1982) 'Continuity and innovation in Sweden in the 1970s', in Andrew Cox (ed.), Politics, Policy and the European Recession, London: Macmillan.
Esping-Andersen, Gosta (1985) Politics Against Markets, Princeton, NJ: Princeton University Press. Flanagan, Robert, David Soskice and Lloyd Ulman (1983) Unions, Economic Stabilization and Incomes Policies, Washington, DC: The Brookings Institution. Gennard, John and Michael Wright (1978) 'Incomes policy', in Derek Torrington (ed.), Comparative Industrial Relations in Europe, Westport, CT: Greenwood
Press. Gillingham, John (1995) The European Coal and Steel Community: an object lesson?', in Barry Eichengreen (ed.), Europe's Postwar Recovery, Cambridge: Cambridge University Press. Griliches, Zvi (1988) 'Productivity puzzles and R & D: another nonexplanation', Journal of Economic Perspectives, 2, pp. 9-21.
Grout, P.A. (1984) 'Investment and wages in the absence of binding contracts: a Nash bargaining approach', Econometrica, 52, pp. 449-60. Gruchy, A.G. (1966) Comparative Economic Systems, Boston, MA: Hough ton Mifflin. Hoel, K. (1990) 'Local versus central wage bargaining with endogenous investments', Scandinavian Journal of Economics, 92, pp. 453-69.
Ingham, G.K. (1974) Strikes and Industrial Conflict, London: Macmillan. International Labour Office (1950) Labour-Management Co-operation in France,
Geneva: ILO. Kaldor, Nicholas (1966) Causes of the Slow Rate of Economic Growth of the United
Kingdom, Cambridge: Cambridge University Press. Katzenstein, Peter (1984) Corporatism and Change: Switzerland, Austria and the
Politics of Industry, Ithaca, NY: Cornell University Press. Kindleberger, Charles P. (1967) Europe's Postwar Growth, New York: Oxford University Press. Kurzer, Paulette (1993) Business and Banking: Political Change and Economic
Integration in Western Europe, Ithaca, NY: Cornell University Press. Lancaster, Kelvin (1973) 'The dynamic inefficiency of capitalism', Journal of Political Economy, 81, pp. 1092-110. Lindbeck, Assar and Denis Snower (1988) The Insider-Outsider Theory of Unemployment, Cambridge, MA: MIT Press. Lohman, Susanne and Sharyn O'Halloran (1993) 'Divided government and US trade policy', unpublished manuscript, Stanford University and Columbia University. Lorwin, V.R. (1954) The French Labor Movement, Cambridge, MA: Harvard University Press. Maddison, Angus (1976) 'Economic policy and performance in Europe, 1913-1970', in Carlo Cipolla (ed.), The Fontana Economic History of Europe, vol. 2, London:
Fontana. Maier, Charles S. (1975) Recasting Bourgeois Europe, Princeton, NJ: Princeton University Press. (1984) 'Preconditions for corporatism', in John Goldthorpe (ed.), Order and Conflict in Contemporary Capitalism, New York: Oxford University Press.
72
Barry Eichengreen
McCain, Roger A. (1989) 'Codetermination, collective bargaining, commitment, and sequential games', in Hans Nutzinger and Jurgen Backhaus (eds.), Codetermination, Berlin: Springer-Verlag. Meade, J.E., H.H. Liesner and SJ. Wells (1962) Case Studies in European Economic Integration, London: Oxford University Press. Newell, Andrew, T. and James Symons (1987) 'Corporatism, laissez-faire and the rise in unemployment', European Economic Review, 31, pp. 567-602. North, Douglass C. (1993) Institutions and credible commitment', Journal of Institutional and Theoretical Economics, 149, pp. 11-23. North, Douglass C. and Barry Weingast (1989) 'Constitutions and commitment: the evolution of institutions governing public choice in seventeenth century England', Journal of Economic History, 49, pp. 803-32. Olson, Mancur (1982) The Rise and Decline of Nations, New Haven, CT: Yale University Press. Panitch, Leo (1979) T h e development of corporatism in liberal democracies', in Philippe C. Schmitter and Gerhard Lehmbruch (eds.), Trends Toward Corporatist Intermediation, Beverly Hills, CA: Sage. Pastor, Robert (1980) Congress and the Politics of US Foreign Economic Policy, Berkeley, CA: University of California Press. Pollard, Sidney (1981) The Integration of the European Economy Since 1815, London: Allen and Unwin. Postan, M.M. (1964) An Economic History of Western Europe 1945-1964, London: Methuen. Przeworski, Adam and Michael Wallerstein (1982) 'The structure of class conflict in democratic societies', American Political Science Review, 76, pp. 215-38. Roberts, B.C. (1958) National Wages Policy in War and Peace, London: George Allen and Unwin. Schelling, Thomas (1960) The Strategy of Conflict, Cambridge, MA: Harvard University Press. Schmitter, Philippe C. (1981) Interest intermediation and regime governability in contemporary Western Europe and North America', in Suzanne Berger (ed.), Organizing Interests in Western Europe, Cambridge: Cambridge University Press. Schmitter, Philippe C. and Gerhard Lehmbruch (1979) Trends Toward Corporatist Intermediation, Beverly Hills, CA: Sage. Sheahan, John (1963) Promotion and Control of Industry in Postwar France, Cambridge, MA: Harvard University Press. Shonfield, Andrew (1965) Modern Capitalism, Oxford: Oxford University Press. Streeck, Wolfgang (1984) Industrial Relations in West Germany, London: Heinemann. Thelan, Kathleen (1991) Union of Parts: Labor Politics in Postwar Germany, Ithaca, NY: Cornell University Press. (1994) 'Beyond corporatism', forthcoming in Comparative Politics. van der Ploeg, Frederick (1987) 'Trade unions, investment and employment: a non-cooperative approach', European Economic Review, 31, pp. 1465-92. van Waarden, Frans (1993) 'Introduction: crisis, corporatism and continuity', in Wyn Grant, Jan Nekkers and Frans van Waarden (1993) Organising Business for War, New York: Berg. Windmuller, J.P. (1969) Labor Relations in the Netherlands, Ithaca, NY: Cornell University Press.
The varieties of Eurosclerosis: the rise and decline of nations since 1982 MANCUR OLSON
1
Introduction
Valuable as they are, conventional economic models have not succeeded in explaining the great differences in economic performance in different countries or historical periods. The 'old' growth theory is called into question by the absence of the convergence in per-capita income levels across the globe that it leads one to expect. The 'new' growth theory readily accommodates continuing differences in per-capita incomes across countries, but does not yet provide much insight into why the particular countries that became rich were the ones that grew. Neither does it explain why a few poor countries have led the world in economic growth at the same time that the poor countries as a whole have not been catching up. In a book, The Rise and Decline of Nations,1 some article-length publications that foreshadowed it, and some subsequent papers and publications that focused on Eastern Europe and on the Third World,2 I presented a theory that combines the insights of familiar neoclassical economic models with a model of collective choice. The model of collective choice enables the theory to comprehend certain aspects of political and organizational life. The theory succeeds in explaining the most dramatic and puzzling variations in economic performance across countries. The present chapter will differ from the aforementioned book and articles in two ways. First, this chapter will attempt to show how the same fundamental forces that influence economic performance everywhere show up in a very different forms in countries with different initial conditions. Even when one considers only the related and similar societies of Western Europe, the institutional and historical differences across countries not only have a significant independent effect on outcomes, but they also make the same fundamental forces take on a totally different appearance. Just as a change in temperature can turn water into ice or steam, so international differences in institutions can make a single, general force operating in different countries appear to be a variety of unrelated phenomena. This chapter will show that some apparently unrelated developments in Europe are, in fact, manifestations of the same process. 73
74
Mancur Olson
Second, the analysis will focus only on economic growth since Rise and Decline was written. Any theory should be able to explain data beyond the data from which it was derived. Obviously, I could not have known, when writing the book, about developments that occurred after the book was published. Thus the theory could not possibly have been adjusted to take account of the developments that are the focus of the present study. Since my discussion will not be comprehensible unless the reader has some knowledge of the theory developed in prior works, section 2 outlines some of the ideas from prior publications that will be used and tested in this essay. Readers who know both Rise and Decline and my subsequent papers on the Communist and post-Communist societies can skip this section. 2
A recapitulation of the theory
The theory begins with collective action - concerted efforts to lobby the government or to combine in the marketplace to influence prices and wages. Such action occurs through professional associations, labour unions, farm organizations, trade associations, and oligopolistic collusions of firms in concentrated industries. The benefits of the governmental favours and the monopolistic or monopsonistic prices or wages obtained by organized action go to everyfirmor individual in some group or category, a tariff or tax loophole favours every firm in some industry or group, and cartelization raises the price or wage for every seller of a good. It follows that any sacrifice an individual makes to support a lobby or cartel for his group will benefit others as much as himself. A group with a common interest will be able to overcome this collective good problem only if it has the advantage of small numbers or is blessed with access to 'selective incentives'. The advantage of small numbers is clearest in a concentrated industry containing only a few largefirms.If there are twofirmsof the same size in an industry, each firm will obtain half of the benefit of any action in the interest of the group. Even though eachfirmmust bear the whole cost of whatever it does for this group of twofirms,its large share of the benefit will often be sufficient to give it an incentive for some unilateral action in the interest of the group. If there are, by contrast, a million individuals in a group with a common interest, a representative individual in the group will receive only a millionth of the benefits of any action he or she takes in the interest of the group. Each individual will still have to bear the whole cost of whatever he or she does in the interest of the group. As these examples make obvious, the incentive to act in the interest of the group must be less in large than in small groups, and in really large groups, the incentive for an individual to engage in spontaneous collective action is vanishingly small. When a large number of people share a common interest, they will be able to act collectively to serve this interest only if they have 'selective incentives'. A selective incentive is a reward or punishment that, unlike the benefits of the collective good itself, selectively apples to individuals according as they do or do not contribute to the provision of the collective good, and therefore gives the individuals an incentive to act in the group interest. Probably the best-known selective incentives are the closed shop, the union shop and the coercive picket line - those who do not share the costs of the collective action are threatened with loss of a job or are subject to social
The varieties of Eurosclerosis
75
or physical intimidation. Though the selective incentives used by other kinds of large organization for collective action are usually less conspicuous, they are no less important.3 Since collective action is difficult and problematical, it usually takes quite some time before a group can engage in collective action, even when it has small numbers or the opportunity to devise selective incentives. The bargaining that is required for those in small groups to organize or collude with full effectiveness often takes a while. Organizing large groups is incomparably more difficult and time consuming. It is only in the fullness of time that many groups will have had the able leadership and the favourable circumstances needed to organize for collective action. 2.1
The incentives facing organizations for collective action
What difference does it make for the prosperity of a society how many groups, and what kinds of group, are organized for collective action? The answer is evident when we look at the incentives that organizations for collective action face. The constituents of any organization that represents only a narrow segment of the society will virtually always be better off if the organization shifts the distribution of income in the society in their favour. An organization for collective action can shift the distribution of income in the society in its favour through lobbying for special-interest legislation - for subsidies, tariffs, tax loopholes or regulations that limit entry and competition. It can shift the distribution of income in its favour by selling less and charging more for it - that is, by collusion or cartelization. In general, special-interest legislation and monopolistic combination make an economy less productive than it would have been, and the constituents of an organization for collective action will share the losses. But an organization for collective action that represents only a narrow segment of the society will bear only a minuscule share of these losses. A special-interest organization whose constituents earn 1 per cent of the national income will, on average, bear only 1 per cent of the loss in national output that occurs because of the inefficiency its activities bring about. But the special interest obtains the whole of the amount redistributed to it. Thus it pays our hypothetical special-interest group to seek to redistribute income to its own members, even if this redistribution reduces the national income by up to 100 times the amount redistributed. Therefore, organizations that represent only a minute percentage of an economy's income-earning capacity are really 'distributional coalitions' - coalitions whose purpose is to redistribute more of society's income to themselves. If an organization for collective action encompasses a large part of the income-earning capacity of a country, its incentives are very different. This is immediately evident if one considers an organization that represents, say, half of the income-earning capacity of a society; its clients will on average obtain one-half of any increase in the national income that it brings about, and bear half of the social loss that results from redistributions to itself. There have been such encompassing interest organizations in Austria, Norway and Sweden, and to a lesser extent in Germany and Japan. The concept of the encompassing interest applies to many different types of situations, offices and individuals. A dictator has an encompassing interest in any
76
Mancur Olson
domain which he securely controls: his tax receipts, which will increase with the productivity of his country, give him this encompassing interest.4 Though a large organization cannot usually optimize as effectively as a single individual, a large and well-disciplined political party, such as the Conservative Party in Great Britain or the Social Democratic Party in Sweden, also has an encompassing interest. So does a president in strong-presidency countries like France and the United States; a president normally needs majority support to be re-elected, and a majority is an encompassing constituency. Some of the conclusions of the present argument hinge on this concept of the encompassing interest, and much of the latter part of this essay is devoted to an analysis of the dynamics of encompassing special-interest organizations. 2.2
Testable implications of the theory
The ideas that have just been evoked can explain the most striking anomalies in modern economic growth. We have seen that organizations for collective action with narrow constituencies have uniquely perverse incentives, and that it takes a long time before a society accumulates many organizations for collective action. Revolutionary upheavals, totalitarian repressions and foreign occupations destroy organizations for collective action. In any brief interval of stability, few if any groups can overcome the difficulties of collective action. By contrast, in a long-stable society, many groups will have overcome these obstacles. Once these organizations have worked out the selective incentives or agreements needed for collective action, they rarely disappear unless they are violently repressed. Thus only long-stable societies are thick with organizations for collective action. Distributional coalitions are, if my argument is correct, uniquely harmful to economic efficiency and dynamism. It follows that societies that set up a good legal order, after a catastrophe has destroyed organizations for collective action, will, for a time, grow extraordinarily rapidly. Similarly, long-stable societies ought to grow much less rapidly than societies that are in other respects comparable. Thus we can test the theory by asking whether it fits the facts about economic performance in different countries. The society that has had the longest period of stability and immunity from invasion and institutional destruction is Great Britain. As the theory predicts, Great Britain for most of the postwar period has also had the 'British disease' - the poorest economic performance of the major developed democracies. The economic miracles of Germany and Japan after World War II are also consistent with the argument. In Italy, the institutional destruction in World War II, while considerable, was less complete than in Germany and Japan. The economic miracle in Italy, though there definitely was one, was correspondingly somewhat shorter and smaller than those in Germany and Japan, and this again is in accord with the theory. With appropriate elaboration,5 the aforementioned theory also explains the general pattern of regional growth in the United States since World War II. Consider also the most remarkable examples of economic growth in previous centuries: the growth of Germany after the Zollverein or customs union was established in 1834 and after German unification was completed in 1871; the growth of Japan after the Meiji Restoration of 1867-8; the growth of the United States in the
The varieties of Eurosclerosis
77
nineteenth century; the growth of Holland during its Golden Age in the seventeenth century; the growth of Britain during the Industrial Revolution from about 1760 to about 1840; and the commercial revolutions in England and France in the sixteenth century. All of these cases illustrate what I called 'jurisdictional integration': in each case, a wider market was created within which there was at least internal free trade, and at the same time a new jurisdiction or government was established that could be influenced only by lobbies that were of a larger scale than most of those that had influenced the parochial jurisdictions that existed before. After the creation of a much larger jurisdiction and the wider market, there was in every case rapid economic growth. Jurisdictional integration undercut the guilds and other special-interest groups of the day. A detailed examination reveals that not only the speed of development after the jurisdictional integration, but also its organizational features and regional patterns, are consistent with the theory, as is much of the subsequent literature.6 2.3
Generalizing the theory for Communism and the transition
Though Rise and Decline said nothing about the Soviet-type countries, the theory in it was soon generalized to cover these societies and then later to deal with the transition from Communism.7 Under Soviet-type dictatorships, there was, of course, no freedom of organization, so there were no formal lobbying organizations, independent labour unions or other explicitly cartelistic organizations. Thus it might seem that under Communism there could not have been any collective action that impaired economic performance. I claim that, in fact, the dictator in a Soviet-type system is utterly dependent for information on bureaucratic competition among the subordinates in individual industries and enterprises, and that the subordinates in each industry and enterprise can, if they covertly collude, distort the informationflowgoing tp the centre in ways that enable them to obtain surplus resources that serve their own interests. Thus covert collective action ultimately creates innumerable nomenklatura collusions and, in fullness of time, this devolution reaches the point where large enterprises and industry associations are as much insider lobbies and monopolies with vested interests as they are instruments of production to serve the Communist regime. Each collusion, enterprise and industry has so small a stake in the productivity of the society as a whole that each tends to ignore its impact on the society. In this the individual collusions and enterprises are very different from the general secretary of a Communist country, who has an encompassing interest in the productivity of his domain and therefore a motive to make it as productive as possible. Thus, in the absence of organizationally destructive events like China's cultural revolution, the theory predicts that the economic performance of stable Soviet-type societies deteriorates over time. Since the enterprises and industries are not destroyed by the transition to democracy, but are, on the contrary, given a new freedom to lobby for their sectional interest, this problem is magnified during the transition. Since the focus here is on Western Europe, I will not discuss the other testable predictions of the theory for Soviet-type and transitional societies, and will return to developed democracies.
78
2.4
Mancur Olson
Better economic understanding as the antidote
The theory argues that it is mainly narrow as opposed to encompassing organizations that repress economic growth. Each narrow special-interest organization represents only a tiny minority of the population. I can therefore, in any democracy, easily be outvoted. A special interest can get its way only because most of the society does not notice or understand what is happening. Thus Rise and Decline emphasized the quality of economic thought and the extent of economic literacy. It argued that, if societies came to understand and believe the argument in Rise and Decline, the predictions in the book would be refuted. Since both the level of economic understanding and the density of organizations for collective action are important determinants of economic performance, each needs to be kept in mind in any attempt to trace the influence of the other. A country may grow faster or slower than would be expected from the density of organization for collective action because its understanding of the problem - and of what is required for an efficient and dynamic economy - is better or worse than in comparable countries. It is even conceivable that societies with more serious cases or less opaque forms - of institutional sclerosis apprehend the problem more quickly than other societies.8 This means that, in analysing economic growth since 1982, we must keep both the character of collective action and the prevailing economic ideas in a country in mind. 3
Economic growth since Rise and Decline
Some of the best evidence in support of the theory that has been outlined above comes from the Communist and formerly Communist countries. In the early postwar period, the Communist countries grew relatively rapidly. But as time went on they suffered a gradual and continuing deterioration in their growth rates, in spite of their continued opportunities for fast catch-up growth. It was only after a long period of relative decline that Communism collapsed. The sharply contrasting consequences of the defeat of Fascist and Communist regimes fits the theory especially nicely. Whereas the defeat of Fascism destroyed most organizations for collective action, the defeat of Communism gave the lobbies of large enterprises and industry associations the opportunity to use their political muscle openly. The defeat of Fascism was followed by economic miracles in all of the formerly Axis countries, but the defeat of Communism has so far often been followed by even poorer performance than under Communism. The only Communist or once-Communist country that has enjoyed an economic miracle is the one that suffered the cultural revolution (and after this revolution China was ruled by a pragmatic dictator with an encompassing interest in a productive domain). This is indeed the exception that proves the rule. But the focus here is on the West, so I turn to the less dramatic changes that have occurred in the developed democracies. A most interesting pattern of gradual change through time has been evident in the countries whose Fascist regimes were defeated in World War II. The logic of the argument in Rise and Decline implied that, as time went on, the same accumulation of organizations for collective action that had troubled the long-stable English-speaking countries would, if these countries remained stable, also continue to occur in
The varieties of Eurosclerosis
79
Germany and Japan. This prediction was made explicitly: The theory here predicts that with continued stability the Germans and Japanese will accumulate more distributional coalitions, which will have an adverse influence on their growth rates/9 The same argument was made separately with respect to Italy. Germany and Italy were already, when Rise and Decline was being written, experiencing some reduction in relative growth rates, but at that time this relative decline had not attracted much attention outside of those countries. The almost universal forecast at that time was that the Japanese economic miracle would continue. By now it is beyond dispute that the Japanese as well as the German and Italian economic miracles are over. In the 1950s and 1960s, real per-capita incomes of Germany and Japan grew more then three times as fast as those of the UK, the USA, Ireland, Australia and New Zealand. Since 1980 the average rate of growth of per-capita income in the English-speaking OECD countries has been as great as, and often even greater than, that of Germany and Japan. Rise and Decline also argued that there was some institutional destruction in the continental countries that were under Axis occupation during World War II, and that many distributional coalitions in France and in some of the other previously protectionist countries were rendered ineffective by jurisdictional integration as the initial six countries created the Common Market. The jurisdictional integration due to the formation of the Common Market gave the six original members some growth advantage, for a period of time, over the long-stable English-speaking OECD countries. Again, the theory in Rise and Decline predicted that, with continued stability in the Common Market countries, the growth-enhancing effects of the jurisdictional effects would gradually wear off as new patterns of collective action accumulated. Again, this is what seems to have happened. In the 1960s per-capita incomes in France, Italy, Belgium and the Netherlands all grew at more than 4 per cent per year, whereas the English-speaking OECD countries grew at far slower rates. Since 1980, by contrast, these countries have grown no faster - and sometimes slower than the English-speaking OECD countries. Countries with the most encompassing interest organizations have also slowed down over time. In the 1950s and 1960s Sweden, in spite of its already relatively high level of per-capita income, grew faster than the English-speaking countries. The growth rates of Austria, Norway and (as we already know) West Germany were also well ahead of those of the English-speaking countries. In the 1980s (and especially the 1990s) Swedish economic growth has fallen far behind that of the English-speaking countries, and that is also true of most of the other European countries that had enjoyed relatively encompassing interest group structures. Whether we look at the growth rates of the Communist countries, or at Germany, Japan and Italy, or at the continental countries that were occupied by the Nazis in World War II, or at countries like Sweden that have had encompassing interest-group structures, we see that the changes are not only in the direction predicted by the theory, but also have the gradual character that would be expected from a sclerotic institutional accumulation. Nonetheless, even though the aggregate evidence on national growth rates that has emerged since 1982 is certainly supportive of the theory, this aggregate evidence by itself is by no means compelling. As we know, economic growth is influenced by
80
Mancur Olson
many different factors. It is obviously possible that other factors that influence growth rates could have explained the observed pattern, and that the conformity with the theory in Rise and Decline is spurious. There is also another and more intriguing possibility. We recall that, if the theory in Rise and Decline were generally understood and accepted, the predictions of theory would necessarily be refuted, and, more generally, that the intellectual climate influences choices about economic policy and is therefore extremely important. It is possible that the countries that suffered the most (or suffered in the least opaque ways) from special-interest cartelization and lobbying would be more susceptible to analyses of the problem, and come to have a bit more of an apprehension of it, than societies that had suffered less (or in ways that were harder to understand). At least if the issue is explicitly discussed, the degree of understanding of the damage done by distributional coalitions could be, at least to some extent, endogenous. Rise and Decline pointed out that two countries that had suffered extraordinarily severe institutional sclerosis, not only because of their long-standing stability, but also because of their small size and then-exceptional propensities to industrial protectionism (the opposite of jurisdictional integration) were Australia and (especially) New Zealand. Interestingly, Labour governments in both New Zealand and Australia have undertaken widespread market-opening and deregulatory reforms that seem greatly to have weakened special-interest organizations on the business side. As might be expected, these Labour parties, dependent in part on support from labour unions, have not taken the same strong stand against cartelization in the labour force. Similarly, the deregulation of the thoroughly cartelized and lobby-intensive industries such as airlines, trucking, telephone and railroads in the United States, mainly during President Carter's Democratic administration, owed something to the conspicuous economic irrationalities that special-interest pressures had brought to these industries. The beginnings, during the Callaghan Labour government in the UK, of a reliance on limiting the money supply rather than only on incomes policies to restrict inflation, could be interpreted in a similar way. Despair, even in the British Labour Party, about the UK's anything-but-encompassing trade unions also led a substantial part of the Labour leadership to leave that party and create a new political party. This step - which was all the more extraordinary because the British electoral system makes the defeat of any third party almost inevitable - could also be analysed as a response to the severity of institutional sclerosis in the UK: as we know, the phenomenon there has been prolonged and serious enough to come to be called the 'British disease' and this surely generated some demand for cures, even if they should be painful. Thus, in the same way, Thatcherism can also be taken to be endogenous - even as the mirror image of the Labour or left-inspired reforms in New Zealand and Australia. Just as the reforms inaugurated by the Labour governments of the South Pacific focused disproportionately on reversing the losses arising from cartelization and lobbying by business, naturally the Thatcher government focused on limiting the harm done by the special-interest groups that were linked with the left: the labour unions, which are also the most visible part of the special-interest iceberg. Intriguing as the foregoing possibility of endogenous response to institutional
The varieties of Eurosclerosis
81
sclerosis is, I must emphasize that I have not been able to undertake the multicountry study of political, economic and intellectual history that would be needed to know how much truth it contains. The more important point for now is that no monocausal explanation of complex historical developments can be correct, including any such explanation drawn from the theory that I have offered. A series of factors that are not accounted for by the theory in Rise and Decline could have made the observed pattern of growth rates since 1982 spuriously consistent with the theory. Some of these other factors could even have operated in such a way as to generate the gradual character of the observed changes. There is some additional evidence in the economic literature on continental and Nordic Europe in the 1980s and 1990s. In a series of papers and lectures starting in 1976, and in Rise and Decline, I used the phrases 'institutional sclerosis' and 'institutional arthritis' to label the process that my theory predicted.10 Thus there is perhaps a little support for my argument in the emergence, in continental Europe in the 1980s, of the term 'Eurosclerosis', which I have borrowed for the title of the present paper. The German economist Herbert Giersch appears to have coined the term, which also has currency in the popular press, and Asaar Lindbeck gave sustained attention to the increasing sclerosis in continental and Nordic Europe in 1982.11 In Sweden, Stahl and Wikstrom have used the more targeted label of 'Suedosclerosis' in the title of a recent book, and this coinage also appears to be spreading.12 Though the economists who use the sclerotic analogy all have somewhat different emphases, the varied accounts nonetheless resonate with each other and with the argument here. Many expert observers with detailed knowledge of the economies of Germany, Sweden and continental Europe find an economic inflexibility and an irrationality in economic policy that they deem to be increasingly serious. There were few if any comparable complaints on the continent in the early postwar period - there is apparently a widely observed accumulation of distortions with social ageing. 4
Distinctive institutions and common processes
Those who are sceptical about whether there are any general principles governing economic growth in countries with very different institutions and histories will want to question whether the general model in Rise and Decline and its collateral works can be reconciled with the great institutional and historical differences across countries. Thus I face the challenge of showing how the general theory could be true, given the significance that the idiosyncratic characteristics of each country can have for economic growth. But this challenge is also an opportunity. The institutional and historical differences across countries can also be used to test, in a different way, whether the argument in Rise and Declinefitsthe facts of experience. If we agree about the important special institutional and historical features of a country, and if it is also clear how the general theory, if true, should be manifest in a country with those special institutional or historical characteristics, then we can test to see if exactly this manifestation is in fact observed. If the general principle shows up in the theoretically predicted way in each of a number of very different institutional and historical settings, then that provides additional evidence that the principle is true.
82
Mancur Olson
Just as we know that a fluid is water when we know the temperatures at which it freezes and turns to steam, so we can have more confidence in a theory if the sclerotic processes vary across countries in a way that is consistent with that theory. In testing the argument in this way, it will be convenient to distinguish different classes of cases. Some of these classes have been analysed elsewhere and will only be mentioned here. Another class of cases is sufficiently straightforward that it can be evoked in a couple of paragraphs. Afinalclass, the Nordic-Teutonic class, is more complex and will require a more elaborate analysis. Most of the rest of this paper will be devoted to dealing with this last type of sclerosis, but I shall touch first on those cases that can here be dealt with more briefly. 4.1
Anglo-American
The sclerotic process that manifested itself in the developed English-speaking countries has already been analysed in Rise and Decline, so it is now only necessary to give it a separate name. It can conveniently be called the 'Anglo-American* or (since it covers the developed democracies that were once part of the British Empire) the 'English-speaking' form. The way in which the hyper-pluralistic organizational structures and collusional patterns evolve in the Anglo-American type of context has also been described elsewhere, and thus can readily be compared with the other patterns of evolution that will be referred to or described in the rest of this chapter. 4.2
Mercantilistic
Most of the countries of the Third World constitute another class of cases. The less developed countries have poor transportation and communication systems that, in combination with the difficulties of collective action, normally make organization of rural interests impossible. Their governments tend to be influenced disproportionately by organized small-group interests in the major urban areas, and especially in the capital city. These organized interests gain from protection and subsidies to activities in which most of the Third World countries do not have a comparative advantage. Because of this, they suffer from what I call a 'perverse policy syndrome'. Since I have dealt with this class of cases in Rise and Decline and in some articles,13 I will say no more about it here. Since Adam Smith first dealt with this class of cases (in analysis of pre-industrial Europe in the Wealth of Nations) and diagnosed it in a similar way, let us use his term, 'mercantilism', to label this form of sclerosis. 4.3
Red devolution
This form of institutional sclerosis has also been dealt with elsewhere.14 The types of special-interest collusion that the theory predicts will emerge in societies without freedom of organization and with a Soviet-type economic organization are, of course, dramatically different from those that are predicted to emerge in the developed democracies with market economies. Most of the specific features that the theory predicts will characterize the evolution of Soviet-style societies are also different from those that characterize the sclerotic process in democracies with market economies. Because the Soviet-type societies lack various countervailing
The varieties of Eurosclerosis
83
and corrective forces that characterize the market democracies, they develop a more severe sclerosis than Western societies do. 4.4
Healing-of-divisions sclerosis: southern Europe
Societies differ in the degree to which they are divided over fundamental beliefs. Though there are divisions about government policy in all societies, in some there are also intense disagreements about what fundamental principles should govern the organization of society. The dangers and disadvantages for a society of intense divisions about fundamental issues are well known. Societies that are deeply divided are less stable, and this instability not only directly endangers these societies, but also generates uncertainties that reduce business confidence and especially limit the inflow of foreign investment. What is not so widely understood is that social divisions can also favour economic growth by inhibiting some collective action that reduces social efficiency and dynamism. If collective action is to occur, there must be some limit to the intensity of divisions among those who would engage in collective action. Collective action by definition requires cooperation and concerted effort. Normally, cartelization and collusion to fix prices and wages can occur only if everyone in some industry, craft or occupation is willing to go along with it. A cartel cannot monopolize the supply of any type of good or labour if any suppliers who can supply a large part of total demand refuse to cooperate. This is one of the reasons why organizations for collective action prefer to have a socially homogeneous membership. By and large, the most seriously divided societies in Europe in the early postwar years were the societies that had simultaneously large numbers of Roman Catholics and large Communist Parties: the Mediterranean societies, especially Italy and France. No doubt these societies suffered some loss of foreign investment and some capital flight and other problems because of the uncertainty and tension arising from their social divisions. In general, the social divisions in these societies did not prevent collusion or cartelization among large private firms, since almost all of those who owned or managed substantial private firms were anti-Communist. Since the Communists were never in the central government, the state-owned firms also were not under Communist management (though in Italy apparently there were systematic disagreements at a more encompassing level about some matters between the often left-Catholic-managed government-owned firms in INTERSIND and the privately owned firms in CONFINDUSTRIA).15 So generally the southern European societies, especially in the years before the Common Market, had many powerful trade associations, cartels and oligopolistic collusions. The social divisions did, however, work against effective cartelization of the labour force in the early postwar years. There were separate Communist, Catholic and socialist unions, and these unions were often in conflict not only in the society at large, but also in particular industries, crafts and enterprises. These divisions made it much more difficult for unions to obtain the powerful selective incentives needed to build up stable dues-paying memberships and large strike funds - that is, they worked against closed shops and other forms of compulsory membership. Union dues and union density tended to be lower in southern Europe than in the UK,
84
Mancur Olson
Scandinavia or Germany. (Interestingly, as the present argument would predict, the religious-ethnic division in Northern Ireland is said to have made labour unions at times less strong there than in other parts of the United Kingdom.) The competing unions in each workplace also sometimes negated each other. The social divisions, and each group's fear of being permanently repressed, worked against the solidarity needed for strikes that could outlast employers. The labour unions in the southern European countries were able to stage conspicuous demonstrations, and they could often organize brief strikes, even in the early postwar period. They made the left-wing forces in politics more important than they would otherwise have been. For the most part, however, the societies that were most sharply divided between Catholics and Communists did not have what in the United States has been called 'business unionism'. That is, they did not have unions whose members shared a consensus about the acceptability of the existing social order and thus were able to focus single-mindedly on obtaining the strike funds and control over the workforce needed for effective cartelization of the supply of labour in somefirm,craft or industry. At the same time, the unacceptability to the majorities of a Communist Party in the governing coalitions (until President Mitterrand in France organized a coalition with the Communists that left them impotent) meant that the Communists did not have as large an effect on economic policy as might be expected from their numbers. Thus there are some respects and circumstances in which, paradoxically, 'Communism is good for capitalism'. This paradox should not be overdrawn. The southern European societies were endangered and, in some important respects, damaged by their deep divisions over Communism, Catholicism and what constitutional order should be chosen. Their ideological-religious division could not have had any large impact on the cartelization of firms, and these societies appear to have lost a good deal from that type of cartelization. Still, there can be no doubt that there is an important and neglected element of trjath in the paradox. The foregoing argument about social divisions applies mainly to the early postwar period in France and Italy. The divisions over the constitution of society in southern Europe diminished greatly even before the collapse of Communism. With the collapse of the Soviet-type societies, both Communism and fear of Communism have nearly disappeared. The logic of the theory leads one to expect to see certain special features in the sclerotic process in postwar southern Europe. Cartelization of labour markets would be expected to proceed more slowly than it would otherwise have done. The political focus of labour unions would be relatively larger and the cartelistic element somewhat smaller than in societies of lesser social divisions. The healing of divisions would make the devolution of these societies resemble the English-language and north European societies more as time goes on. As social divisions heal, these societies would reap a direct gain from improved confidence in the investment climate. They would also lose from an increasing institutional sclerosis. In short, the ageing process in southern Europe would, by the logic of the argument, tend to be accompanied by a healing of social divisions that would bring diverse social benefits at the same time that it brings new types of damage from cartelistic collective action.
The varieties of Eurosclerosis
4.5
85
The devolution of encompassing interest organizations: northern Europe
As we recall, our basic logic implies that, if an organized interest is sufficiently encompassing, it will not have an incentive to engage in the anti-social behaviour that a narrow special-interest group has. A disciplined labour union or employers' association that represents, say, 50 per cent of the income-earning capacity of a country will, on average, bear half of any losses it imposes on society. Thus it will cease any redistribution to itself when the marginal social losses that result from this redistribution are twice as large as its gain at the margin from the redistribution. It will have no incentive to engage in the extremely anti-social redistributions that are advantageous for a group that represents only a narrow interest, such as an industry or occupation. Indeed, it has been proven elsewhere16 that if an interest is sufficiently encompassing - if it is a 'super-encompassing' interest - then self-interest, paradoxically, keeps it from taking any redistribution whatever from the rest of society, and motivates it to provide socially optimal supplies of public goods. In a few (mostly small and homogeneous) societies, encompassing interest organizations were created by quasi-constitutional settlements utilizing, among other things, the power of government (or of occupation authorities). In Sweden, the encompassing arrangements were worked out before World War II and (since Sweden was not occupied or even a combatant in the war) continued unbroken through the war. In the case of Norway, Austria and Germany, these arrangements date from after the war. At the beginning, in the cases of Norway and Sweden, and certainly of Austria, it was at most a moderate exaggeration to say that there was only one big union that represented all manual workers and only one big employers' organization representing all substantialfirms.This was never true in Germany, but that country's structure of labour unions and business organizations was nonetheless relatively encompassing, especially by the standards of the English-speaking countries. As the foregoing logic would suggest, the countries with encompassing interest organizations worked very well in the 1950s and 1960s. But as time has gone on, their economic performance has deteriorated considerably and, in the case of Sweden, dramatically. This raises the question: how do societies with encompassing interest organizations change over time? When I was writing Rise and Decline, I had no idea what governed the evolution of such organizations, so that book is cautious and agnostic about how such organizations would work out in the long run. But, as I realized when writing some later articles on this subject,17 the logic of the matter is so clear that I ought to have seen it from the start. Though encompassing political parties and offices in certain types of electoral system can remain encompassing for indefinitely long periods, encompassing interest organizations, such as business organizations and labour unions, are eventually bound to devolve, implicitly or explicitly, into narrow special-interest groups. The underlying logic can best be understood by comparing the incentives that face political parties in countries with electoral rules that generate two-party systems, with the incentives facing organized groups in some other situations. In elections such as those for the House of Commons in the United Kingdom, or in presidential elections in France or the United States, the candidate with a plurality -
86
Mancur Olson
the 'first past the post' - wins. There is no reward for the party that represents a block of voters but never comes infirst.Therefore, though small parties can thrive under proportional representation, there is no incentive under winner-take-all electoral rules to set up political parties to serve the interests of thefirmsin any one industry, or the workers in any one occupation, or any other group too small to have a chance of coming infirst.As has long been known from Duverger's Law, in such electoral systems there is a tendency towards a two-party system. Thus encompassing political parties in countries with plurality-winner electoral systems can last indefinitely. The situation over the long run is dramatically different for encompassing interest organizations such as employers' associations or labour unions. Lobbying and cartelistic organizations do not need a majority or even a plurality of the society to profit from collective action. All a collusion or cartel needs is control over the supply in a single market. All a lobby needs is enough resources to hire a lobbyist, or to make campaign contributions, or to provide enough campaign workers to make a difference in crucial districts, and so on. A trade association offirmsin a particular industry, or the union that represents the workers in a single craft or industry, will normally be large enough to have significant lobbying power, even if the members of the organization in the aggregate obtain only a minute share of the national income. They are therefore in a very different situation from a political party in a country with electoral rules that favour a two-party system. To understand the decisive significance of this for the evolution of encompassing organizations, we begin by assuming an ideally encompassing system with one encompassing labour union representing all workers and one encompassing organization representing all of business. A union representing all workers in the country, if it acts in the best interests of its clientele, strives in the labour market to obtain the highest possible real wages for workers as a whole. The highest possible incomes for workers in this society will be attainable only if the economy is efficient, which it will not be if workers in particular occupations have monopolistic wage levels. Thus an encompassing labour union representing all workers would work for uniform wage levels for each skill level throughout the economy. An encompassing union will often maximize the real incomes of its constituency by using its political power to obtain some income redistribution through government to the income categories in which its membership mainly falls. But it cannot serve its membership by seeking special protection or subsidies for workers in particular industries or occupations: industry-specific and occupation-specific favouritism will reduce the efficiency and dynamism of an economy. In the same way, an encompassing business organization will best serve its clients by opposing any industry-specific orfirm-specificprotection or subsidies, unless they repair specific market failures, and it will also resist any efforts to block entry into any industries, including especially those that have come to have unexpectedly high profits. Consider now an individual industry or occupation in the domain of an encompassing business organization or labour union. Even if the encompassing entity represents business as a whole or labour as whole perfectly, an individual industry or occupation will obtain additional gains if it wins special-interest legislation or a monopoly price or wage for itself. Though special-interest legislation
The varieties of Eurosclerosis
87
and cartelistic prices or wages for an industry or occupation are contrary to the interests of the encompassing constituency as a whole, they are still advantageous for a narrow interest within it. Remember that, so long as an industry or occupation earns only a tiny percentage of aggregate income, it will bear only a tiny share of the social losses from the distortions brought about by a redistribution to itself, but will receive all of the redistribution. Thus, no matter how well an encompassing organization serves the encompassing constituency in which a given industry or occupation falls, industry-specific or occupation-specific lobbying or cartelization can obtain something extra. Since the amount that an encompassing interest group will redistribute to its members is, as we found earlier, limited by its encompassing interest, whereas there is almost no limit to the amount of redistribution that is advantageous for a narrow interest, the activities of the narrow interest will normally be worth much more to its beneficiaries than the activities of an encompassing organization. Paradoxically, this
is true even though many if not all of the narrow interests would have been better off if there had been no special-interest favouritism or cartelization by any group! The reason is that, when the number of narrow interests is large, so the narrow groups do not interact strategically, predation by one narrow group does not affect the likelihood of predation by another. Thus the unilateral action of each group takes no account of the losses to them all from their actions in the aggregate, so they continue with anti-social action even if all of them would have been better off if there had not been any such action. Unexpected developments that bring large gains or losses to thefirmsor workers in a given market can often provide exceptional openings for cartelization or subsidy.18 As a society with encompassing organizations matures, more of its firms and workers will have found themselves in exceptional circumstances where the rewards from abandoning or opposing the encompassing organizations are unusually great. A political entrepreneur who promotes an organization that obtains a surplus by organizing collective action for a narrow group can consume part of it him or herself. Therefore, whenever the firms or workers in an industry or occupation have the small numbers or access to selective incentives that make collective action possible, there is always an incentive to organize for collective action, whether or not the industry or occupation is already represented by an encompassing organization. Nevertheless, it takes a long time before a society comes to have a large number of narrow cartelistic and lobbying organizations. Whether an encompassing organization already exists or not, each industry, occupation or other narrow group must overcome the difficulties of collective action in order to organize. This process slows down the devolution of societies with encompassing interest organizations. On the other hand, the foregoing logic - and especially the logic that shows that it pays narrow interests to redistribute much more than any encompassing organization would - implies that the branch organizations of an encompassing business or labour organization have a powerful incentive to push for the interests of their own branch, even when this is very harmful to the interests of the membership of the encompassing organization as a whole. As subordinate units gain autonomy, and as separate industry or occupational caucuses are formed, an encompassing organization
88
Mancur Olson
will tend to become merely a clearinghouse for the separate interests of its subordinate parts. It might seem that the problem of insubordinate branches and independent coalitions could be controlled by appropriate by-laws for the encompassing interest group, or by government legislation or policies that discourage independent action by branches or other non-encompassing organizations. Corporatist societies such as Fascist Italy and Franco's Spain did that with great severity. In political science and sociology, there are advocates of corporatism (or, as it is now sometimes called, neo-corporatism), and some of them propose government licensing or other devices to give a monopoly to corporatist pressure groups in democratic societies. The governments of some democracies have discouraged independent associations and favoured established encompassing interest organizations in other ways as well, and this has delayed their devolution. But governmental edicts, at least in democracies, cannot permanently prevent the devolution of encompassing interest organizations. There is no way that corporatist legislation in a democracy can prevent a branch of an encompassing organization from pressuring the leadership of an encompassing organization on behalf of the branch's narrow interest. The branch can always press for the right to obtain a monopoly price or wage for itself, or lobby within the encompassing organization to make it support special-interest legislation for the branch. This branch advocacy is bound to have some effect in the long run. The way to get elected to the leadership of an organization is to gain the favour of influential constituents. The leaders of branch units are accordingly important in determining who comes to lead an encompassing interest organization. The leaders of branch units often have an incentive to insist that only those candidates for central office that agree to allow them to set prices or wages separately in their own market, or who promise to support special-interest legislation for those in this market, will receive the branch's support. Those subgroups of an encompassing organization that are not already separately organized in branches can, if they have small numbers or if'selective incentives' can be found, organize a caucus or lobby within the encompassing organization to pressure it to serve the sectional interest at the expense of the encompassing interest. In the long run, how could a democratic society prevent subsets of members of an encompassing organization with a legal monopoly of representation from being controlled in large part by internal lobbies working on behalf of internal subgroups? Even if a democratic government stipulates that only specified organizations are allowed to petition the government, how in practice could it prohibit the creation of internal caucuses or organizations to lobby internally to change the policies of any legally established encompassing lobby? The logic that has just been discussed seems to fit some developments in the Nordic and Teutonic economies. Some observers of Norway and Sweden have emphasized developments that are completely consistent with the logic that I have just set out. Lash has argued, from a Marxist and sociological perspective, 'that there is a long-term trend toward decentralization of Swedish industrial relations'.19 The Swedish economist Assar Lindbeck has argued (personal communication) that the actual policies of encompassing organizations in Sweden have apparently often
The varieties of Eurosclerosis
89
been dictated by relatively small subsets of the membership with special interests; government bail-outs of Swedish shipbuilding were not in the interest of Swedish labour as a whole, but the big union, the LO, presumably at the behest of its constituent units in the shipbuilding industry, favoured such subsidies. Gudmund Hernes has described, in compelling detail, a large number of situations in both Norway and Sweden where once-encompassing interest organizations have been undone by breakaway organizations, by assertive subunits, or by new organizations for collective action.20 No doubt there continue to be significant episodes of encompassing behaviour in some interest groups in Austria, Norway, Sweden, Germany and some other countries. This behaviour has had no counterpart in narrow distributional coalitions. But, as I argued in articles published in 1986,21 we should expect to find fewer examples of encompassing interest group behaviour as time goes on. Though some additional supporting evidence has emerged since then, it remains to be seen whether this prediction will be borne out in the long run. We also need to test the foregoing argument about the devolution of encompassing organizations against evidence on economic performance. Economic growth has gradually been slowing down in all of the pertinent countries, and the deterioration in Swedish economic performance in the last few years has been especially striking. Though this is consistent with the theory in Rise and Decline, that book argues that monocasual theories are wrong and that any complex reality is multicasual; there may well be other factors that account for the apparent consistency of the growth data with the theory. For example, the countries in which encompassing organizations have been devolving tend to have much above average levels of egalitarian or welfare state income redistribution; the exceptionally high levels of income redistribution in Sweden are especially well known, and in the 1950s West Germany redistributed as high a proportion of its income through welfare state programmes as any country in the world. Probably the single best-known explanation of the general deterioration in economic performance in the Nordic and Teutonic economies is that it is due to the magnitude of their programme for egalitarian income redistribution and their correspondingly high tax rates. As we shall see, this extensive redistribution may not be independent of the present argument, but we must nonetheless be wary of the possibility that the economic slowdown is due not to the devolution of encompassing organizations, but to inefficiencies arising from the extensive egalitarian redistributions of income in these societies. Similarly, the severe depression of the Swedish economy in the last couple of years is almost certainly partly due to various macroeconomic mistakes and disequilibria that are in some degree independent of the argument offered here. Therefore, because so many different casual factors are involved, we cannot know, until much more detailed research has been done, how well the available evidence fits the theory. The high level of egalitarian income redistribution in the countries with devolving encompassing interest organizations raises at least three questions. First, why have these countries, and especially Sweden, chosen a more egalitarian redistribution of income than other countries? Second, in view of the huge divergence between marginal private and marginal social returns arising from the
90
Mancur Olson
large welfare state subsidies and taxes, why have these countries done as well as they have? Third, why has the Swedish economy performed worse than the other countries in this category? With respect to why there has been so much income redistribution, other factors may be more important, but I have hypothesized that the early postwar success with encompassing interest organizations led to overconfidence and to 'overshooting',22 especially in the case of Sweden.23 The successful economic performance at the same time that there was a substantial level of income redistribution persuaded many people that very much larger levels of income redistribution could be undertaken with little social cost. Since the deadweight losses from income redistribution almost certainly rise more than linearly with the magnitude of the redistribution, and are multiplied by the distortions from special interest, the overshooting tended to have much higher costs than its advocates expected. On the question of why the countries in this category have managed as well as they have, I have argued elsewhere that 'explicit income redistribution' to low-income people brings about, for several neglected reasons, lower deadweight costs and much less retardation of innovation than the 'implicit income redistribution' in the form of protection, regulation and cartelization that is normally brought about by narrow special-interest groups.24 Therefore, in the early postwar period before their encompassing organizations had devolved much, Sweden and the other countries with encompassing organizations had relatively less deadweight loss from income redistribution in relation to other countries than might have been expected. This helps to explain why Sweden and the other Nordic and Teutonic countries worked as well as they did for so long. On the third question of why Sweden is lately doing worse than the other countries with devolving encompassing organizations, we must remember that none of the other countries, because of their participation or occupation in World War II, have any encompassing organizations with continuity from before World War II. But Sweden does - the 'Swedish model' or 'middle way' has enjoyed continuity (and great international visibility) since the 1930s. Thus the country in which the devolution of encompassing interest organizations has apparently gone on the longest is not only a country that seems to have started overshooting with the welfare state relatively early, but also the country where this devolution has brought about the greatest deterioration in economic performance. 4.6
Different beginnings, similar endings
When one puts the devolution of encompassing interest-group organizations in perspective, the parallel with the Anglo-American or English-speaking sclerotic processes becomes clear. Just as political entrepreneurs have an incentive to organize narrow organizations for collective action in a society without any such organizations, so they have an incentive to create such organizations - or to establish them by breaking off from larger organizations - in a system of encompassing organizations. In both situations, the political entrepreneurs must overcome the great difficulties of collective action, so they will not be able to create many organizations for collective action quickly. But some groups have the small numbers or the access to selective incentives needed for organization, and will, in a
The varieties of Eurosclerosis
91
stable society, eventually succeed in organizing. When thefirmsor workers in some industry or occupation are separately organized, either by the creation of a new organization or by breakaway from an encompassing organization, they have an incentive to use cartelistic and lobbying power to shift the distribution of income to their advantage, and to do this even if the deadweight costs to society are large in relation to the amount they gain. This is true no matter how well the unorganized society or the society with encompassing organizations works. So the political entrepreneur's incentive to establish narrow coalitions for collective action is the same both in unorganized societies and in those starting with ideal encompassing organizations. Though I did not realize this when I wrote Rise and Decline, encompassing organizations may delay - but they cannot by themselves prevent -the emergence of narrow special interests. The result in the long run, if we abstract from any differences across societies in how well they understand the problem, is likely to be much the same whether societies set up encompassing organizations or not. In 1970 Sweden was substantially ahead of the UK, for example, in per-capita income, even though it started modern economic growth much later and had a similarly high level of organization for collective action. Its superior performance surely owed something to the fact that its organizations for collective action were encompassing rather than narrow. But there is no reason whatever to suppose that the societies that have had highly encompassing organizations will stay in front. On the contrary, unless they come to have a better appreciation of the problem than other countries, they seem destined to operate more like societies with narrow organizations for collective action as time goes on. 5
Distinctive institutions and inescapable logic
If the foregoing analysis of 4healing-of-divisions sclerosis' and of the 'devolution of encompassing interest organizations' is read along with other publications on the sclerotic processes in Soviet-type societies, on Third World mercantilistic environments, and on the sclerotic process in the English-language countries, it becomes clear that, important as the differences in the structure of interest organizations across societies are, the similarities in the sclerotic process are nonetheless more fundamental. Encompassing organizations and social divisions are conspicuous, but the inconspicuous accumulation of industry-by-industry and occupation-by-occupation lobbying and collusion is likely to be more telling in the long run. The healing of social divisions and the devolution of encompassing organizations tends to make the continental European societies resemble the English-speaking countries more as time goes on. Important as the institutional and historical differences across societies are, they cannot alter the laws of logic. The logical principle that organized interests, when they have the capacity to influence or control coercive power, have incentives to act in anti-social ways if they have only a narrow or minuscule stake in society, but incentives to act in less harmful and sometimes even socially beneficent ways when they have a sufficiently encompassing stake, applies to all societies. So does the logic of collective action: it is only small groups and those with access to selective incentives than can organize for collective action in any society. Thus groups like
92
Mancur Olson
consumers, taxpayers and the poor are not organized in any society, but in all societies more of the groups with the potential for collective action organize as time goes on. The two logical principles on which the present theory is based are manifest in different forms in different types of society, but they are always (along with other important factors in this complex and multicausal world) evident. The diverse manifestations of the theory in different types of society indicates that the theory has not only parsimony and explanatory power, but also what the nineteenth century scientist William Whewall called 'consilience': the capacity to explain quite diverse phenomena. The foregoing analysis shows that, though a society cannot preserve an encompassing interest-group structure over the very long run, encompassing political parties can be viable indefinitely with the right electoral rules. Thus the logic of encompassing interests argues for electoral rules that discourage small political parties and, ideally, even promote two-party systems. In the United States, where the problem is the weakness of political parties rather than their number, it is institutional rules that make the political parties stronger and more responsible that are needed. The most important implication of the analysis, however, is that the only real solution is for societies to acquire a better understanding of economics and of the present argument. The problem, even in the societies that once had only encompassing interest organizations, is small organized minorities. Each of these tiny minorities will easily be defeated if the public wises up. No historical process that is understood is inevitable. NOTES I am grateful to the US Agency for International Development for support of my research through the Center for Institutional Reform and the Informal Sector (IRIS) at the University of Maryland. This support has facilitated my development of the general theory used here and particularly its application to the formerly Communist countries and the Third World. I have also benefited from a travel grant from the Thyssen Stiftung, and from helpful criticisms by Christopher Bartlett, Carol Kaplan, Nicholas Crafts, Lars Jonung and R.C.O. Matthews. 1 New Haven, CT, and London: Yale University Press (1982). 2 The devolution of centrally planned economies' (with Peter Murrell), initially presented in tentative form to various audiences in the mid and late 1980s and published in the Journal of Comparative Economics, 15, pp. 239-65 (1991); The logic of collective action in Soviet-type societies', Journal of Soviet Nationalities, 1(2), (1990), pp. 8-33; The exploitation and subsidization of agriculture in the developing and developed countries', in Agriculture in a Turbulent World Economy: Proceedings of the Nineteenth International Conference of Agricultural Economists, held at Malaga, Spain, 26 August-4 September 1985, edited by Allen Maunder and Ulf Renborg (Brookfield, VT: Gower, 1986); The hidden path to economic development', in Christopher Clague and Gordon Rausser (eds.), The Emergence of Market Economies in Eastern Europe (Oxford: Blackwell, 1992). 3 The logic of the foregoing argument is demonstrated in The Logic of Collective Action (Cambridge, MA: Harvard University Press, 1965). The findings of much of the best subsequent literature on the topic are cited in Russell Hardin, Collective Action (Baltimore, MD: Johns Hopkins University Press, 1982) and in
The varieties of Eurosclerosis
93
Todd Sandier, Collective Action (Ann Arbor, MI: University of Michigan Press, 1992). 4 The concept of encompassing interests has only been extended to dictatorships and analysed in a formal way since Rise and Decline was published. See, for example, Martin C. McGuire and Mancur Olson, Jr, 'Dictatorship, democracy, and the provision of public goods', Department of Economics and IRIS, University of Maryland, December 1990; The economics of autocracy and majority rule: the invisible hand and the use of force' (IRIS, University of Maryland and University of California, Irvine, 1994'; forthcoming in Journal of Economic Literature); Mancur Olson, Jr, 'Autocracy, democracy and prosperity', in Richard J. Zeckhauser (ed.), Strategy and Choice (Cambridge, MA: MIT Press, 1991); 'Dictatorship, democracy and development', American Political Science Review, 87 (3) (September 1993), pp. 567-76. 5 See 'The South will fall again: the South as leader and laggard in economic growth', Southern Economic Journal, 49 (April 1983), pp. 917-32. 6 See, for example, Kwang Choi, Theories of Economic Growth (Ames, I A: Iowa State University Press, 1983); Richard Vedder and Lowell Galloway, 'Rentseeking, distributional coalitions, taxes, relative prices, and economic growth', Public Choice, 51 (1) (1986), pp. 93-100; Steve Chan, 'Growth with equity: a test of Olson's theory for the Asian Pacific-Rim countries', Journal of Peace Research, 24 (2) (1987), pp.135-49; Erich Weede, 'Catch-up, distributional coalitions and government as determinants of growth and decline in industrial democracies', British Journal of Sociology, 37 (1986), pp. 194-220; Jan-Erik Lane and Svante Ersson, Comparative Political Economy (New York: Pinter, 1990); Todd Sandier, Collective Action: Theory and Applications (Ann Arbor, MI: University of
Michigan Press, 1992); Jonathan Rauch, Demosclerosis (New York: Times Books, 1994); and many of the contributions in the following collections of assessments and tests of the Rise and Decline of Nations: Dennis C. Mueller (ed.), The Political Economy of Growth (New Haven, CT: Yale University Press, 1983); International Studies Quarterly, 27 (1983); Scandinavian Political Studies, 9
(March 1986). , 7 See the citations in note 3 and in my forthcoming book on Capitalism, Socialism and Dictatorship. 8 This possibility, and the role of economic ideas generally, is discussed in 'How ideas affect societies: is Britain the wave of the future?' in Ideas, Interests and Consequences (London: Institute of Economic Affairs, 1989), also republished in The LSE Quarterly, Winter, 1989. 9 Rise and Decline, p. 76.
10 In 'The political economy of comparative growth rates', in US Economic Growth, vol. 2, Joint Economic Committee, Congress of the United States, Hearings of the Joint Economic Committee for November 10, 1976, pp. 105-12; and 'Comment', in Thomas Wilson and Andrew Skinner (eds.), The Market and the State: Essays in Honour of Adam Smith (Oxford: Oxford University Press, 1976), pp. 105-12. There was an international conference in December 1978, at the University of Maryland, to assess and criticize an early draft of Rise and Decline. Though that 1978 draft and some of the papers of commentary presented at that conference were circulated widely, the proceedings of this conference were published only later in Dennis Mueller (ed.), The Political Economy of Growth (New Haven, CT: Yale University Press, 1983). My use of the term 'institutional sclerosis' in the 1978 paper is on pages 35 and 46-9 of that book, and other participants in the conference, such as K. Choi (pp. 57-78), J.-C. Asselain and C. Morrison (p. 158) and D. Mueller (p. 268) also used the term to label my argument.
94
Mancur Olson
11 Herbert Giersch, 'Gegen Euro-Sklerose', Die Wirtschaftswoche, 33 (12 August 1983), and 'Eurosclerosis', Institut fur Weltwirtschaft Discussion Paper No.l 12 (1985). See also the Financial Times (11 September 1984) and Time (30 January 1984). Lindbeck's essay is The emerging sclerosis in the Western economies' (Institute for International Economics, University of Stockholm, 1982). 12 Ingemar Stahl and Kurt Wickman, Riv bostadspolitiken (Stockholm: SNS, 1992). 13 The exploitation and subsidization of agriculture in developing and developed countries', op. cit., and 'Space, agriculture and organization', American Journal of Agricultural Economics, 67(5) (December 1985), pp. 28-37. 14 Peter Murrell and Mancur Olson, The devolution of centrally-planned economies', Journal of Comparative Economics, 15 (1991), pp. 239-65. 15 I am thankful to G. Toniolo for helpful criticism on this point. 16 Martin C. McGuire and Mancur Olson, Jr, The economics of autocracy and majority rule: the hidden hand and the use of force' (IRIS, University of Maryland and University of California, Irvine, 1994; forthcoming in Journal of Economic Literature). 17 'A theory of the incentives facing political organizations: neo-corporatism and the hegemonic state', International Political Science Review, 1 (2) (April 1986), pp. 165-89; and 'An appreciation of the tests and criticisms', Scandinavian Political Studies, 9 (1) (1986), pp. 65-80. 18 The workers in a line of work that is enjoying a boom will gain the most if they receive not only the wage premia that firms want to offer to attract additional labour, but also the far higher increases that come from blocking entry into the booming line of work, or making the wage so high that there is little or no employer demand for additional workers. During the oil boom, the Norwegian helicopter pilots that served the oil rigs in the North Sea could make themselves far better off by demanding the exceptional rises their employers could afford than by sticking with the policy of the encompassing labour union. When an unexpectedly large devaluation of the Swedish krona once made some manufactured exports exceptionally profitable, the Swedish engineering union decided to abandon centralized wage bargaining. Clearly, the encompassing interest of workers as a whole is normally best served by allowing those in other sectors to move into the booming lines, thereby expanding the workforce that is too small and making wages higher than they would otherwise be in the areas that were not so lucky. Similarly, when adverse shocks reduce the demand for, say, shipbuilding labour, the workers in this industry have an incentive to demand subsidies for the industry in which they work. By contrast, workers as a whole lose when a society maintains uneconomic industries that cannot cover their costs without protection or other subsidies. Obviously, there are exactly parallel arguments showing similar conflicts of interest between the firms in a booming or declining industry and the owners of capital as a whole. 19 S. Lash, The end of neo-corporatism? The breakdown of centralized bargaining in Sweden', British Journal of Industrial Relations (July 1985), pp. 215-39. 20 Acta Sociologica, 34 (4) (1991), pp. 239-60. 21 'An appreciation of the tests and criticisms', op. cit:, and The incentives facing political organizations', op. cit. 22 See Assar Lindbeck, 'Overshooting, reform and retreat of the welfare state', De Economist, 142 (1994), pp. 1-19. 23 This is argued in my 'Devolution of the Nordic and Teutonic economies', American Economic Review (May 1995). 24 This and other arguments that bear on the present problem are set out in my book, How Bright are the Northern Lights? Some Questions about Sweden (Lund, Sweden: Institute of Economic Research, Lund University Press, 1990).
4
Why the 1950s and not the 1920s? Olsonian and non-Olsonian interpretations of two decades of German economic history KARL-HEINZ PAQUE
1
Introduction
A major puzzle of German economic history in the twentieth century is the fact that the records of the two postwar decades have been so vastly different - a fast and sustained growth leading to full employment in the 1950s, a slow and rather volatile expansion with persistent unemployment in the 1920s. To explain this puzzle, Mancur Olson and some other prominent scholars - notably the economic historians Knut Borchardt and Harold James - have advanced a theory of institutional discontinuity: in essence, they claim that the Weimar Republic of the 1920s suffered from some major institutional weaknesses which gave interest groups a much strong influence to pursue their distributional objectives than was to be the case in the different institutional environment of the later Bonn Republic.1 In fact, Mancur Olson himself has regarded German economic history as a most important piece of empirical evidence in favour of his celebrated theory of the rise and decline of nations. In this sense, the German case delivers more than just another string of idiosyncratic national history; it is rather a paradigmatic playground for testing one of the most elegant and parsimonious growth theories advanced in the last few decades. This chapter is no more than a critical note on this 'Olsonian interpretation' and an attempt to offer an alternative view that takes the role of interest groups seriously without running into conflict with the basic facts of history. In section 2,1 argue that the major premise of the Olsonian interpretation - the existence of a relatively sharp institutional break between the Bonn and the Weimar Republics concerning the role of distributional coalitions - is untenable. In section 3,1 outline the skeleton of my own interpretation of the puzzle, which may be called a theory of corporatist rigidity under different conditions of unanticipated trade integration and productivity growth. 95
96 Karl-Heinz Paque 2
An Olsonian interpretation of the two postwar records
An Olsonian account of West German economic history after World War II as against the background of the Weimar economy can be summarized in basically three propositions: 1. Through Nazi dictatorship and wartime physical and social destruction and the subsequent Allied occupation of the country, the traditional network of distributional coalitions in Germany was fatally weakened in the decade or so after the war, so that the economy could for a while grow unburdened by the static and dynamic efficiency losses induced by institutional sclerosis. 2. To the extent that the distributional coalitions reappeared early on, they did so in a much more encompassing form that helped to prevent a persistent divergence between the respective coalitions' and society's interests. 3. Over time, the economy was gradually subjected again to the growth-impeding network of distributional coalitions which typically gain ground in stable societies; thus, after all, the German miracle began to fade from the 1960s on. 2.1
A liberation from interest groups?
Any serious discussion of the role of World War II as a historical watershed of whatever kind must first recognize the fact that, in its most fundamental legal, political and economic characteristics, the Bonn Republic as it emerged through the re-establishment of a market economy in 1948 and a parliamentary democracy in 1949 was a descendant of the Weimar Republic of 1919-33. This is not to say that there may not have been substantial differences between the two, which may also matter in the sense of Olson's theory; it is to say, however, that the basic point of reference and even the starting point for deliberate deviations were the Weimar institutions. In this sense, both the Nazi period and the Allied occupation, with their systems of administrative control of political and economic life, were pauses though very different ones - in an otherwise continuous tradition. Both left traces that changed the track of German society for good; but for neither can it be claimed that they influenced the institutional framework of later West Germany more than the 'hysteretic' shadow of the Weimar Republic did. After all, the great institutional transformations in Germany had happened at earlier times: politically from autocracy to democracy towards the end of World War I, legally to a modern Rechtsstaat with an elaborate national system of codified private and public law in imperial times, economically towards a 'corporatist' industrial society with a prominent role for collective bargaining in the early 1920s. Thus there was never really an institutional void that had to befilledwith newly invented laws. For the Allies and the Germans after the war, it was more a question of whether, when and to what extent the pre-Nazi regime would be reimplemented, and which particular changes would have to be made to avoid some of the unfortunate developments of the Weimar period. In this respect, it is the period of Allied occupation from May 1945 until the economic liberalization in June 1948 and the election to thefirstfederal parliament
Olsonian and non-Olsonian interpretations of German economic history
97
in September 1949 that deserves careful examination. Taking a bird's-eye view of this period, it is remarkable how much leeway for collective action the Allies always excluding the Russians in the eastern zone - left to newly founded or refounded German interest groups, and how strongly these organizations represented their traditional pre-Nazi interests. Politically, this can be explained by the specific aims of the Allied forces: after a virtually complete standstill of German political and economic activities in mid-1945 right after the end of World War II, the country was simply to be put to work again, and this was to be done without reactivating a Nazi influence; for this purpose, the extensive use of personnel and expertise of pre-Nazi vintage and origin was obviously indispensable, and the price to be paid in terms of a resurgence of lobbies, coalitions and collusions was viewed to be bearable as long as no overly centralized German organization could challenge the Allies' internal enforcement power. In the light of Olson's theory, three types of organization stand out in importance as distributional coalitions: unions, political parties and business associations. All three re-emerged early with ideas, ideologies, practical purposes and personnel that were in full continuity with Weimar times.2 Most important is the case of the unions. Due to their consistent opposition to the Nazi regime, the unions were viewed very favourably by the Allies - above all, by the union-friendly British Labour government - and they were accepted as an important pillar of a new German economy and society. On a local level, constituent meetings of unions took place as early as summer 1945, and a conference of union deputies of the British zone (which included the industrial heartland of the Rhine/Ruhr valley) was held in December 1945. At this conference, the basic principles of organization of unionism in Germany were laid out in a first statute, which was approved by the British in January 1947. In the American zone, the development was a bit slower and a bit more decentralized - no genuine zone-wide organization was permitted to operate but except in Bavaria, all major industrial unions were established in these regions in the course of 1946 (in Bavaria not before spring 1947). Interzonal conferences of unions were held regularly from December 1946 on. In addition, union membership grew very fast: in the first full year of its existence (1950), the DGB had 5.5 million members - two years later about 6 million - which amounted to a higher density rate than at any time up to the present.3 In short, at least with respect to the quality and the degree of organization, it is difficult to uphold the view that German unions were in any sense weak for more than a very brief initial period of Allied occupation. In a much more narrow sense, however, the unions were handicapped by a very mundane event: the virtual destruction of their strike funds through the currency reform of June 1948. For a while - maybe something like one or two years - most unions were simply notfinanciallypotent enough to carry out any major strike, and that may well have contributed substantially to a rather smooth working of the bargaining process and relatively modest wage settlements (see Giersch et al, 1992: 73). However, this morefinancialconstraint falls well short of any more fundamental weakness of the unions' position that could help to explain a larger chunk of the beginning West German economic miracle of the 1950s. After all, with unions being organizationally strong, one might have expected sharp wage increases and labour conflicts as soon as the strike funds werefilledagain, and this obviously did not happen.
98
2.2
Karl-Heinz Paque
Encompassing unions?
Historically, there were three major attempts of the unions after World War II to make their organizations more encompassing in the Olsonian sense. The first and most radical one was to establish one 'central union' (Zentralgewerkschaft), comparable to the Swedish LO, that was supposed to carry out wage negotiations on a nationwide basis in the future. These efforts failed because of Allied resistance to any such large-scale concentration of power in whatever hands. Recognizing the Allied veto as insurmountable, the union leaders settled for the construction of autonomous industry unions under one umbrella organization, which later became the DGB. Like its predecessor in the Weimar Republic, the ADGB, the new umbrella organization did not in general have a mandate to conclude wage agreements, so the actual economic clout remained with the industry unions, among them the large metalworkers' union (IG Metall, the largest industrial union in the world), which was to play a dominant role as a pacemaker in the mostly annual bargaining rounds during the following decades. Like the ADGB in the Weimar Republic, the DGB grew into the political arm of the union movement, coordinating general aims and targets without interfering in actual bargaining matters. Hence, in this respect, the discontinuity with earlier times appears to be not very significant. The second attempt was to replace the Weimar-type crafts-based unions by strictly industrial ones. In this sense the union leadership was fully successful, and not much controversy arose about this shift of organizational principles (see Hemmer and Schmitz, 1990: 27-8). The main reason for the smooth transition was that, in the early 1930s, there had already been a consensus on the necessity of this kind of reform, but the rise of Hitler and the subsequent suppression of the union movement prevented its being realized. Atfirstglance, this postwar reform seems to be a far-reaching step in the Olsonian sense towards a more encompassing representation of interests (see Olson, 1982: 49). However, the extent of the organizational innovation should not be overrated because major branches of industry had adopted a factual industry organization of labour for a long time. For example, metalworkers, who at times made up 30 per cent of all union members in the Weimar Republic, were organized in the German Metalworkers' Union (the predecessor of IG Metall and already at that time the largest single industrial union in the world), which was founded in 1891 as an industrial union and which carried out its collective bargaining consistently on an industrial basis. Of course, the backbone of this union (as of all others) was skilled workers, who made up about two-thirds of the membership, but actual negotiations covered all skill groups, including unskilled and semi-skilled workers.4 Formally, there were still many different craft associations or unions within the industry, but they were usually united under special cartel arrangements (e.g. in the branches related to metal manufacturing, the Metalkartell), which explicitly delegated their right to conclude collective agreements to the German Metalworkers' Union. Apparently, they no longer played any significant role in collective bargaining.5 Hence, after all, the gap between the Weimar and the Bonn collective bargaining culture was probably much narrower than, say, between the traditional crafts-based British and the German system, be it of the Weimar or the Bonn variety. The third attempt was to absorb various smaller unions of different ideological
Olsonian and non-Olsonian interpretations of German economic history
99
standings - notably the more conservative Christian and the tiny liberal ones - as well as the union of white-collar workers in the mainstream of the DGB. In the end, after painful deliberations, this attempt failed mainly because the large DGB quickly developed an identity of blue-collar worker representation with strong political sympathies for reformist socialism, i.e. in the range of the German party system, for social democracy.6 Hence, just like in the Weimar Republic, the union movement remained split, although it is fair to say that the relative weight of the dominating blue-collar umbrella organization, the DGB, was becoming even more decisive than it had been in Weimar times with respect to the ADGB. To sum up, there were certainly trends towards a more encompassing union organization in the early postwar period, but it is hard to see anything revolutionary in them. They appear to be more like thefinalconclusion of developments that had been well under way in the Weimar Republic. In any case, there was no sharp 'Olsonian' discontinuity that may explain a dramatic shift from aggressive wage demand based on narrow group interests towards moderation in the interest of the whole economy. Note that there was in fact one institutional change in the legal framework of collective bargaining from Weimar to Bonn which was probably much more important that the marginal shifts of union organization: the end of compulsory arbitration.7 Whereas, in the Weimar framework, an industrial dispute could be settled by a rather complicated procedure which culminated in compulsory arbitration under the auspices of the Federal Ministry of Labour, the new law on collective bargaining of 1949 did not contain any such provisions of state intervention in industrial disputes. Again, the explicitly wide interpretation of private bargaining autonomy was a consequence of the bad Weimar experience with this system. Instead of being used as an instrument of last resort, it actually removed the pressure for agreement from the parties and thus induced them to carry on with maximalist positions and to speculate on the arbitration's likely bias. In all major industries, at least half of the 'agreements' in the later years of the Weimar Republic were implemented via compulsory arbitration (Hartwich, 1967:418-20).8 Industrial relations remained bad, not least because most compulsory agreements were explicitly disapproved by one party, in the majority of cases the employers' side, and sometimes even by both parties. It is important to realize that the removal of compulsory arbitration and its likely consequences are outside the scope of Olson's theory. If anything, his theory would predict them to have a negative impact on growth because they dissolve the link between the behaviour of distributional coalitions and the public interest via government intervention. 2.3
Worsening distributional sclerosis over time?
If it is difficult to defend the view that there was a wholesale demise of distributional coalitions after World War II, it is equally difficult to identify an Olsonian process of sclerosis thereafter: once established, interest groups were able to lobby for their purposes just as well in the early 1950s as in later decades.9 This is particularly obvious for unions because the framework of collective bargaining remained remarkably constant over time: there were only very few labour law extensions and
100 Karl-Heinz Paque
court decisions specifying the conditions of strikes and lockouts, and these were due to technical progress and the ever more perfect logistics of firms. At any rate, these changes cannot remotely explain the vast variations in wage behaviour between the 1950s and the 1980s on the one hand - with a persistent decline of real unit labour costs, but different unemployment records - and the first half of the 1970s on the other, when labour costs rose sharply and aggravated the labour market plight. 3
A non-Olsonian mode of interpretation
By its very logic, a non-Olsonian interpretation of the German economic performance of the 1920s and the 1950s has to be based on a twin pair of assumptions: (1) that the relevant institutional framework remained by and large unchanged between the two periods, and (2) that there are forces external to this constant institutional framework which can be made responsible for the difference of economic performance in the two periods. 3.1
Collective bargaining and unemployment inertia
Focusing again on the system and practice of collective bargaining as one central element of the relevant institutional structure, one may say that a simple monopoly union or a bargaining model can do a reasonably good job in portraying the main features of wage determination both in the Weimar and in the Bonn economy. A typical characteristic of these models is that they imply a stable long-term relationship between the level of the real wage and the level of unemployment. This means that - unlike in a Phillips curve world - a once-for-all rise in unemployment has no long-run depressing effect on real wage growth, so there is no self-correcting mechanism to bring unemployment back to a prior 'natural' level. Thus unanticipated shocks to labour demand have strong persistence effects which tan be explained in the last resort by the working of insider interests of the incumbent employed labour force vis-a-vis unemployed outsiders.10 It is important to realize that these models have in principle symmetrical implications with respect to 'good times' and 'bad times', depending on which way the unanticipated shock goes. If, for example, there is a series of unanticipated positive supply shocks - say, labour productivity or the terms of trade following a stochastic trend, but with a few successive random changes with a positive sign then there may be a non-inflationary expansion for a longer period of time (provided, of course, there is a sufficiently large labour surplus to be re-employed). Hence, in such a framework, the issue of the different output and employment growth performance of the 1920s and the 1950s comes down to the question of whether there was a sequence of unanticipated positive shocks in the latter which was lacking in the former period. Note the important fact that our non-Olsonian framework takes the institutions of collective bargaining very seriously as an element of what might be called 'passive rigidity' in the system - unlike Olson, who interprets them as a factor of 'active sclerosis'. In an Olsonian setting, the gradual cartelization of the labour market leads to an equally gradual strangling of market-type adjustment mechanisms; it thus turns a good world into a bad one, and an abrupt decartelization turns a bad
Olsonian and non-Olsonian interpretations of German economic history
101
world back into a good one. In our non-Olsonian setting, the cartelization amplifies economic performance: in terms of employment growth, it makes good times better and bad times worse. Note also that this non-Olsonian concept of passive rigidity need not be confined to the labour market, but may re-emerge in other types of cartelization as well - basically in all areas which Olson (1982) touches upon. For instance, a series of positive supply shocks to industry may ease structural change between the primary and the secondary sector of an economy so as to make part of farmers' protectionism pointless because many farmers voluntarily leave agriculture to accept emerging well-paid jobs in industry. Just as Olson developed a general theory of active sclerosis, so one may draw the outlines of a theory of passive rigidity that helps to account for many phenomena of social and economic life. However, this kind of theory can never be complete because it explains only why the system has no strong capacity to absorb particular positive or negative shocks. One still needs an answer to the question of why these shocks came about or did not come about in the first place. 3.2
External forces
Giersch et al. (1992: section 3.A.3) and Paque (1994: section 3) have shown that there is considerable empirical evidence for interpreting the West German 'economic miracle' of the late 1940s and the 1950s as a long sequence of positive, unanticipated productivity and terms of trade shocks, which, by turning outsiders into insiders in the labour market, made unit labour costs fall in real terms over a full decade. Only in the very early postwar years were these shocks the immediate effect of a genuine reconstruction in the sense of a very high profitability of (mostly repair) investments in a war-damaged industrial capital stock. From the early 1950s on, they must rather be regarded as the consequence of the extremely rapid integration of the West German economy into an international division of labour, which assigned an ever more prominent place to intra-European, intra-industry trade. With its traditionally strong manufacturing industries, notably in the production of investment goods and in chemicals, the West German economy was particularly well placed to take full advantage of this intensification of trade links. Given the large supply of skilled and highly mobile workers that had been added to the West German labour force through the postwar influx of ethnic Germans from Eastern Europe, the export-led expansion of manufacturing did not meet hard supply constraints and could thus turn into a genuine ^industrialization of the economy, with the share of industry in total employment increasing until the early 1960s, which was an unusual phenomenon for a highly developed industrial country, even at that time.11 Concerning the difference between the 1920s and the 1950s, the question thus becomes: why did a series of positive shocks happen in the 1950s and not in the 1920s? To be sure, recent research in the wake of the so-called Borchardt controversy12 has shown that, by international standards, the German productivity growth performance of the 1920s was not bad at all, and that, towards the end of the decade, a large part of the shares in international markets that had been lost in the course of World War I and its inflationary aftermath was regarhed by German firms. While the evidence is not without ambiguities,13 it does indicate that the German economy was well on its way to resuming its proper place in the international
102
Karl-Heinz Paque
division of labour. Even before the Great Depression, however, the unemployment rate was stuck around 8 per cent, not far off the scale of the 1980s. Although not bad by other standards, productivity growth was apparently not powerful enough to surpass the prior expectations that underlay collective wage agreements. Hence we are still left with the question: why did such a formidable series of (unanticipated) positive shocks from trade integration occur only in the 1950s and not twenty years earlier? In actual historical sequence, it was of course the Great Depression and its consequences, notably the subsequent Nazi autarky policy, which definitely stopped the further reintegration of Germany into world markets. As it happened, the Great Depression was first and foremost a gigantic macroeconomic accident originating in the United States, and hitting the German economy particularly hard because of its high dependence on short-term transatlantic capital flows, which in turn were an indirect consequence of a whole bunch of complex unresolved political issues going back to the ill-designed and ill-fated Versailles treaty.14 One can hardly deny that the breakdown of international coordination and the resort to protectionist means in the course of the Great Depression destroyed an international division of labour and a network of international trade relations which had been beneficial to its participants and which was per se not doomed to fail. If this is so, however, it is legitimate counterfactually to assume away the Great Depression and ask what might have been the path of economic history if this most unfortunate macroeconomic event had not taken place. Was there a realistic chance for an export miracle in the style of the 1950s? Probably not, at least not in the near future after 1929. The reasons lie in what might be called a lack of trade potential in technology and a lack of cooperative spirit in international politics. Technologically, there was no general backwardness in continental Europe and, for that matter, in Germany vis-a-vis the United States in all industrial branches, but there was a very differentiated pattern, with Germany still having the lead in chemicals and electrical engineering, but clearly lagging behind in those branches like vehicle production where the revolutionary productivity advances through the innovation of the assembly line and related production techniques had been made in the United States during the 1920s. While these techniques were increasingly applied in Germany as well, the country was probably still some way off the threshold level at which production costs fall so low that a mass market for middle-class income earners emerges, with all the backward and forward linkages that the production of vehicles implies. After all, the American car industry had already produced an impressive 4.5 million passenger cars in its first peak production year of 1929, which translates into an annual supply of 37 cars per thousand inhabitants, a level not reached in France and Germany, the main continental European car-producing countries, until the early 1960s. Behind this assessment, there is the implicit idea that waves of productivity improvements in the course of a rapid process of integration, which goes along with a rapid rise in intra-industry trade, require the passing of some threshold levels of technological implementation and of income levels of the population before a mass market and the means to supply it can simultaneously develop. It is doubtful whether these conditions were already prevailing in Europe in the early 1920s. Politically, Europe in the 1950s enjoyed the decisive advantage of being subject to
Olsonian and non-Olsonian interpretations of German economic history
103
a 'trade hegemon', the United States, which had a strong interest in free trade both across the Atlantic and, for reasons of economic and political stability, within Europe itself (see Giersch et a/., 1992:95-105). Again, in this respect, the experience of World War II and the subsequent Cold War provided an almost ideal background to overcome traditional political frictions. A much more realistic ex-post scenario as an alternative to the Great Depression is probably that Germany would have continued its reintegration into the world economy at about the pace of the late 1920s, reaching and then surpassing the prewar export share some time in the mid-1930s. However, in such a scenario, there would have been no inherent mechanism to reduce unemployment substantially below the level inherited from the late 1920s. Without the wartime backlash, this moderate growth might even have accelerated later, say, in the 1940s, by which time the continental European countries would have gradually grown into the threshold range of technology and income, where the 'expected' cumulative process of mechanization and motorization of their economies would have set in. Again, to the extent that the relevant productivity advances would have been unanticipated, there would have been a decline in unemployment. However, given the assumed continuity of institutions, the degree of unanticipation would probably have been much less pronounced than it turned out to be in the actual history of the 1950s. Hence, while the productivity level may end up to be the same in this counterfactual world as it actually was in reality - once all transitory disturbances like wars and subsequent periods of 'reconstruction' or 4catching-up' have run their course - the long-term persistent changes in the equilibrium unemployment rate would have been much more moderate.15 Admittedly, all this is highly speculative - more so than professional historians might find acceptable. However, it should at least suggest that the future path of research into the causes of the rise and decline of economies or even nations may well lead in a direction which is slightly different from Olsonian thinking.
NOTES This chapter draws heavily on a CEPR discussion paper by the author (Paque, 1994). In that paper, all the major lines of reasoning of this chapter are presented in a much more extended form, and additional quantitative evidence on the main points is given. 1 See Olson (1982, 1987) which are in essence applications of Olson's theory of collective action as advanced in Olson (1965). For a brief summary account of Olson's theory as applied to economic growth and some conceptual difficulties it encounters, see Paque (1993b: 1-4). In more detailed analyses of the Weimar economy, some major ideas of Olson have been taken up by James (1986) and Borchardt (1990), although these authors do not engage in an explicit comparison of the post-World War I and II periods. Implicitly, however, they do assume some major institutional discontinuity between Bonn and Weimar because otherwise they would not be able to account for the difference in the relevant economic records. 2 For an authoritative historical account of the resurgence of these organizations after World War II, see Eschenburg(1983:171-218). For a summary evaluation of the re-emergence of business associations and political parties in the light of
104
Karl-Heinz Paque
Olson's theory, see Paque (1994: section 2). 3 In 1950-2, about 38-9 per cent of all employees were organized in the DGB, a share that gradually declined to roughly 30 per cent by the mid-1960s and - after a sharp rise in the early 1970s - levelled off at about 33 per cent. If the union of white-collar workers (DAG) and the small Christian unions are included, the density rate was somewhat higher throughout, with about the same intertemporal pattern. For statistical details, see Hemmer and Schmitz (1990). In the Weimar Republic, the density rate reached a peak of about 33 per cent in 1920-2, but declined sharply after the stabilization crisis of 1923/4 to reach a low of 15 per cent in 1925. By 1929, it had risen again to 22 per cent. Even if (low-unionized) agricultural employment is excluded, the density rate in the period 1925-9 never surpassed one-third. (Own calculations based on trade union membership data by Moses (1982: vol. II, 512) and employment data by Hoffmann (1965:205-6).) 4 For details, see Hartwich (1967: 70-2). 5 For details of the structure of organization of metalworkers in one important regional bargaining district, the metropolitan area of Berlin, see Hartwich (1967: 65-70). 6 After a brief period of merger with the DGB, the Christian unions split off in the early 1950s as they did not recognize their more conservative stance sufficiently represented in the official union position. As to white-collar workers, a separate nationwide union (the DAG) was founded in 1949 after unsuccessful negotiations with the DGB. 7 In all other respects - including, for example, special provisions concerning the possibility of declaring a collective agreement generally binding under specific circumstances - the legal frameworks for collective bargaining in the Weimar and Bonn Republic are remarkably similar. See Paque (1993a). 8 For a detailed account of the Weimar experience with compulsory arbitration, see Bahr (1989). 9 As early as 1954, the renowned German historian Theodor Eschenburg held a much celebrated public lecture series entitled 'Herrschaft der Verbande' (later published as Eschenburg (1955)) in which he described and criticized the strong influence of organized interests in public decision making on the political and administrative level. 10 On monopoly and bargaining models of unionism in general, see Farber (1986), Oswald (1986) and Calmfors (1990); by now classical insider/outsider models had been developed by Blanchard and Summers (1986) and Lindbeck and Snower (1986). On their being quite realistic descriptions of the German labour market and industrial relations in the 1920s and 1950s, see in detail Paque (1994: section 3). 11 For a detailed descriptive account of this non-inflationary process of expansion, (see Giersch et al, 1992: section 3.A), where it is also argued on the basis of casual empirical evidence that there was a strong and persistent tendency in the public to underestimate the growth dynamics of this process until well into the late 1950s; hence the trend path of growth can reasonably be regarded as unanticipated. 12 Notably Balderston (1993). The main Borchardt hypothesis that the economy of the Weimar Republic suffered from inherent supply-side deficiencies - above all, a badly working institutional framework - was developed and expounded in Borchardt (1979, 1990); major contributions apart from Balderston (1993) have been made by Holtfrerich (1982, 1983), James (1986) and Ritschl (1990); Kruedener (1990) presents a useful collection of essays on the major issues of the controversy.
Olsonian and non-Olsonian interpretations of German economic history
105
13 For details, see Paque (1994: sections 1 and 3). 14 See, among others, the accounts by Aldcroft (1977) and Kindleberger (1973). 15 There may, of course, be many more differences in details between our counterfactual scenario and the real world. For example, without World War II, there would not have been a footloose labour supply of refugees in western Germany, but rather a larger share of agricultural in total employment. To the extent that agricultural workers would be less ready to leave their land and to search for industrial jobs than the refugees in actual history, the process of reindustrialization would have been more protracted than it actually was in the 1950s. REFERENCES Aldcroft, D.N. (1977) From Versailles to Wall Street, London: Allen Lane. Bahr, J. (1989) Staatliche Schlichtung in der Weimarer Republik, Berlin: Colloquium. Balderston, T. (1993) The Origins and Course of the German Economic Crisis 1923-1932, Berlin: Hande & Spener. Blanchard, O. and L.H. Summers (1986) 'Hysteresis and the European unemployment problem', NBER Macroeconomics Annual, 1, pp. 15-78. Borchardt, K. (1979) 'Zwangslagen und Handlungsspielraume in der groBen Wirtschaftskrise der friihen dreiBiger Jahre: Zur Revision des iiberlieferten Geschichtsbildes', in Bayerische Akademie der Wissenschaften Jahrbuch 1979, Munich. (1990) 'A decade of debate about Briining's economic policy', in Kruedener (1990). Calmfors, L. (ed.) (1990) Wage Formation and Macroeconomic Policy in the Nordic Countries, Oxford: Oxford University Press. Eschenburg, T. (1955) Herrschaft der Verbdnde?, Stuttgart: Deutsche Verlags-Anstalt. (1983) Geschichte der Bundesrepublik Deutschland, Volume 1: Jahre der Besatzung 1945-1949, Stuttgart: Deutsche Verlags-Anstalt. Farber, H.S. (1986) T h e analysis of union behaviour', in O. Ashenfelter and R.G. Layard (eds.), Handbook of Labour Economics, vol. 2, Amsterdam. Giersch, H., K.-H. Paque and H. Schmieding (1992) The Fading Miracle: Four Decades ofMarket Economy in Germany, Cambridge: Cambridge University Press. Hartwich, H.H. (1967) Arbeitsmarkt, Verbdnde und Staat 1918-1933, Berlin: Walter de Gruyter. Hemmer, H.-O. and K.T. Schmitz (1990) Geschichte der Gewerkschaften in der Bundesrepublik, Cologne: Bund-Verlag. Hoffmann, W.G. (1965) Das Wachstum der deutschen Wirtschaft seit der Mitte des 19. Jahrhunderts, Berlin, Heidelberg, New York: Springer. Holtfrerich, C.-L. (1982) 'Alternativen zu Briinings Wirtschaftspolitik in der Weltwirtschaftskrise?', Historische Zeitschrift, 235, pp. 605-31. (1983) 'Zu hohe Lohne in der Weimarer Republik? Bemerkungen zur BorchardtThese', Geschichte und Gesellschaft, 10, pp. 122—41. James, H. (1986) The German Slump, Oxford: Clarendon Press. Kindleberger, C.P. (1973) The World in Depression, London: Allen Lane. Kruedener, J. Baron von (ed.) (1990) Economic Crisis and Political Collapse: The Weimar Republic 1924-33, New York, Oxford, Munich: Berg. Lindbeck, A. and D. Snower (1986) 'Wage setting, unemployment and insider-outsider relations', American Economic Review, 76, pp. 235-9. Moses, J.A. (1982) Trade Unionism in Germany from Bismarck to Hitler 1869-1933, Vols. I and II, London: Priorr
106 Karl-Heinz Paque Olson, M. (1965) The Logic of Collective Action: Public Goods and the Theory of
Groups, Cambridge, MA: Harvard University Press. (1982) The Rise and Decline of Nations: Economic Growth, Stagflation and Social
Rigidities, New Haven, CT, London: Yale University Press. (1987) 'Some questions about the Weimar republic and possible parallels to the developed democracies today', in P. Koslowski (ed.), Individual Liberty and Democratic Decision-Making, Tubingen: J.C.B. Mohr (Paul Siebeck). Oswald, A. (1986) The economic theory of trade unions: an introductory survey', in L. Calmfors and H. Horn (eds.), Trade Unions, Wage Formation and Macroeconomic
Stability, Basingstoke: Macmillan. Paque, K.-H. (1993a) 'Germany: living with tight corporatism', in J. Hartog and J. Theeuwes (eds.), Labour Market Contracts and Institutions: A Cross-National
Comparison, Amsterdam: North-Holland. (1993b)'How clean was the slate? Some notes on the Olsonian view of the postwar German economic miracle', Kiel Working Paper No. 588, Kiel: Institut fur Weltwirtschaft. (1994) The causes of slumps and miracles: a critical evaluation of Olsonian views on the German economic performance in the 1920s and the 1950s', CEPR Discussion Paper No. 981. Ritschl, A. (1990) 'Zu hohe Lohne in der Weimarer Republik? Eine Auseinandersetzung mit Holtfrerichs Berechnungen zur Lohnposition der Arbeiterschaft 1925-1932', Geschichte und Gesellschaft, 16, pp. 375-402.
5
Convergence, competitiveness and the exchange rate ANDREA BOLTHO
1
Introduction
Recent years have seen a burgeoning literature on the issue of why countries' per-capita income levels converge (or do not, as the case may be). This topic is hardly new, of course, and much of the present research is merely rediscovering arguments that had long been known (the advantages of relative backwardness, the role of skills, the importance of investment or of learning by doing, etc.). There is, however, one feature of the older literature which recent writings have hardly exploited: that is, the role which government intervention (through, for instance, macroeconomic, industrial or exchange rate policies) could play in promoting growth. This neglect is understandable. In the context of the dominating orthodoxy, industrial policies can only lead to inferior outcomes, while macroeconomic and exchange rate policies, when not instantly ineffective, must be so in the medium term. Similarly heretic, and hence ignored, are approaches of the 'export-led growth' variety which dare to suggest that demand factors may contribute to longer-run growth differences. Yet, just a cursory look at the economic history of the OECD countries since World War II suggests that the role of demand and policies cannot be ignored, notably in the 1950s (witness the importance that many attribute to the indicative planning of France, to the investment subsidies of Germany, to the industrial policies of Japan, and so on). In particular, policies may have influenced one area that seems crucial for an understanding of the convergence process, that of international competitiveness. Traditional approaches usually view the latter as an entirely endogenous function of the growth process itself. Should competitiveness have an exogenous (e.g. policy-determined) component, however, then the convergence process may, in part at least, be explained by forces other than the usual supply factors advanced in the literature. The text that follows, by looking at one such (partly) exogenous force - the exchange rate - tries to see to what extent devaluations of the currency can generate not only short-term improvements in price competitiveness, but also longer-term improvements in non-price competitiveness. This is done in the context of Western 107
108
Andrea Boltho
Europe's 'Golden Age', a period in which convergence was rapid and exchange rates were, at times, allowed to vary. After a methodological introduction to the issue in section 2, Europe's overall experience is surveyed in section 3 and briefly contrasted with that of the interwar period. Section 4 considers in somewhat greater detail the economic histories of Germany and Italy in the 1950s, and those of France and Spain in the 1960s. Thefinalsection summarizes the main arguments and provides what turn out to be some rather inconclusive conclusions! 2
Competitiveness and the exchange rate
The proposition that catch-up in per-capita incomes implies a high degree of competitiveness would seem to be almost self-evident for the open economies of postwar Europe. Converging on a richer country means growing relatively rapidly; above average growth, in turn, usually leads to high import penetration; external constraints (be they imposed by a fixed exchange rate system and/or policy unwillingness to countenance rising external indebtedness) therefore require a parallel expansion in exports and/or in import-substituting capacity.1 In either case the country must be able continually to preserve the international competitiveness of its tradable sector. Much less obvious are the conditions that lead to such a state of high competitiveness. Standard demand theory might suggest that this should be a function of differences in relative costs or prices between countries. Yet, international monetary theory usually argues, with some plausibility, that differences in price competitiveness occur only in the short run (e.g. because of differential inflation rates or exchange rate changes). Over the longer run, however, prices are unlikely to vary much across countries because purchasing power parity re-establishes itself. By default, therefore, longer-run cross-country differences in competitiveness are attributed to so-called non-price factors. Partly, perhaps, because such factors are very difficult to quantify, their explanation has usually been left fairly vague. Overall, however, the consensus is that they are largely a function of the supply side, a consensus that embraces writers as different as Kaldor and Krugman. Thus, the former explains high income elasticities of demand for a country's exports in terms of'the innovative ability and adaptive capacity of its manufacturers' (Kaldor, 1981:603). The latter, on the other hand, explains the same phenomenon via a model in which, thanks mainly to the operation of scale economies, 'fast growing countries expand their shares of world markets . . . by expanding the range of goods that they produce' (Krugman, 1989: 1039). And some empirical backing is provided for these views by results that link longer-run changes in export and import market shares to technological and investment variables, rather than to price or cost competitiveness (Fagerberg, 1988). While all this sounds eminently plausible, it is not very helpful for policy-making purposes. As so often in the growth literature, the outcome seems to boil down to virtuous (or vicious) circles - rapidly growing countries enjoy high competitiveness which reinforces their rapid growth, and vice versa. Models of export-led growth had, long ago, come to similar conclusions (Beckerman, 1962). The interesting policy-oriented question is whether tools exist that could improve the performance of a country's tradable sector and allow an entry into such a virtuous circle.
Convergence, competitiveness and the exchange rate
109
Endogenous growth theories do, of course, suggest a number of variables that might contribute to international competitiveness and hence to rapid growth (e.g. investment in R & D, in human capital or even in infrastructures), but the empirical backing for these theories is, as yet, very weak. In any case, for the developed market economies of Western Europe, in which spending on R & D is usually subsidized and human capital or public infrastructure levels are already relatively high, such a list is not particularly helpful. And the number of remaining instruments may be limited. Industrial policy could be a potential candidate, but it is one that has increasingly received a bad press, East Asia's apparent successes notwithstanding. Alternatively, one could select the exchange rate. This may sound paradoxical. It was argued above that relative prices are unlikely to alter in the longer run. Hence currency devaluations, for instance, can have only temporary effects, as is the case, more generally, for monetary policy changes. Indeed, this view of neutrality is embedded in standard international economics textbooks (e.g. Krugman and Obstfeld, 1988). Output and employment may be temporarily boosted by devaluation, but ultimately everything returns to its 'natural' level except for prices, which end up permanently higher. Yet, just as money neutrality may not be quite as axiomatic as standard theory suggests, similarly exchange rate changes may, after all, have more than merely temporary effects, even if one accepts that any price advantage is eventually eroded by higher inflation. Thus, according to one (non-orthodox) school of thought, the short-run improvements in price competitiveness conferred by devaluation could well be more than swamped in the longer run by devaluation-induced worsenings in non-price competitiveness. These would arise, inter alia, from a shift in production towards price-sensitive goods that are less likely to be in high demand on world markets, from a slowing down in resource reallocation away from low value-added activities, now that the latter have been given a temporary lease of life, and from the protection-like effects on X-inefficiency and on the dulling of competitive pressures that could follow from a devaluation. Taken to their logical conclusions, such views would imply that currency appreciation rather than depreciation might be the road to longer-run competitiveness. An alternative (and equally non-orthodox) view is, however, also plausible. At the microeconomic level it has recently been suggested that the presence of sunk entry costs in imperfectly competitive industries may lead to persistent effects on competitiveness coming from transitory (if large) exchange rate changes (Baldwin, 1988). Once foreignfirms,for instance, have entered domestic markets in the wake of home currency appreciation, 'they will find it profitable to remain in [these] markets even at a lower exchange rate' (Baldwin and Krugman, 1989: 635). Conversely, 'once...firmshave abandoned markets, a mere return of the exchange rate to the former level will not be enough to make the expensive recapture of these markets worthwhile' (Baldwin and Krugman, 1989: 635-6). And there may be more to changes in currency values than these effects on the numbers of firms that are present in a market. Each individual firm's non-price competitiveness may also be raised, since a depreciation of the exchange rate almost inevitably generates higher profits. While these profits could, of course, be squandered in high dividends and/or high salaries, they could, alternatively, be used to raise capacity, to launch advertising campaigns, to improve product quality, to
110
Andrea Boltho
finance R & D spending, etc. As was rightly said in a similar context: 'relative selling effort and availability determine relative sales. Selling effort and availability [in turn] are greatest when the seller has the resources, which means the biggest . . . profits' (Brown, 1979: 8). If the short-term opportunities that devaluation provides, in the form of such higher profits, are seized upon, then arguably, by the time the price advantage has evaporated,firmscould still find themselves more competitive than before. Similarly, there may also be macroeconomic channels of transmission. Models of export-led growth, for instance, have long suggested that an undervalued exchange rate could have beneficial effects that went beyond a short-run boost to exports and the balance of payments. Thus, higher exports could stimulate investment via the accelerator, while a favourable current account could allow governments to follow more expansionary policies. Both these trends might further add to investment by raising business confidence (Beckerman, 1962; Lamfalussy, 1963). The virtuous circle could then be completed by a supply response in the form of scale economies and more rapid productivity growth, which would perpetuate the competitive advantage (Cornwall, 1977). In other words, what is being proposed is another version of what has recently been called hysteresis and was, earlier, known as path-dependence - the idea that temporary shocks may have more than merely temporary consequences. For appreciation this has often been argued in the recent past: 'If the level of... productivity depends on past levels of output, then a fall in output today, due to, say, a loss in competitiveness, will lower productivity in the future and reduce supply' (Bean, 1988: 59). A similar mechanism in reverse could also apply to depreciation, with higher profitability generating longer-run beneficial effects through higher exports, capital formation and productivity growth. In principle, as with any investment, one might expect that diminishing returns would set in to 'investment in exporting', but learning by doing and the displacement of foreign rivals could offset such tendencies and raise the growth rate of exports (or of import substitutes) over a prolonged period. 3
Western Europe in the 'Golden Age'
Whether the hysteresis effects suggested in section 2 can be detected in practice is more difficult to determine. The following few paragraphs will look at Western Europe's experience in the 1950s and 1960s, a period that would seem ideal for testing the approach. For one thing, these were years during which convergence in per-capita income and productivity levels was very rapid. Between 1950 and 1970, for instance, labour productivity grew at more than A\ per cent per annum in countries such as France, Germany, Italy and Spain, but at only 2 per cent per annum in the United States. More importantly, these were thefixedexchange rate years of the Bretton Woods system, in which parity changes did occur but were, on the whole, few and far between. If devaluation of the currency is to have favourable longer-term effects, it would seem that pegged systems offer far more opportunities than floating ones. In the latter, the likelihood of strong supply responses may well be limited if, as has been the case since the early 1970s, the persistence of real exchange rate changes is in doubt.
Convergence, competitiveness and the exchange rate
111
Experience of floating so far suggests that all major currencies have risen and fallen in real terms, and often by large amounts and in quick succession. In such circumstances, the response of risk-averting firms to sudden depreciations that are likely to be viewed as temporary may be a quick attempt to increase sales, but not a commitment to a long-run switch into tradable production. The short-run apparent price elasticities could actually be somewhat higher as a result of such behaviour than they would have been in a more predictable exchange rate environment, but the longer-run ones would be lower. Conversely, in the Bretton Woods days a devaluation of the currency would probably have been viewed by most firms as conferring a semi-permanent competitive advantage, particularly in circumstances in which labour markets were stillflexibleand money illusion was widespread. This, in turn, could have led to long-term changes in strategy as firms reallocated not only production, but also investment, to the more profitable export and import-competing activities. If this had indeed been the case, one would expect to see exchange rate changes affecting the growth performance of various European countries in these two decades. The first, and very simple, procedure followed to test for this hypothesis was to begin by establishing which major forces accounted for convergence during the period and then to see whether the exchange rate was able to throw light on the behaviour of any of these major forces. Differential productivity growth rates in this sample of relatively developed economies have traditionally been explained in terms of two main variables. The first one, known since time immemorial, is the advantage of relative backwardness. The second one, which also has a relatively long history (e.g. Hill, 1964) is the importance of investment and, particularly, of investment in machinery and equipment (De Long and Summers, 1991). Indeed, these two variables seem to be the only ones that are consistently robust in the econometric literature, irrespective of sample size and period investigated, according to a recent thorough examination of the issue (Levine and Renelt, 1992). The evidence for Western Europe in the 1950s and 1960s confirms these findings. Initial productivity levels and investment ratios in machinery and equipment throw a good deal of light on the productivity growth experience of this period.2 Indeed, if the sample of countries is enlarged to encompass the United States and Japan, the estimated relationship improves further, with almost 98 per cent of the variance in productivity growth rates 'explained' by catch-up factors and investment efforts. It may be interesting to note that this seems not to have been the case in the interwar period when, admittedly on the basis of a very imperfect data set and of a smaller country sample, no relation whatsoever is found between these various variables. If the exchange rate matters, it should presumably have some impact on the investment ratio. To see whether this was the case, a set of nominal and real effective exchange rate indices was constructed from 1950 to 1970, using weights derived from the IMF's 'Multilateral Exchange Rate Model' (Artus and McGuirk, 1981) for all the eighteen OECD countries which that model covers. The effective exchange rate index that was derived considers the various bilateral nominal rates; the real one deflates these with wholesale prices.3 Unit labour costs would, no doubt, have been preferable, but these were not available for all the countries considered here. Equally, an earlier set of weights might have been more appropriate (the one used
112 Andrea Boltho
here, based on 1977 tradeflows,seems to give too much importance to, for instance, Japan), but again this seemed not to be available. Linking, in a very simple way, the investment share to an accelerator variable and to the real exchange rate index that had been derived gave totally insignificant results.4 At this level of aggregation, in other words, it would appear that exchange rate effects on investment, and hence on income convergence, cannot be detected. While this may, prima facie, seem surprising, it should not be forgotten, from the discussion in section 2 above, that the impact of longer-run exchange rate changes on investment and non-price competitiveness need not necessarily be favourable, so that a nil result for this sample of countries may not, after all, be that unexpected. A more appropriate procedure might thus be one that looked at single countries' experiences in somewhat greater depth. This has been done fairly exhaustively in the case of the United Kingdom. Studies of the 1967 parity change have, for instance, provided favourable assessments (for a survey, see Tew (1978)). Similarly, there is tentative evidence suggesting that the pound's sharp appreciation in 1978-81 had long-run negative effects on British exports (Bean, 1988), as may also have been the case for the uncompetitive exchange rate that the UK chose in 1925, at the time of its return to the pre-World War I gold parity. Yet, such conclusions have been questioned in a broader examination of the country's longer-run history of currency changes, which has suggested that postwar depreciations on balance worsened rather than improved non-price competitiveness (Thirlwall, 1980). Indeed, a more detailed econometric investigation has shown this to have been the case for the machine tools branch in the 1970s (Brech and Stout, 1981). The text that follows will eschew the conflicting evidence on the United Kingdom's experience and concentrate, instead, on some other countries that through the period under review showed very rapid convergence and either recorded low values of the exchange rate at the beginning, or consciously devalued their currencies at some time or other during the 'Golden Age'. An indication of where Western Europe's currencies may have stood at the outset is provided by Table 5.1, which shows various estimates of purchasing power parities vis-d-vis the United States in 1950 (following, therefore, the sharp devaluations of late 1949). Data are provided both on an absolute and on a relative basis (vis-a-vis 1938). While the latter year may not have been one of full payments equilibrium, a selective reading of the literature suggests that it may not have been too far from it either.5 The overwhelming impression the table conveys is that, vis-d-vis the United States at least, most European countries had acquired at the time a significant competitive price advantage (the major exceptions being Belgium, Switzerland and possibly France).6 Thisfindingstands in sharp contrast to the experience that followed World War I. The parities that were fixed in the 1920s, at different dates for different countries, seem often to have been above what a simple purchasing power parity rule might have suggested. Table 5.2 provides a tentative comparison of where real exchange rates may have stood in the mid-1920s and in 1950 relative to the prewar levels, on the basis of wholesale price developments. While, as mentioned above, virtually every Western European economy seemed competitive vis-d-vis the United States in 1950, very few looked similarly competitive a quarter of a century earlier. Moreover, activity in Western Europe in the early 1950s received a further boost
Convergence, competitiveness and the exchange rate 113 Table 5.1. Indicators of European competitiveness, 1950 Absolute competitiveness Cost of tradable basket0 (USA = 100) France Germany Italy United Kingdom Spain Austria Belgium Denmark Netherlands Norway Sweden Switzerland Mean
85 87 83 81 _ 89 83 76 82 83
Relative competitiveness (relative to USA) GDP
deflator 97 52 82 62 67 115 78 66 63 72 94 77
Wholesale Export Unit lab. prices prices costs (1938 = 100) 114 101 (80) 53 (70) 106 73 (87) 72 57* 72 112 85 75 58 77 100 78
79 71C 107 85 100 94 96 125 98
_ _ -
a
Obtained by averaging the United States and own country basket figures shown in the original source. b 1936 = 100. c 1937 = 100. Sources: Armstrong et al (1991); Gilbert and associates (1958); Triffin (1957); United Nations, Statistical Yearbook, 1954; United Nations Economic Commission for Europe, Economic Survey of Europe in 1951, Geneva, 1952, and Economic Survey of Europe Since the War, Geneva, 1953; US Department of Commerce, Long Term Economic Growth: 1860-1965, Washington, DC, 1966.
from the Korean War. Conversely, the achievement of afixedparity in the 1920s had been preceded (or accompanied) in a number of countries by domestic deflation aimed at reducing wages and prices to levels compatible with the new exchange rate. This had clearly been the case in Italy (Toniolo, 1980) and the UK (Lewis, 1949), and had probably also been true of the Netherlands, Norway, Sweden and Switzerland (Eichengreen, 1989). The debilitating effects of such deflations in some countries, combined with the lack of competitiveness in these and many others, may provide yet another reason for why Europe's growth experience after World War I was so much less successful than the one that followed World War II (Boltho, 1982b). Returning to the period under examination, the greatest degree of undervaluation in 1950 is that seen for Germany, an undervaluation that probably reflected the choice of a favourable exchange rate in spring 1948 (when Germany's economic situation still looked extremely precarious), and the subsequent devaluation of September 1949. This state of affairs hardly changed through the following decade, in view of the stability of the real exchange rate indicator shown in Figure 5.1. While
114 Andrea Boltho Table 5.2. Selected indicators of European competitiveness, 1920s and 1950 Relative to United States Wholesale prices Mid-1920sa (1913 = 100) France Germany Italy United Kingdom Spain Austria Belgium Denmark Finland Netherlands Norway Sweden Switzerland Czechoslovakia Hungary Mean
87 94 105 120 91 111 97 103 113 112 106 107 114 105
1950 (1938 = 100) 101 53 73 72 57* 72 112 85 — 75 58 77 100 78
a
The years chosen depend on when the country stabilized its exchange rate. For each country, the figure shown is an average of the real exchange rate in two years - the year of stabilization and the year immediately following it (e.g. in the French case, 1926 and 1927). b 1936 = 100. Sources: League of Nations, International Statistical Yearbook (various issues); Svennilson (1954); United Nations, Statistical Yearbook, 1954; United Nations Economic Commission for Europe, Economic Survey of Europe in 1951, Geneva, 1952; US Department of Commerce, Long Term Economic Growth: 1860-1965, Washington, DC, 1966.
Italy's starting advantage seems no more pronounced than that of the rest of Europe, Italy's real exchange rate dropped significantly in the early 1950s (Figure 5.2), and remained low throughout the decade. This reflected a very low rate of price inflation (wholesale prices fell by nearly 1 per cent per annum between 1951 and 1960, the sharpest decline in the OECD area), helped by both subdued wage developments and rapid productivity growth. In addition, two further rapidly growing countries (France and Spain) devalued in the late 1950s, in contrast to their experience in the 1930s. At the time, defence of the currency had been an obsessive aim of the authorities (Lewis, 1949; Fontana and Nadal, 1976). In the 1950s, on the other hand, such monetary orthodoxy was jettisoned when both countries opened their very protected economies to the forces of international trade. The competitive advantages that were acquired thanks to
Convergence, competitiveness and the exchange rate
95 1950
1955
1960
1965
115
1970
Source: Author's calculations using IMF data.
Figure 5.1 Effective and real exchange rate: Germany, 1950-70 (1950 = 100) 110 i
105 -
80 1950
1955
1960
1965
Source: Author's calculations using IMF data.
Figure 5.2 Effective and real exchange rate: Italy, 1950-70 (1950 = 100)
1970
116
Andrea Boltho
120 -
Effective ER Real ER
110 -
100
90 -
1950
1955
1960
1965
1970
Source: Author's calculations using IMF data.
Figure 5.3 Effective and real exchange rate: France 1950-70 (1950 = 100)
parity changes were then maintained, particularly in the French case, through most of the following decade (Figures 5.3 and 5.4). For these four countries, prolonged periods of low real exchange rates could have led to higher investment and structural transformation. Some inkling that this may have been the case is provided by a re-estimation of the simple econometric link between investment ratios, output growth and the exchange* rate, unsuccessfully tried above for the full sample of countries. Limiting coverage to France, Germany, Italy and Spain, it would now appear that exchange rate depreciation did raise investment shares.7 This is, however, far from conclusive proof of anything. Slight changes in periodization or in the definition of the real exchange rate variable could provide different outcomes. And even if the results were robust to somewhat different specifications, there is no insurance that the (relatively small) favourable effect that the exchange rate seems to have had on investment necessarily improved longer-run non-price competitiveness. Hence the need for a somewhat more detailed analysis of the four countries' experience. 4
Country experience
The present section looks at selected episodes of the postwar economic history of the four countries chosen above. In the German and Italian cases in the 1950s, the emphasis is not so much on nominal parity changes in 1949, since these were relatively modest,8 but on the low real exchange rate levels that seem to have prevailed in the decade. For France and Spain, on the other hand, the accent is on the sharp devaluations of the late 1950s.
Convergence, competitiveness and the exchange rate
601950
1955
1960
1965
117
1970
Source: Author's calculations using IMF data.
Figure 5.4 Effective and real exchange rate: Spain, 1950-70 (1950 = 100)
4.1
Germany and Italy
For Germany and Italy, despite obvious differences in maturity and industrial structure, the 1950s were in many ways a period of relatively similar 'miracles'. Both countries saw rapid output growth, rising investment propensities and sharp increases (particularly in Germany) in their shares of the world market for manufactured goods. And for both countries the explanations provided for this 'success story' are broadly similar - the combination of an elastic supply of labour and of cheap imports of American technology generated relatively high profits; these, in turn, stimulated investment and productivity growth. The interesting issue in the present context is to see to what extent external developments contributed to this virtuous circle. Despite the evidence of Table 5.1, there seems little acknowledgement in the literature that Germany's exchange rate level in 1950 was particularly favourable. Indeed, the Bizonal authorities apparently thought that the rate which had been agreed upon in 1948 was too high (Erhard, 1954). Despite the subsequent devaluation of 1949, there is a hint in the literature that suggests the presence of an overvalued exchange rate in the 1950s (Giersch et a/., 1992: 95). Yet, the evidence of mounting trade surpluses through that decade later convinced many that the currency had been clearly and increasingly undervalued (Gatz, 1963; Borchardt, 1967; Hardach, 1976). And such a conclusion is indirectly corroborated by the numerous interpretations of the period which suggest that both growth and cyclical upswings in these years were export-led (Michalski, 1970; Hennings, 1982; Giersch et al, 1992). This is even truer for Italy. Here, too, there seems to be little reference to the appropriateness or not of the 1950s nominal exchange rate, but there is virtual
118 Andrea Boltho
unanimity in acknowledging that the real exchange rate fell significantly in the 1950s, thereby stimulating exports (D'Antonio, 1973; Fua, 1981). And this has led a number of authors to argue that Italy too benefited from an export-led growth mechanism (e.g. Stern, 1967; Graziani, 1979; for a dissenting voice, see Rey, 1982). Yet, neither country was subject to revaluation pressures through the 1950s, in contrast to the many such pressures that were exerted in the 1920s, particularly in the Italian case, to restore parities to an overvalued rate. The stance of demand management policies may have helped. Fiscal policy was relatively orthodox, but there seems to have been little pressure to follow tight monetary policies. While both the Bundesbank and the Bank of Italy were, of course, concerned with inflation, modest wage rises, consequent upon the already mentioned elastic labour supplies, allowed a monetary stance that can be described as accommodating. The average annual growth rates of broad money supply from 1952 to 1960 were 16.6 per cent in Germany and 14.3 per cent in Italy (Budd and Dicks, 1982), as against rises in nominal GDP of only 9.3 and 8.2 per cent per annum respectively. Looking at overall performance in the 1950s, it is likely that in both countries a low exchange rate was an important ingredient in the virtuous circle of strong competitiveness, high investment, rapid productivity growth and scale economies, leading to even stronger competitiveness and more confident expectations. By the early 1960s, however, this price-cost advantage had begun to wane (Figures 5.1 and 5.2). In Germany it was appreciation of the currency, as well as a tighter labour market, that eroded some of the economy's external competitiveness; in Italy it was the effects of overheating in 1962-3. In both countries growth decelerated and inroads on foreign markets slowed down (Table 5.3). Arguably, however, the advantages conferred by earlier real depreciation had become permanent. After all, export market shares were not lost, and German and Italian competitiveness remained strong. To see whether these impressionistic observations can be confirmed, Table 5.4 presents indices of'revealed comparative advantage' (RCA) for the two countries in a number of product groups at the outset and end of the period under review.9 Data are shown for several broad SITC categories and for a subset of goods called 'growth commodities', which consists of twelve products (chosen at the three-digit SITC level) of the chemical, machinery, and transport and equipment industries whose share in OEEC exports grew fastest in the years 1954-61 (NIESR, 1963).10 Evidence of increasing RCAs in this group, as well as in the broader SITC categories 5 and 7, would presumably suggest that the two countries had moved 'up-market' in the period, thus making subsequent real appreciation easier to resist. As they stand, the changes between the 1950s and the 1960s show a clear improvement in Italy's RCAs from a relatively low starting point (particularly in the machinery branch) and a consolidation by Germany in a number of growth sectors. For both countries the impressionistic verdict presented above would seem to be confirmed - non-price competitiveness is likely to have improved in a period in which price competitiveness was strong. Yet, even this relatively innocuous conclusion need not be fully warranted. For one thing, the benefits of a low real exchange rate have been subject to fairly radical criticisms in both countries - whatever advantages export-led growth may have bestowed in terms of output, employment and the balance of payments, it also led to
Convergence, competitiveness and the exchange rate
119
Table 5.3. Changes in selected macroeconomic indicators: Germany and Italy, selected dates (%) Germany
G D P growth 0 Manufactured exports share* Machinery equipment investment share c Profit share: manufacturing sector0" Profit share: corporate sector0*
Italy
1951-2
1960-1
1951-2
1961-2
to
to
to
1960-1
1969-70
1960-1
to 1969-70
7.5 8.8
4.5 0.8
6.0 1.8
5.8 1.9
2.0 —9.9* -3.2e
2.0 —6.9 -6.4
2.0 — 4.0* 4.8 e
-0.9 —2.1 -0.1
0
Average annual percentage charges. In the exports of the eleven largest industrialized exporters. c In constant price G D P . ^In value added. * 1954 to 1960-1. Sources: Armstrong et al. (1991); M a d d i s o n (1991); N I E S R , Economic b
Review
(various issues); OECD, National Accounts of OECD Countries, 1950-68, 1953-69 and 1960-71. Table 5.4. Indices of revealed comparative advantage: Germany and Italy, selected dates Germany Primary products (0-4) Manufactures (5-8) of which: Chemicals (5) Machinery, etc. (7) 12 'growth products'0
Italy
1951-2
1960-1
1951-2
1960-1
0.51 1.29
0.37 1.28
0.93 1.06
0.87 1.07
1.64 1.32 1.60
1.38 1.45 1.43
0.92 0.76 0.83
0.93 0.95 0.91
0
For composition, see Statistical Appendix. Note: Figures in brackets refer to SITC categories. Sources: OEEC, Statistical Bulletins - Foreign Trade (Series II, III and IV).
a distortion in resource allocation whose longer-term structural effects were unfavourable. Thus, in Germany, it has been argued that the low exchange rate led to an 'oversized manufacturing sector' (Giersch et ai, 1992: 222). In the longer run this was untenable, and the country was forced into too rapid a phase of deindustrialization once the currency rose in the 1970s. Fluctuations in output shares, in other words, were magnified with, arguably, negative consequences for welfare. As for Italy, a low exchange rate favoured the manufacturing North at the expense of the more agricultural South. By exacerbating industrial and regional
120 Andrea Boltho
dualism, this paved the way for some of the difficulties the country experienced in the later 1960s (Graziani, 1979). Be this as it may, it can also be plausibly argued that the two countries' transformations owed more to forces other than low real exchange rates. In Germany's case, for instance, the view has been put forward that the weight of history mattered most - Germany's industrial structure had already, before the war, been heavily concentrated on investment goods. Events in the 1950s may have merely represented a return to earlier patterns of production and penetration of world markets (Altvater et al, 1980). More importantly, perhaps, the period of low real exchange rates coincided in both countries with a period of significant trade liberalization. Italy, for instance, eliminated its quantitative, trade barriers very rapidly - thus, by 1952 it was, with Portugal, the only OEEC country to have fully liberalized intra-European trade (Table 5.5). Liberalization vis-a-vis the dollar area was somewhat slower and Italy's tariffs remained high in the decade, but the Messina treaty, signed in 1955, was a clear signal of impending tariff dismantling. Germany, similarly, had lifted quota restrictions from 90 per cent of its intra-European trade by 1953 and of its dollar-area trade by 1956, well ahead of the rest of Europe. In addition, it was also aggressively engaged in unilateral tariff reductions through most of the 1950s (Giersch et al, 1992). That this opening to international trade may have mattered seems indirectly confirmed by the evolution of profitability in both countries (Table 5.3). A striking feature of the period is the less favourable movement of the profit share in the manufacturing sector if compared to that of the corporate sector as a whole, despite the currency's undervaluation. Overall profit shares decline sharply in Germany in the 1950s, but the fall in the open industrial sector is well above that of the more sheltered rest of the economy, in marked contrast to the relatively uniform trends of the 1960s. Similarly in Italy, the manufacturing sector's profitability is significantly weaker in the 1950s relative both to the profitability of the overall corporate sector and to developments in the 1960s. In other words, the improvements in the two countries' international competitiveness that have been noted above may, after all, have been prompted just as much, if not more, by the coming down of protectionist barriers as by the low real exchange rates of the period. Indeed, a number of observers have argued that the spur to modernization provided by trade liberalization was a powerful stimulant to investment and competitiveness in both Germany and Italy (e.g. Giersch et al, 1992; Ackley, 1979). 4.2
France and Spain
Not unlike Germany and Italy in the 1950s, France and Spain in the 1960s also shared some similarities. France saw a 'profound renovation' (Fohlen, 1976: 114), while Spain witnessed 'true industrialisation' (Fontana and Nadal, 1976: 522). In both countries, growth was more rapid than in the previous decade, there were gains in market shares abroad, and ratios of investment to output rose, when they could be expected to fall, at least in France (Carre et al, 1982) (Table 5.6). Yet, in contrast
Convergence, competitiveness and the exchange rate
121
Table 5.5. Trade liberalization in OEEC countries, 1952-8 (% of imports freed from quota restrictions; end years)
Intra-OEEC trade Germany France Italy OEEC average Trade with dollar area Germany France Italy OEEC average
1952
1954
1956
1958
81 0 100 65
90 65 100 83
92 82 98 89
91 90 98 89
a
90 11 39 61
85 51 68 72
0 0 0 11
54
0° 24* 44*
a
End-September 1954. Sources: OEEC, Annual Reports of the OEEC (9th, 10th, 12th).
Table 5.6. Changes in selected macroeconomic indicators: France and Spain, selected dates (%) France 1950-1 to 1957-8 G D P growth 0 Manufactured exports share 6 Machinery and equipment investment sharec Profit share: manufacturing sector** Profit share: corporate sector^ a
Spain 1957-8 to 1967-8
1954 to 1957-8
1957-8 to 1967-8
4.5 -1.7
5.2 0.3
5.8 -
6.2 0.3
1.2
2.1
1.5
4.7
-6.2'
-1.1 2.7/
-0.6/
-4.6e
2.3
Average annual percentage changes. In the exports of the eleven largest industrialized countries. c In constant price GDP. **In value added. e 1954 to 1960-1. /Inverse of adjusted wage share in GDP. Sources: Armstrong et al. (1991); EEC, European Economy, no. 54, 1993; Maddison (1991); NIESR, Economic Review (various issues); OECD, National Accounts of OECD Countries, 1950-68 and 1953-69; UN, Monthly Bulletin of Statistics (various issues). b
122 Andrea Boltho Table 5.7. Indices of revealed comparative advantage: France and Spain, selected dates Spain
France Primary products (0-4) Manufactures (5-8) of which: Chemicals (5) Machinery etc. (7) 12 'growth products'0
1957-8
1967-8
1961
1967-8
0.91 1.02
1.04 1.00
2.40 0.43
2.16 0.64
1.12 0.73 0.61
1.18 0.83 0.84
0.62 0.12 0.18
0.67 0.39 0.34
fl
For composition, see Statistical Appendix. Note: Figures in brackets refer to SITC categories. Sources: OEEC, Statistical Bulletins - Foreign Trade (Series II, III and IV); OECD, Foreign Trade Statistics (Series B and C).
to the German and Italian experiences, these developments seem to have owed less to elastic labour supplies and subdued wage developments. Low population growth in France and emigration in Spain meant that the industrial areas of the two countries were experiencing conditions approaching full employment (Merigo, 1982; Sautter, 1982). It is tempting, therefore, to advance the devaluations of 1957-8 in France and of 1959 in Spain as possible important causes of subsequent developments. Both devaluations were triggered off by earlier inflationary developments, both were accompanied by sharp absorption-reducing policies, and both led very quickly to current account improvements and pronounced rises in investment. It is true, however, that both countries found themselves forced to devalue again towards the end of the decade. Spain followed the UK in October 1967, after its mid-1960s boom had led to generalized overheating, while France devalued in 1969, not long after the May 1968 events. Though neither of these two inflationary episodes can be directly linked to the earlier devaluations, they still raise the issue of how temporary were the advantages that these parity changes conferred. For France, the literature on the devaluation is, on the whole, sceptical about its favourable impact. Some argue that price advantages had already been lost by 1964 (Carre et al., 1982), something that Figure 5.3, however, does not fully confirm. Others point, instead, to the lack of a strong investment response in the tradable sector (Mistral, 1975). All lament (in predictable French fashion), the continuing presence of an unfavourable taux de couverture in manufactured goods (i.e. a trade deficit), often forgetting that this was associated with an acceleration in output growth. For Spain, the judgement would seem to be that the immediate response to devaluation had been favourable (OECD, 1966; Martinez Serrano et al., 1982), but that subsequent demand-pull inflation undid much of the initial gains (Anderson, 1970). In an attempt to go beyond, Table 5.7 replicates the RCA information already provided above for Germany and Italy.11 It is interesting to note how relatively underdeveloped the export structures of both countries were at the outset. France's
Convergence, competitiveness and the exchange rate
123
Table 5.8. Indicators offoreign trade changes: France and Spain, selected dates (%) 1950-1
1957-8
1967-8
38.0 17.5 1.9U
35.9 23.6 2.6
13.4 42.2 5.9
(18.6) (20.3) -
13.7 29.0 3Ab
12.5 30.2 6.7
France
Share of exports to franc area Share of exports to EEC 6 Share of manufactured imports in GDP Spain Share and Share Share
of exports to African possessions Latin America of exports to E E C 6 of manufactured imports in G D P
a
1954. M961. Sources: OECD, Statistical Bulletins - Foreign Trade (Series I and IV), Foreign Trade Statistics (Series A and B) and National Accounts of OECD Countries, 1950-58 and 1953-69; Annuario Estadistico de Espana (various issues).
RCAs in both SITC 7 and the twelve 'growth products' were, in 1957-8, below those of Italy in 1951-2. As for Spain, its manufactured exports were virtually absent from world markets. Over time, there is a clear improvement in both countries' RCAs in what can be called the 'growth areas' (in France's case there is also an improvement for primary products, totally concentrated, of course, in agriculture). In fact, these figures probably underestimate the transformation of the two countries' tradable sectors. Spain was able to make its first entry on the world markets for manufactures, strongly helped in this by an overflow of foreign direct investment, while France radically changed the geographic structure of its exports. Since the Great Depression, the country had concentrated its sales abroad on the protected markets of the franc area (Table 5.8). In the space often years, it succeeded in massively reorienting its exports towards the EEC. Even if some felt that the changes in France's RCAs were not sufficient (INSEE, 1974), the data shown in Table 5.8 would seem to provide a very clear indication of a structural improvement in the country's competitiveness. Yet, devaluation may not have been the only cause of these developments, since France in 1958 and Spain in 1959 also saw massive programmes of market opening. Indeed, it is likely that the devaluations were an integral part of the trade liberalization packages - industry, as well as labour, would probably have baulked at the prospects of a sharp decline in protectionism in the absence of some compensating measures. France, which had been slow in implementing trade liberalization until then (Table 5.5), now entered the Common Market and dismantled all its tariffs vis-a-vis the EEC countries between January 1959 and July 1968. As for Spain, it abandoned the virtually autarkic model of development that it had followed since the early 1940s. Before 1959, up to 90 per cent of imports had been subject to quantitative restrictions. This share had fallen to 30 per cent by 1966 (Wright, 1977), although
124 Andrea Boltho
effective tariff protection remained high in the decade (Martinez Serrano et a/., 1982). And as in Germany and Italy, manufacturing profits in France fared much less well in this decade than in the 1950s relative to total corporate sector profits. For Spain a very rough indicator of overall profits also suggests a worsening. Similarly to Germany and Italy, therefore, changes in international competitiveness in France and Spain coincided with, and may well have been prompted by, the opening to foreign competition, as indeed is argued by a number of writers (Carre et al., 1982; Lieberman, 1982). Two relatively protected economies abandoned their semi-autarkic ways, allowed much greater import penetration, and at the same time were able to improve the commodity and country composition of their exports. Whether the primary stimulus of this conversion came from devaluation of the currency or from the opening to free trade is unresolved. Yet, a glance at Figures 5.3 and 5.4 suggests that liberalization may have played a greater role. By the mid-1960s France's real exchange rate (while well below 1957) had returned to its (not very competitive) 1950 starting point. As for Spain, the peseta's real value had virtually jumped back to its (equally non-competitive) pre-devaluation levels. 5
Conclusions
The preceding pages have explored the possibility that the rapid growth experienced by several Western European countries in the 1950s and 1960s may have been partly linked to their exchange rate developments. In particular, the chapter has tried to investigate, in an admittedly very informal and ad hoc way, the hypothesis that devaluation of the currency, in addition to conferring well-known, but ultimately short-lived, cost and price advantages, may also improve longer-term non-price competitiveness. The evidence provided for Germany and Italy in the 1950s, and France and Spain in the 1960s, predictably perhaps, does not really allow a clear-cut answer to this question. All the four countries surveyed did benefit from relatively favourable exchange rates through some of the period - Germany thanks mainly to the postwar fixing of the Deutschmark at what turned out to be a very low level; Italy in the wake of subdued domestic wage and price developments; France and Spain following deliberate and relatively large devaluations of their currencies. All four countries also experienced rapid output and investment growth as well as rising export market shares in the period during which their real exchange rates remained low. More tellingly, all four countries also went through significant structural transformations. Changes in indices of revealed comparative advantage for products that experienced very rapid growth on world markets suggest, for instance, that improvements in non-price competitiveness were widespread, particularly in Italy. Equally strikingly, France was able sharply to restructure the geographic composition of its sales abroad, away from largely protected and slow-growing overseas markets to the much more dynamic and competitive Western European outlets. All this may suggest that, by the time the advantages of low real exchange rates had been eroded (be it by appreciation in Germany or by accelerating wage and price inflation in France, Italy and Spain), these countries found themselves in a stronger position than they would otherwise have achieved. A low real exchange rate, in other words, may have had more than merely transitory effects.
Convergence, competitiveness and the exchange rate
125
Yet, however plausible such a conclusion may sound, it need not be fully warranted. One important feature of the four countries' experiences at the time at which they gained high competitiveness was an extensive programme of trade liberalization. For Germany and Italy, this coincided with the general freeing of intra-OEEC exchanges in the 1950s, but both Germany and Italy seem to have gone faster down the liberalization road than other countries. For France and Spain, liberalization was an essential component of the late 1950s devaluation packages, and led to a sharp opening in economies that, until then, had followed highly protectionistic or even semi-autarkic policies. In standard economic theory, trade liberalization would hardly be expected to generate longer-run growth effects. Freer trade leads to well-known, but merely static gains in welfare. When translated into GDP equivalents, these often turn out to be pitifully small - not quite 0.2 per cent of EEC GDP, for instance, following the completion of the Common Market (Balassa, 1974). Yet, just as devaluation may have more than a transitory impact, so too could trade liberalization. Thus, estimates which allow for scale economies and increased competitive pressures suggest that the formation of the EEC may have raised Community GDP by 3 to 6 percentage points (Owen, 1983). And the new 'growth economics' states that further, and even greater, gains are possible because of induced higher investment (Baldwin, 1989). It would seem plausible, therefore, to argue that both devaluation and trade liberalization can provide longer-term favourable effects on economic growth. The 'warm blanket of protection' which a devaluation provides may stimulate profits and investment, rather than retard structural change. Similarly, the 'cold shower of competition', rather than hurting investment through its squeeze on profits, may raise efficiency and force firms to step up investment. Most probably, in the circumstances of the time, the two favourable effects worked hand in hand - rising external pressures dictated the need for structural change,while low real exchange rates provided the funds with which this structural change could be financed. If this is so, some conclusions can perhaps be drawn. First, a combination of currency devaluation and trade liberalization, such as occurred in the four countries examined here, may, in many ways, be optimal. At the political level, it facilitates the acceptance of market-opening packages (as suggested by the French and Spanish examples of the late 1950s). Without devaluation it may be very difficult to persuade industrial and trade union lobbies to accept a lowering of protective barriers.12 And at the economic level, a joint package supplies three crucial ingredients for the success of the tradable sector: increased competition (which devaluation alone does not provide), increased profits (which liberalization alone does not provide), and increased opportunities both at home and abroad.13 Second, the (partial) successes which the exchange rate instrument may have had in ensuring structural change were perhaps limited to the particular period under consideration. One reason for this is to be found in the international monetary and trading conditions of the time. These allowed for occasional parity changes, in contrast to much of the interwar period and to the post-1973 world, when floating was more prevalent. In addition, they also made for a rapid process of trade liberalization that was very different from the trade wars that prevailed in the 1930s, and even from the 'creeping protectionism' of the post-1973 years. Finally, United
126 Andrea Boltho
States hegemony in the 'Golden Age' allowed a form of free-riding behaviour in both Europe and Japan. Undervaluation of major currencies would not have been possible for long in the interwar years and is similarly unlikely today. It was feasible in the very special conditions of the Bretton Woods system, given the 'benign neglect' with which the United States viewed its balance of payments developments. These two conclusions, in turn, have an important implication for policy. Devaluing so as to ensure export-led growth may have been a viable strategy provided it was accompanied by market-opening measures and taken in an environment of quasi-fixed exchange rates. In today's Europe, however, most trade barriers (outside agriculture) have fallen to very low levels, and efforts at trade liberalization, such as those of the 1992 Programme or of the Uruguay Round, are much smaller in scope. Similarly, greater currency flexibility is likely to have diminished the investment response to what are now often seen as merely transitory changes in real exchange rates. In other words, longer-run effects on competitiveness stemming from temporary exchange rate changes (or from radical market opening) may be much less likely at present than they were in the 'Golden Age'.14 Devaluations may still, of course, be indispensable to restore purchasing power parity following nominal shocks. They can also provide a welcome shot in the arm (as they seem to have done in 1993 in both Italy and the UK). They may even lead to prolonged periods of price or cost competitiveness, as suggested by Belgian experience since the franc's downward realignment of 1982, but their effects in altering non-price competitiveness may be much less secure.
Statistical Appendix Data for regressions: 1950 and 1970 GDP and employment data for twelve European countries were obtained from Maddison (1991); 1950 GDP data for Spain and Ireland came from Maddison (1989) and were linked to OECD statistics (from OECD, National Accounts, 1960-1990) to obtainfiguresfor 1970. Spanish and Irish employment data for 1950 and 1970 came from OECD, Manpower Statistics and Labour Force Statistics (various issues). Productivity growth was derived by dividing GDP by employment growth. Productivity levels in 1950 for the United States and twelve European countries came from Maddison (1991); thefiguresfor Spain and Ireland were very tentatively obtained by regressing Bairoch's (1976) estimates of per-capita GDP on to Maddison's data for per-capita productivity for twelve European countries, and applying the resulting coefficients to Bairoch's Spanish and Irish data. The 1950-69 GDP shares of investment in machinery and equipment (at constant 1963 prices) came from the OECD's National Accounts of OECD Countries, 1950-1968 and 1953-1969; for Italy and Norway thefiguresbegin in 1951, for Belgium in 1953, and for Spain in 1954. Definition of'growth products': the twelve 'growth products' are SITC 266 (synthetic fibres), SITC 512 (organic chemicals), SITC 521 (crude coal and petroleum chemicals), SITC 581-99 (plastic and chemical materials), SITC 711 (non-electric power generating machines), SITC 714 (office machines), SITC 715 (metal working machines), SITC 717-19 (other industrial machines), SITC 722-9 (electric machines), SITC 734 (aircraft), SITC 861 (scientific, optical and photographic instruments) and SITC 891 (musical instruments and sound recorders).
Convergence, competitiveness and the exchange rate
127
NOTES 1 This assumes little change in the terms of trade - an assumption that seems broadly acceptable in the Western European context until the oil shocks of the 1970s. 2 Thus, linking the growth rate of productivity over the two decades (ChPr) to initial productivity levels relative of those of the United States (Pr/Prus) and to the share of machinery and equipment investment in output (lme/Y) gives the following result for a sample of thirteen European countries (f-ratios in brackets): ChPr = 5.0 - 6.9Pr/Prus + 0.21J Y (5.5) (8.0) (2.8)
R2 = 0.90 S£ = 0.35
(1)
The data sources for the fourteen countries initially selected are provided in the Statistical Appendix. Norway was, however, excluded from the regression. The country is an outlier with respect to its investment share in output, which (at 14.5 per cent of GDP) is some 50 per cent above that of the sample's mean - largely a reflection of the highly capital-intensive nature of some of its industry (e.g. aluminium and shipping). 3 The 1950 to 1970 data for the nominal exchange rate and for wholesale prices come from the IMF's International Financial Statistics Yearbook (various issues). 4 The output share of investment in machinery and equipment was made dependent on a simple accelerator variable (the change in GDP over the two decades, or ChY) and on various indicators of the real exchange rate (e.g. its average cumulative deviation over the period from its 1950 level, or DvRE). An example of the type of result obtained is given by the following: IJY
= 5.8 + 0.1 ChY + O.ODvRE (3.6) (2.2) (0.4)
R2 = 0.33 SE = 1.20
(2)
5 Thus, according to Nurkse, 'At the end of 1936 . . . exchange relationships between the principal free currencies were not widely different from what they had been in 1930' (League of Nations, 1944: 129), while Kindleberger suggests that the 1937 exchange rates were 'broadly back to the 1929 pattern' (1986:269). This pattern, in turn, bar the overvaluation of the pound, seemed a reasonably stable one (Lewis, 1949). 6 This is most obvious for the wholesale price indicator and least obvious for the export price one or for the absolute measure. Arguably, however, these last two indicators are less significant, since in 1950 most of Europe's economies were price-takers on the world markets. Hence, they would have set their tradables' prices at, or close to, United States levels. Unit labour cost data would have been more appropriate, but unfortunately they are not available for most countries. Even the indices shown in Table 5.1 (for three countries only) must be considered as very tentative. 7 Investment shares were now calculated for cyclically comparable subperiods (four for France and Germany, three for Italy and Spain), and linked again to output growth rates, lagged one year, to cumulative real exchange rate deviations from their level at the outset of each subperiod, and to a (debatable) time variable (t) that could proxy, inter alia, for a trend towards increasing capital-output ratios or even for the perceived lower riskiness of investment as growth expectations firmed between the 1950s and 1960s: IJY
= 2.6 + 0.6ChYt_] - O.lDvRE 4- 1.3* (1.7) (2.8) (2.1) (5.6)
R2 = 0.79 S£ = 0.91
(3)
The subperiods were chosen to cover full cycles, with the cyclical peaks taken
128 Andrea Boltho
8 9
10 11 12
13 14
from OECD (1973). Initial and terminal years are as follows: France, 1951-7, 1957-61,1961-6, 1966-70; Germany, 1950-5, 1955-60,1960-5, 1965-70; Italy, 1951-5, 1955-62, 1962-9; Spain, 1954-8, 1988-66, 1966-70. While most of Europe devalued by 30 per cent vis-d-vis the dollar, Germany devalued by just over 20 per cent and Italy by less than 10 per cent. Indices of 'revealed comparative advantage' measure the ratio between a country's share of world exports in a particular product and its share of total world exports. The 'world' is here defined as the OEEC area (Canada, the United States and Western Europe, excluding Finland, Spain and Switzerland, for which early 1950s data are not available). For the definition of the twelve 'growth products', see the Statistical Appendix. The products represented 18.4 per cent of total OEEC exports in 1951-2 and 30.3 per cent in 1967-8. The initial year for Spain's RCA calculations is 1961, rather than 1957-8, for lack of comparable figures before that date. This is true even if, with the benefit of hindsight, it appears that the short-term costs that one would normally associate with trade liberalization were much smaller in Europe than feared, thanks to the unexpected growth in intra-industry trade. I am grateful to Stephen Nickell for this point. United States experience following the dollar's steep appreciation to 1985 would seem to confirm this conclusion. Despite much discussion of hysteresis effects, these receive little empirical support from the evidence (Krugman, 1991).
REFERENCES Ackley, G. (1979) 'Lo sviluppo economico dal 1951 al 1961', in A. Graziani (ed.), L'economia italiana dal 1945 a oggi, new edn, Bologna: II Mulino. Altvater, E., J. Hoffmann and W. Semmler (1980) Vom Wirtschaftswunder zur Wirtschaftskrise: Okonomie und Politik in der Bundesrepublik, Berlin: Verlag Olle & Wolter. Anderson, C.W. (1970) The Political Economy of Modern Spain, Maddison, WI: University of Wisconsin Press. Armstrong, P., A. Glyn and J. Harrison (1991) Capitalism since 1945, Oxford: Blackwell. Artus, J.R. and A.K. McGuirk (1981) 'A revised version of the Multilateral Exchange Rate Model', IMF Staff Papers, 28 (2), pp. 275-309. Bairoch, P. (1976) 'Europe's gross national product: 1800-1975', Journal of European Economic History, 5 (2), pp. 273-340. Balassa, B. (1974) Trade creation and trade diversion in the European Common Market: an appraisal of the evidence', Manchester School, 42 (2), pp. 99-135. Baldwin, R. (1988) 'Hysteresis in import prices: the beachhead effect', American Economic Review, 78 (4), pp. 773-85. (1989) 'The growth effects of 1992', Economic Policy, 9, pp. 247-81. Baldwin, R. and P. Krugman (1989) 'Persistent trade effects of large exchange rate shocks', Quarterly Journal of Economics, 104 (4), pp. 635-54. Bean, C. (1988) 'Sterling misalignment and British trade performance', in R.C. Marston (ed.), Misalignment of Exchange Rates: Effects on Trade and Industry, Chicago, IL: University of Chicago Press. Beckerman, W. (1962) 'Projecting Europe's growth', Economic Journal, 72 (288), pp. 912-25. Boltho, A. (ed.) (1982a) The European Economy: Growth and Crisis, Oxford: Oxford
Convergence, competitiveness and the exchange rate
129
University Press. (1982b) 'Growth', in Boltho (1982a). Borchardt, K. (1967) T h e Federal Republic of Germany', in G. Stolper (ed.), The German Economy, 1870 to the Present, London: Weidenfeld and Nicolson. Brech, M.J. and D.K. Stout (1981) The rate of exchange and non-price competitiveness: a provisional study with UK manufactured exports', Oxford Economic Papers, 33 (supplement), pp. 268-81. Brown, A.J. (1979) 'Inflation and the British sickness', Economic Journal, 89 (353), pp. 1-12. Budd, A. and G. Dicks (1982) Inflation - a monetarist interpretation, in Boltho (1982a). Carre, J.-J., P. Dubois and E. Malinvaud (1982) La croissance francaise: Un essai d'analyse economique causale de I'apres guerre, Paris: Le Seuil. Cipolla, C M . (ed.) (1976) The Fontana Economic History of Europe, vol. 6, Glasgow: Collins/Fontana. Cornwall, J. (1977) Modern Capitalism: Its Growth and Transformation, London: Martin Robertson. D'Antonio, M. (1973) Sviluppo e crisi del capitalismo italiano, 1951-1972, Bari: De Donato editore. De Long, J.B. and L.H. Summers (1991) 'Equipment investment and economic growth', Quarterly Journal of Economics, 106 (2), pp. 445-502. Eichengreen, B. (1989) 'International monetary instability between the wars: structural flaws or misguided policies?', CEPR Discussion Paper No. 348, November. Erhard, L. (1954) Germany's Comeback in World Markets, London: George Allen and Unwin. Fagerberg, J. (1988) 'International competitiveness', Economic Journal, 98 (391), pp. 355-74. Fohlen, C. (1976) 'France, 1920-1970', in Cipolla (1976). Fontana, J. and J. Nadal (1976) 'Spain, 1914^1970', in Cipolla (1976). Fua, G. (1981) 'II "miracolo economico", in una interpretazione dell'epoca', in G. Fua (ed.), Lo sviluppo economico in Italia, vol. I, Milan: Franco Angeli editore. Gatz, W. (1963) 'Grunde und volkswirtschaftliche Wirkungen der D-MarkAufwertung', Weltwirtschaftliches Archiv, 90 (2), pp. 379-432. Giersch, H., K.-H. Paque and H. Schmieding (1992) The Fading Miracle: Four Decades ofMarket Economy in Germany, Cambridge: Cambridge University Press. Gilbert, M. and associates (1958) Comparative National Products and Price Levels, Paris: OEEC. Graziani, A. (1979) 'Introduzione', in A. Graziani (ed.), L'economia italiana dal 1945 a oggi, new edn, Bologna: II Mulino. Hardach, K. (1976) 'Germany, 1914-1970', in Cipolla (1976). Hennings, K.H. (1982) 'West Germany', in Boltho (1982a). Hill, T.P. (1964) 'Growth and investment according to international comparisons, Economic Journal, 74 (294), pp. 287-304. INSEE (1974) Fresque historique du systeme productif, Collections de 1'INSEE, Serie E, no. 27. Kaldor, N. (1981) T h e role of increasing returns, technical progress and cumulative causation in the theory of international trade and economic growth', Economie appliquee, 34 (4), pp. 593-617. Kindleberger, C.P. (1986) The World in Depression, 1929-1939, Harmondsworth: Penguin. Krugman, P. (1989) 'Differences in income elasticities and trends in real exchange
130 Andrea Boltho
rates', European Economic Review, 33 (5), pp. 1031-46. (1991) Has the Adjustment Process Worked?, Washington, DC: Institute for International Economics. Krugman, P.R. and M. Obstfeld (1988) International Economics: Theory and Policy,
Glenview, IL: Scott, Foresman and Co. Lamfalussy, A. (1963) The United Kingdom and the Six: An Essay on Economic Growth in Western Europe, London: Macmillan. League of Nations (1944) International Currency Experience: Lessons of the
Inter-war Period, Princeton, NJ: Princeton University Press. Levine, R. and D. Renelt (1992) 'A sensitivity analysis of cross-country growth regressions', American Economic Review, 82 (4), pp. 942-62. Lewis, W. A. (1949) Economic Survey, 1919-1939, London: George Allen and Unwin. Lieberman, S. (1982) The Contemporary Spanish Economy: A Historical Perspective,
London: George Allen and Unwin. Maddison, A. (1989) The World Economy in the 20th Century, Paris: OECD. (1991) Dynamic Forces in Capitalist Development, Oxford: Oxford University Press.
Martinez Serrano, J.A., M. Mas Ivers, J. Paricio Torregrosa, F. Perez Garcia, J. Quesada Ibanez and E. Reig Martinez (1982) Economia espahola: 1960-1980 Crecimiento y cambio structural, Madrid: H. Blume ediciones.
Merigo, E. (1982) 'Spain', in Boltho (1982a). Michalski, W. (1970) Export und Wirtschaftswachstum, Hamburg: Verlag Weltarchiv. Mistral, J. (1975) 'Vingt ans de redeploiement ducommerce exterieur', Economie et statistique, no. 71, pp. 23-40. NIESR (1963) 'Fast and slow growing products in world trade', NIESR Review, no. 25, pp. 22-39. OECD (1966) Economic Survey of Spain, Paris.
(1973) The measurement of domestic cyclical fluctuations', OECD Economic Outlook - Occasional Studies, July, pp. 3-68. Owen, N. (1983) Economies of Scale, Competitiveness, and Trade Patterns within the
European Community, Oxford: Carendon Press. Rey, G.M. (1982) 'Italy', in Boltho (1982a). Sautter, C. (1982) Trance', in Boltho (1982a). Stern, R.M. (1967) Foreign Trade and Economic Growth in Italy, New York: Praeger. Svennilson, I. (1954) Growth and Stagnation in the European Economy, Geneva:
UNECE. Tew, J.H.B. (1978) 'Policies aimed at improving the balance of payments', in F.T. Blackaby (ed.), British Economic Policy, 1960-74: Demand Management,
Cambridge: Cambridge University Press. Thirlwall, A.P. (1980) Balance-of-Payments Theory and the United Kingdom
Experience, London: Macmillan. Toniolo, G. (1980) L'economia dell'Italia fascista, Bari: Laterza.
Triffin, R. (1957) Europe and the Money Muddle, New Haven, CT: Yale University Press. Wright, A. (1977) The Spanish Economy, 1959-1976, London: Macmillan.
6
British economic growth since 1945: relative economic decline . . . and renaissance? CHARLES BEAN AND NICHOLAS CRAFTS
1
Introduction
It is a commonplace that the growth and productivity performance of the British economy has been disappointing in the postwar period, at least until the 1980s, which can be regarded as still controversial. Explanations for this 'failure' abound in the literature, as Alford's (1988) introductory essay well reflects. It will be useful at the outset to distinguish between three categories of analysis to which we can return in the concluding section, after the details of the historical record have been reviewed. First, we can look at proximate sources of growth in the manner of growth accounting, which shows both low investment and low productivity growth in the UK compared with its peer group. Second, we can explore the correlates of these aspects of growth performance in terms of their association with particular forms of institutions or policies, such as vocational training systems and ill-conceived industrial policy. Third, and most difficult, we can consider the fundamental factors which sustain the relatively low growth path as an equilibrium. Here we need to consider the incentive structures facing decision-makers both in business and in government, which produced market failure and/or precluded effective supply-side policy responses. In exploring the UK's growth performance we can take advantage of insights from recent work in growth economics. In particular, it will be useful to bear in mind the following points. First, new models of growth have tended to reassert the importance of capital accumulation and have, in effect, proposed various ways of endogenizing technical progress and productivity change. Second, broad capital is now seen as being at the heart of the growth process, with a strong emphasis on the roles of human capital formation and research and development. Third, investment decisions must be thought of in terms of optimizing decisions; taxes and institutional arrangements influencing the ability to appropriate quasi-rents will affect investment and, thus, growth outcomes. Fourth, comparative growth performance across countries may be influenced by the catching up of leading countries by followers, as obstacles to technology transfer are removed in a world where catching-up is not automatic. 131
132 Charles Bean and Nicholas Crafts
The problem of relative economic decline dominates the literature on postwar British economic growth and, from at least the late 1950s, has been a constant preoccupation for policy-makers. Both the right- and left-wing policy recommendations drawn from new growth theory have their echo in all this discussion. British industrial policy from the mid-1950s to the early 1970s has been described as dominated by attempts to promote investment-led growth (Morris and Stout, 1985), while the Thatcher reforms of the 1980s sought to lower direct taxes and to support an enterprise culture. Academic analyses have for a long time found fault with policy in ways which anticipate some of the key themes of recent theorizing on growth. The famous broad-brush approaches of Bacon and Eltis (1976) and Kaldor (1966) exemplify this point. The former proposed an explanation for slow growth in the UK based on an excessive expansion of the non-marketed sector, leading to high taxes and to the detriment of investment, while the latter stressed the need for policies to switch activity from services to manufacturing, to reap the benefits of learning-by-doing effects. At a more detailed level, Tanzi (1969) is representative of a large group condemning high marginal income tax rates as a significant impediment to growth, while Prais (1981) marks the beginning of a concerted critique on the deficiencies of human capital formation in the UK. It is obviously important to bear in mind that the UK had less scope for rapid catch-up or reconstruction-based growth than most European countries in the early postwar period. Even so, it seems clear that the UK was less successful in exploiting what opportunities there were for catch-up, and attention therefore needs to be directed at Abramovitz's (1986) emphasis on the role of 'social capability' in this context. Here, particular importance attaches to industrial relations, competition policy and capital markets. In each of these, the UK exhibited significant idiosyncrasies which had important effects on the environment in which British company management - perhaps the most common scapegoat^of all - operated. In this chapter, for example, we consider both short- and long-run implications of bargaining between firms and their workers, which took place for most of the postwar period in a traditional, unreformed context. It is well recognized that productivity levels are one of the outcomes of such bargaining, and that circumstances such as weak competition in product markets, low unemployment and nonencompassing unions can lead to low productivity equilibria, and this implies a relatively low steady-state income level at the end of a catching-up process. In a context of new growth theory, however, it should be recognized that firm-worker bargaining arrangements can have long-run growth rate effects through their influence on the accumulation of broad capital (which will usually be measured in practice as an impact reflected in total factor productivity growth). To a substantial extent, the framework within whichfirmsand workers bargain is determined by government either directly or indirectly. Here it should be noted that, in the postwar UK, governments have traditionally looked to short-run macroeconomic performance as the economic underpinning to their re-election chances, rather than to productivity or long-run growth outcomes. Price and Sanders (1994: 302-3) suggest that this heavy concern with unemployment and inflation in the short term was perfectly rational; their econometric estimates show substantial effects of these variables on electoral popularity. This tended to push governments
British economic growth since 1945 133 Table 6.1. Growth of real output per head and productivity, 1950-89 (% per year)
1950-73 GDP/head Labour productivity TFP in business sector 1973-9 GDP/head Labour productivity TFP in business sector 1979-89 GDP/head Labour productivity TFP in business sector
UK
12-country median
2.4 2.5 2.3
3.4 3.6 3.0
1.5 1.3 0.6
2.0 2.2 1.4
2.1 1.7 1.5
1.9 1.6 1.2
Sources: Maddison (1991) and OECD (1991b). Labour productivity is measured using persons employed, and TFP estimates are for 1960-73 not 1950-73. towards formal or informal incomes policies and 'social contracts' which both precluded reform of industrial relations and lowered the competitive pressures on firms. The important implication of recent work in growth economics is that this policy preference may have had a long-run growth rate effect, not simply reduced the level of productivity. Table 6.1 examines the raw data of British growth in a comparative context. Relatively slow growth of both output and productivity through to the end of the 1970s shows up prominently, while relative but not absolute performance looks rather better in the 1980s, as other countries' growth slowed down. Although growth was disappointing by international standards in the Golden Age of 1950-73, it was most respectable relative to the UK's own historical performance. In 1873-1913 and 1924-37, real GDP grew at 1.8 and 2.3 per cent, with TFP growth at 0.5 and 0.7 per cent respectively (see Table 6.8 below). Table 6.2 reports purchasing power parity adjusted estimates of relative productivity levels. This makes more explicit economic decline relative to France and Germany, both of which were more successful in catching up the USA. Indeed, British productivity was overtaken by each of these countries during the Golden Age, and an increasing gap developed during the 1970s, especially in manufacturing. The 1980s, by contrast, saw a notable reduction in the UK's manufacturing productivity gap. Growth accounting exercises involving comparisons of the UK with France and Germany are reported in Table 6.3. Both Denison and Maddison made their estimates within a framework derived from traditional growth theory. Taken at face value, there are strong messages in Table 6.3. According to Maddison, during 1973-87 the sources of growth in each economy were brqadly similar, whereas during 1950-73 both capital input and, much more importantly, TFP growth were less strong in the UK.
134
Charles Bean and Nicholas Crafts
Table 6.2. Estimates of comparative levels of productivity, 1950-89 (UK =100) (a) Comparisons of real GDP/person employed France Germany 1950 1960 1973 1987
69.7 88.6 110.2 116.4
63.3 90.2 104.7 105.6
USA 167.4 167.5 151.6 128.9
(b) Comparisons of real value added/hour worked in manufacturing France Germany USA 1950 1960 1973 1979 1989
76.7 90.3 113.3 134.1 120.4
88.8 115.7 126.8 151.0 117.4
250.0 226.2 190.8 186.2 165.3
Sources: Maddison (1991); van Ark (1993).
In turn, this was only partly based on greater initial backwardness, and a major weakness in the UK is recorded as lower improvement in residual efficiency. Very similar results were obtained earlier by Denison in his classic study; indeed a major theme of Denison's paper was the substantial shortfall in UK productivity levels in 1960, coming from differences in work effort, restrictive labour practices and management quality. He estimated that this amounted to a shortfall of 14.3 per cent relative to France and 13.2 per cent relative to Germany (1968: 274). Evidently the details of these results may not be robust when confronted with new growth theory and, in any event, rely on assumed rather than estimated relationships. For example, the elasticity of output with respect to capital is assumed to be equal to the share of profits in G D P - which ignores the possibility of externalities to physical investment, as argued by Romer (1986) - while education's contribution is assumed to be through the impact of formal schooling on earnings - which ignores not only the externalities suggested by Lucas (1988), but also the impact of vocational training. An alternative accounting framework based on regression analysis provides an interesting check on these results. In Table 6.4, we have used an equation estimated to 'explain' growth in the international cross-section of countries after 1960, to consider the UK in a European context during the last part of the Golden Age. The equation taken from Levine and Renelt (1992) predicts that growth is increased by higher investment and higher school enrolment, and uses estimated rather than imposed coefficients. The equation includes a negative impact of government consumption spending on growth (presumably through both direct and indirect effects on productivity change), and normalizing for these variables also predicts a, catch-up or conditional convergence effect. This approach produces results which are not so very different from Maddison's, bearing in mind the difference in period covered. The main differences are in a somewhat bigger impact from lower investment and from scope for catch-up. Again
British economic growth since 1945
135
Table 6.3. Differences in the sources of growth, 1950-87 (% per year) (a) Maddisoris estimates France 1950-73 Labour input Education Capital input Total factor productivity Backwardness Other specific Residual efficiency 1973-87 Labour input Education Capital input Total factor productivity Backwardness Other specific Residual efficiency
Germany
UK
0.18 0.39 1.84 3.02 0.71 0.43 1.88
0.15 0.19 2.27 3.50 0.70 0.54 2.26
0.01 0.20 1.75 1.27 0.17 0.35 0.75
-0.24 0.56 1.48 0.92 0.11 0.19 0.62
-0.49 0.05 1.28 1.01 0.28 0.22 0.51
-0.19 0.41 1.12 0.82 -0.25 0.33 0.74
(b) Denison's estimates for 1950-62 Excess over the UK Labour input Capital input Total factor productivity Backwardness Other specific Residual efficiency Total
France
Germany
-0.15 0.28 2.50 1.47 0.31 0.72 2.63
0.77 0.90 3.30 2.14 0.35 0.81 4.97
Sources: (a) Derived from Maddison (1991); capital and labour include both quantity and quality; and 'backwardness' comprises technology diffusion and structural change. (b) Derived from Denison (1968: table 6.4); 'backwardness' is the sum of rows 20 and 24 in Denison's table, 'other specific' is row 14 and residual efficiency is row 29. some part of the shortfall in growth is 'unexplained' and represents a possible productivity failure. Indeed, the hypothesis of relatively poor UK growth performance after allowing for differences in factor accumulation has not been refuted by other investigations based on new growth theory. Crafts (1992: table 8) shows that much lower residual TFP growth for the UK in 1950-73 still shows up strongly after reasonable allowances are made for the Romer and Lucas externalities and a regression-based estimate of the differential impact of backwardness. The residuals from different specifications of regression-based estimates of new growth models allowing for catch-up and factor accumulation conclude that the UK's growth was unusually
136 Charles Bean and Nicholas Crafts Table 6.4. Levine and Renelt accounting for UK growth shortfall, 1960-73 (% per year)
Scope for catch-up Investment Education Public sector Unexplained Total
-0.49 -0.55 -0.08 -0.19 -0.32 -1.63
Source: Crafts (1996); comparisons of the UK with the average offifteenWestern European countries. low in the Golden Age (Crafts, 1992; Do wrick and Nguyen, 1989; van de Klundert and van Schaik, 1996). These studies give an average residual of —0.63 per cent for the UK between 1950 and 1973. Table 6.5 takes a closer look at some aspects of investment. Britain has a relatively low rate of physical investment, particularly in the period 1950-73. This may have been less serious for growth than might be imagined. De Long and Summers (1991) have argued that the key category of investment spending, where social returns are exceptionally high, is machinery and equipment; here the gap relative to other countries is quite small. The main shortfall lies in (residential and non-residential) construction expenditure; in 1960-73 the UK investment share of GDP was 5.4 percentage points below the twelve country median (OECD,1991a). Moreover, recent work at NIESR has found no evidence that social returns were significantly larger than private returns to fixed capital formation in British manufacturing (Oulton and O'Mahony, 1994). Crafts (1991a) also points out that relative to German manufacturing the UK investment rate throughout the postwar period has been fairly similar, and the main UK problem has appeared to be in the incremental capital to output ratio. The UK experienced quite normal European tax rates in the early postwar period, while since then taxation, notably direct taxation, has risen somewhat less than the average. Crafts (1991a: table 6) also shows that in the long run the Bacon and Eltis (1976) hypothesis that the rise of the non-marketed sector replaces investment is incorrect - actually marketed sector consumption has given way over the long run. If it were to be argued that relatively high taxation was the reason for relative economic decline, some more nuanced data would be required, reflecting the disadvantageous structure of the system, as Tanzi (1969) argued. Similarly, as Table 6.5 suggests, the crude data on school enrolment do not show a serious British shortfall and match Maddison's growth accounting findings on education, but a more subtle examination focusing more on training may give more insights. This issue is taken up in sections 4 and 7. In the recent literature, links between macroeconomic policy and productivity have been examined explicitly in terms of a bargaining model (Machin and
British economic growth since 1945 137 Table 6.5. Some comparisons relating to investment UK Fixed capital formation (% GDP) 1950-73 1974-9 1980-9 Equipment investment (% GDP) 1950-73 1974-9 1980-9 Total taxes (% GDP) 1950-73 1974^9 1980-9 Non-consumption taxes (% GDP) 1950-73 1974-9 1980-9 Gross primary school enrolment rate (%) 1950-73 1974-9 1980-9 Gross secondary school enrolment rate (%) 1950-73 1974^9 1980-9
12-country median
16.9 19.4 17.4
24.6 22.4 20.4
8.1 8.8 8.2
8.8 8.8 8.4
31.6 39.0 41.4
31.6 44.0 46.8
22.3 30.7 30.7
19.0 32.0 34.9
95 104 103
105 104 100
67 73 83
58 76 91
Sources: Investment data from OECD (1991a); taxation data from OECD, Revenue Statistics of Member Countries (various issues) - for 1950-73 an average of 1955, 1960, 1965 and 1970; school enrolment rates from Unesco, Statistical Yearbook (various issues).
Wadhwani, 1989). In this chapter it will be argued that the UK's postwar macroeconomic policy stance was to some extent harmful to productivity improvement through the 1970s, and that the experience of the 1980s reveals that willingness to invest in a spell of high unemployment in the short term can pay dividends in the longer term. Given the existence of these short-term trade-offs and the wide range of macropolicy regimes experimented with through time, it is particularly interesting to review short-term macroeconomic performance using Table 6.6. This shows that during 1950-73 the UK Misery index was no better than the European average, that inflation was extremely high in the 1974-9 period without the recompense of lower unemployment than elsewhere, and that during the 1980s the UK's inflation fell to a level similar to that in other European countries, but unemployment was rather higher and by now much above the level of the 1950s and 1960s. It is also noteworthy that the UK appears during the 1970s and 1980s to have moved to a high equilibrium level of unemployment both compared with other
138 Charles Bean and Nicholas Crafts Table 6.6. Inflation and unemployment: UK, 1950-89 (%) UK 1950-73 Inflation Unemployment 1974-9 Inflation Unemployment NAIRU 1980-9 Inflation Unemployment NAIRU
12-country median
4.6 2.8
4.6 2.3
15.6
9.2 4.4 3.2
5.0 5.2 7.3
10.0 7.9
7.0 7.2 5.0
Note: Layard and Nickell (1985) estimated the UK NAIRU as 1.96% for 1956-66. Sources: Maddison (1991) except for NAIRU estimates which are from Layard et al. (1991); the twelve countries are the European entries in the Maddison database. The NAIRU estimates are actually averages for 1969-79 and 1980-8. countries, especially those with more corporatist labour markets (Calmfors and Driffill, 1989), and compared with the early postwar period. 2
The legacy of the 1930s and World War II
As is well known, World War II left the UK with formidable problems of macroeconomic management, which are dealt with in some detail in Cairncross (1985). Key features of this legacy are reported in Table 6.7, which reflects a serious monetary overhang resulting from the forced savings of wartime, a major balance of payments problem, particularly in terms of dollars, and a high ratio of national debt to national income. In addition, of course, the government was well aware of the large productivity gap between the UK and USA (see Table 6.2), and realized that during 1940-4 net non-war capital formation had averaged minus 14 per cent of NNP (Pollard, 1992: table 5.8). Equally widely recognized are the implications of the 'Road to 1945' (Addison, 1975). The 'inescapable experience' of the economic problems of the 1930s and life on the home front during the war generated a large electoral majority for a Labour government committed to major increases in welfare spending and nationalization of a significant fraction of economic activity. Other important aspects of the postwar legacy prominent in the economic history literature are generally less well understood by non-specialists. As Table 6.8 reports, the interwar economy enjoyed some economic growth and, indeed, experienced a small increase in growth relative to the late Victorian/Edwardian period. Encouragement has been derived from the strong recovery from the depression after 1932, together with accelerated TFP growth and restructuring in manufacturing, and a reading of some texts gives quite an optimistic picture (Pollard, 1992).
British economic growth since 1945 139 Table 6.7. National debt, money supply and balance of payments: UK, selected dates (£m, current prices)
1938 1945 1950
Stock of national debt
Ml
8149 23774 25986
1862 4967 5710
Balance of payments Visible
Invisible
Current account
GDP at FC
-285 -250 -51
+ 230 -620 + 524
-55 -870 + 473
4932 8787 11391
Source: Mitchell (1988).
Table 6.8. Growth rates of output, capital stock and TFP: UK, selected dates (% per year)
1870-1913 Whole economy Manufacturing 1924-37 Whole economy Manufacturing
Output
Capital
TFP
1.8 2.0
2.0 2.6
0.5 0.6
2.3 3.2
1.8 1.3
0.7 1.9
Source: Matthews et al. (1982: table 8.3).
Nevertheless, in common with most of Europe, in a number of important facets interwar Britain was still very far from the type of economy to exhibit high growth potential for the years after 1945. Gross investment averaged only 8.9 per cent of GDP between 1923 and 1938 (Feinstein, 1972), and Table 6.8 shows a slow rate of growth of the manufacturing capital stock. Investments in other types of capital stressed by new growth theory were also weak. Research and development spending was only about 0.3 per cent of GDP (Sanderson, 1972), and only 10 per cent of the top 200 British manufacturing firms had their own research laboratory compared with 73 per cent of the American top 200 (Mowery, 1986). Of the cohort born in 1910-29, only 17.6 per cent went into secondary education (Floud, 1954). The UK had a revealed comparative advantage very different from that of the United States. Table 6.9 indicates this and shows clearly the US strength and British weakness in the high-technology, R & D-intensive industries of the period. This partly reflects American advantages in natural resources and a large home market, which raised the payoff to R & D (Nelson and Wright, 1992). It is also associated with the weak educational background of British managers (Gourvish, 1987) and the continued survival of small-scale family capitalism in an era when the hostile takeover was unknown (Chandler, 1990; Hannah, 1974). Last, but by no means least, the postwar legacy included a dangerous combination of traditional, craft-based trade unionism and product markets in which competition
140 Charles Bean and Nicholas Crafts Table 6.9. Revealed comparative advantage rankings: UK and USA, selected dates UK Agricultural equipment Car and aircraft Industrial equipment Electricals Iron and steel Non-ferrous metal Book and film Chemicals Metal manufactures Brick and glass Wood and leather Rail and ship Fancy goods Apparel Alcohol/tobacco Textiles
USA
1913
1929
1937
1913
1929
1937
10 12 5 8 3 16 13 11 7 14 15 1 9 6 4 2
16 14 8 7 4 13 6 10 11 12 15 1 5 9 2 3
16 11 7 5 9 15 8 12 13 10 14 3 4 6 1 2
2 4 3 5 9 1 10 12 6 11 7 8 13 14 15 16
1 2 4 5 8 3 6 9 7 12 10 11 14 13 15 16
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Note: At this time hi-tech sectors might be thought of as cars and aircraft, chemicals, electricals and industrial equipment based on research and development spending. Source: Derived from Crafts (1989: table 1). was very weak. British economic policy reacted to the world recessionary shock by moves to stabilize profitability and raise prices relative to vyages, including the encouragement of collusive agreements among firms and the imposition of tariffs. This was a sensible policy for the short term but had unfortunate implications for productivity improvements (Broadberry and Crafts, 1990,1992). Protectionism and trade associations were further strengthened in the war economy.
3
Reconstruction
On the whole, the Attlee government has been well regarded by recent academic commentators (e.g. Hennessy, 1992). Economic historians have been impressed in particular by its macroeconomic management in the difficult early postwar circumstances (Cairncross, 1985; Tomlinson, 1990). Three questions deserve to be raised in this section, and consideration of them leads to a somewhat less favourable assessment. They are the following: • Was the policy stance as good for long-term growth as it seems to have been for short-run macroeconomic performance? • Would a less gradualist approach to liberalizing the wartime economy have been better? • Was the 'New Jerusalemism' of the postwar welfare state policies a major reason for relative economic decline, as famously alleged by Barnett (1986)?
British economic growth since 1945
141
Table 6.10. The short-term UK macroeconomic position, 1948 and 1951 (1938 = 100)
1948 1951
GDP
IND Y
C
I
G
X
M
111.9 123.4
120 145
100.3 103.6
100.8 116.6
129.1 147.2
113.3 142.2
85.3 99.3
Sources: Feinstein (1972), except for industrial production taken from Eichengreen and Uzan (1992). Table 6.10 reflects the short-run position and the policy priorities of the government. Given the difficulties shown in Table 6.7, policy emphasized rapid increases in exports, tight control of imports and home consumption, and an early return to a balanced budget. The reconstruction years of the late 1940s were a time of moderate inflation (about 5 per cent per year) and very low unemployment rates (around 1.5 per cent as measured at the time). Afixedexchange rate was maintained, apart from a devaluation in 1949 of 9 per cent in the effective exchange rate (30 per cent against the dollar). By 1950 the balance of payments position had been rectified. The increase in industrial output reported in Table 6.10 is average for Western Europe excluding Germany (Eichengreen and Uzan, 1992: table 7). In many respects this was a successful stabilization, bearing in mind the precariousness of the original situation. Nevertheless, there were costs attached to the approach adopted. The most obvious was the preservation of market power of British producers. The tough anti-trust policy which the Board of Trade had originally proposed to introduce was abandoned (Johnman, 1991; Mercer, 1991), and the fear of harming the postwar export drive together with the threat perceived to the balance of payments from superior American industrial productivity also encouraged the retention of import controls. These still applied to 54 per cent of imports in 1951 (Dow, 1964: 175), and the threat of import competition remained remote for most manufacturers - the share of imports in home demand was still only 4.7 per cent in 1955 (Scott, 1963: 41). Eichengreen (1993) estimates econometrically that excess money balances were still around 50 per cent and a monetary overhang persisted until the mid-1950s. The policy choice was to forgo both currency reform and elimination of the overhang by price liberalization and an associated devaluation. Accordingly, the desire to hold down consumption implied retention of rationing and import controls, while the currency remained inconvertible apart from an ill-fated brief experiment in 1947. This gradualist approach to liberalization of the economy fitted with the overall approach of the government to its dealings with organized labour, in a climate where not only was a high priority given to redistribution of income but fears of wage inflation loomed large in the changed employment situation (Jones, 1987). The government sought a cooperative 'social contract* with the trade unions, in which wage restraint was accepted in return for welfarism, a commitment to full employment and non-interference in industrial relations (Tomlinson, 1991). Broadberry (1994) suggests this was conducive to a low NAIRU and shows econometrically that wage-setting behaviour changed substantially from its prewar mode. Nevertheless, the gradualism of the social contract implied that opportunities
142 Charles Bean and Nicholas Crafts
to increase competitive pressures on British managers were passed up - there would be no Thatcher-shock. It would seem likely that the overall approach to macroeconomic stabilization, the approach to improving the balance of payments, the failure to reform industrial relations and the obstacles to increasing product market competition all combined to lower productivity improvement relative to a more thorough and rapid liberalization of the wartime economy. Postwar welfare policy was largely based on the approach of the wartime Beveridge Report, which had argued the need to double benefit expenditures. Comparison of 1938 with 1951 shows that in the latter year government outlays were about 7.6 per cent higher as a share of GDP, about 4.8 per cent of which was accounted for by social services (Peacock and Wiseman, 1961). The share of direct taxes in private income from production doubled, rising from 11 per cent in 1938 to 22 per cent in 1950 (Weaver, 1950: 202). Clearly, the Beveridge approach to social security, relying on a contingency benefit/social insurance structure, was unduly expensive (Dilnot et al., 1984). A sizeable fraction of benefits went to the non-poor - Weaver's estimates suggest that about 25 per cent of expenditures in 1949 were received by the middle or wealthy classes and that, if these flows had been eliminated, direct taxation could have remained at its 1938 level without unbalancing the budget (1950: 202, 208). Nevertheless, the main reason for initiating a very progressive system of income taxation was to hold down the consumption of the rich in the postwar austerity period, not to fund the National Insurance system. It would seem mistaken to put undue blame on the expansion of the welfare state alone for any adverse effects high direct taxation may have had on postwar capital accumulation and growth. Nor would it be right to suggest that the situation was one of complacency on efficiency while concentrating too much on equity, as does Barnett (1986). The government was keenly aware of the advantages of higher productivity for achieving its macroeconomic objectives (Tomlinson, 1993), but had to accept that the social contract and the short-term imperatives reduced its leverage. Finally, how much impact did the Marshall Plan have on postwar British growth? Aid to the UK amounted to about 1.8 per cent of GDP in 1948-51. Eichengreen and Uzan (1992) estimate that the direct effects of this on growth were very small, while recent British research also argues that any impact of conditionality of the aid was also quite limited, apart from pushing the UK into the European Payments Union (Burnham, 1990). The UK did accept the idea of the Anglo-American Productivity Council as a way of seeking to reduce the productivity gap, or at least of placating American sentiment, but this appears to have had little or no success in raising British productivity levels in the prevailing circumstances of firms with market power and traditional industrial relations structures (Carew, 1987; Middlemas, 1986:158-62). In sum, the European Recovery Program probably had no more than a marginal impact on reconstruction in postwar Britain. 4
The Golden Age
The years 1950-73 are a period when British economic growth appears to have been exceptionally disappointing when placed in a comparative international context, as section 1 made clear. At the start of the period, the new Conservative administration
British economic growth since 1945
143
sought to change relatively little and, in particular, continued to aim for a social contract with the Trades Union Congress. By the end of the 1950s, slow growth had started to be perceived as a problem by politicians, and from then on there were a series of policy innovations designed to promote growth by stimulating investment, striving to achieve economies of scale and to improve technology. Two central questions permeate the literature on this growth experience: • How far was slower growth due to low UK investment? • What part did government policy play in inhibiting growth? Answers to these questions are far from agreed, but recent research at least provides some new, if preliminary clues. The simplest approach to the first of these questions is mechanically to apply recent cross-section regression results from studies of OECD growth to the difference in investment rates between the UK and the twelve-country median reported in Table 6.6. Using the Levine and Renelt (1992: table V, equation (v)) regression and accepting their standard approach to measurement, schooling differences are inconsequential and all the action comes from lower physical investment in the UK, which would have had 0.8 per cent higher growth if investment had been at the median level. An alternative approach would be to consider only equipment investment, as proposed by De Long and Summers (1991); here the gap is smaller between the UK and the median country, and an impact of only 0.2 per cent per year would be predicted. Englander and Mittelstadt (1988) also find for OECD countries that growth of the stock of R & D has an impact on growth. Prima facie, this should have been an aspect of British investment favourable to faster growth in the UK. By 1995 UK R & D spending had risen to 1.7 per cent of GDP and by 1960 to 2.5 per cent, second only to the USA (United Nations, 1964). Table 6.11 shows that this relatively high spending has not continued during the past twenty years, but during the Golden Age the UK was a big spender and this marks a substantial change from the 1930s. Unfortunately, much of this effort may have been misdirected from the point of view of economic growth (Pavitt, 1976), and after 1960 the stock of R & D was growing faster elsewhere. Englander and Mittelstadt's estimates imply that from 1960 to 1973 R & D may have accounted for 0.3 per cent per annum of the growth gap between the UK and Germany, while the patent statistics shown in Table 6.11 are clearly disappointing from the late 1950s on. Government accounted for much of the spending, and this concentrated on high-technology defence-related activities, which offered little in the way of externalities to the rest of the economy (ACOST, 1989). A detailed examination also suggests some weaknesses in human capital formation not well captured by either growth accounting or cross-section regression studies. These relate to training rather than education, an aspect of the British economy which has been the subject offiercecriticism lately. Table 6.12 shows the relative shortfall of vocational qualifications which had built up in the British workforce relative to Germany by 1979, the earliest available date. O'Mahony (1992) seeks to estimate the impact of the different proportions of qualified workers on the German-UK productivity gap in a cross-section of industry in 1987, allowing for both direct and Lucas-externality effects. She concludes that these together were responsible for a 13.4 percentage point gap.
144 Charles Bean and Nicholas Crafts Table 6.11. Research and development and patenting, 1938-89 (a) R&D
expenditure/GDP (%) 1964 1970
France Germany Japan UK USA 12-country median
1.8 1.6 1.5 2.3 2.9
1.9 2.1 1.8 2.2 2.6 1.3
1980
1989
1.8 2.4 2.2 2.2 2.4 1.4
2.3 2.9 3.0 2.2 2.8 2.0
(b) Share of non-American patents granted in USA (%) 1938 1950 1958 1979 France Germany Japan UK
9.2 38.2 1.5 22.7
15.5 0.6 0.03 36.0
10.4 25.6 1.9 23.4
7.9 22.7 27.8 10.8
1988 6.8 17.4 41.3 8.1
Sources: OECD (1991c, 1991d); Englander and Mittelstadt (1988); Nelson and Wright (1992); Pavitt and Soete (1982). Table 6.12. Labour force with intermediate vocational qualifications, 1979 and 1988 (%)
France Germany UK
1979
1988
32 61 23
40 64 26
Source: Steedman (1990).
Assuming that this applied across the economy as a whole and had accrued entirely in the period 1950-79, this might have accounted for a difference in growth between the UK and Germany of about 0.4 per cent per year. Given that German training is probably substantially better than the European average, and that there was a similar qualifications gap in the oldest workers in 1978/9 who would have been trained in the 1930s and 1940s (O'Mahony and Wagner, 1994: 16), this no doubt exaggerates the part training played in the UK's growth gap with the twelve-country median. In sum, it is likely that traditional growth accounting somewhat underestimates the role of lower investment in lowering UK growth relative to other European countries in the Golden Age. If so, it may turn out that this is not because of large externalities to physical investment, but because of failure to appreciate fully the role of differences in training of the workforce. It should also be noted that O'Mahony's estimates, while consistent with the Lucas externality, do not confirm the hypothesis of completely endogenous growth; her point estimates for manufac-
British economic growth since 1945 145 Table 6.13. Labour productivity comparisons in international companies: reasons for differentials, 1972 German advantage over UK (%) 'Economic' causes Length of production runs Plant and machinery Other 'Behavioural' causes Strikes and restrictive practices Manning and efficiency Total
5.5 5 2 3.5 8.5 27
American advantage over UK (%) 20.5 6 6 5 6 50
Source: Pratten (1976). turing would imply an elasticity of output with respect to broad capital of less than 1, with a best guess of about 0.6. It seems unlikely on this evidence, however, that the whole of the shortfall of UK growth, discussed in section 1, can be accounted for by elements of broad capital formation. This impression is confirmed by a consideration of the literature on UK productivity performance at this time - and by the events of the 1980s (see section 6). Pratten's study of productivity differences in international companies operating in several locations, summarized in Table 6.13, suggests that 'behaviourial differences' were responsible for about half the observed productivity gap between the UK and Germany. This is consistent with twenty-five studies relating to 1950-73, summarized in Pratten and Atkinson (1976: 574), twenty-three of which suggest that inefficient use of labour lowered productivity. There seems also to have been an impact from industrial relations reducing the rate of growth of productivity from new technology, according to the case studies in Prais (1981). How could such poor performance persist? The answer seems to lie partly in the 'public bad' of unreformed industrial relations, which even incoming American firms had to endure, partly in the weakness of entry threats, and partly in the ineffectiveness of takeover threats as a discipline on management that was still weak on qualifications (Acton Society Trust, 1956). Big changes appeared on the company scene during the 1950s and 1960s. At this point, the move towards corporate capitalism advocated by Chandler (1990) took place, and at a pace faster than on the continent. By 1970,72 per cent of the top 100 companies were M-form compared with 40 per cent in Germany (Channon, 1973; Dyas and Thanheiser, 1976). The 1948 Companies Act increased disclosure requirements; the improved information facilitated mergers and, for the first time, hostile takeovers. Whereas in the 1940s only 788 firms disappeared via merger, in 1960-9 the number was 5635 (Hannah, 1983). By 1970 the share of the largest 100 firms in manufacturing output was about 40 per cent, almost double the level of 1949, while the averagefive-firmconcentration ratio had risen to about 65 per cent (Walshe, 1991: 340-1).
146 Charles Bean and Nicholas Crafts
While these changes appear to have increased market power (Cowling and Waterson, 1976), there is little evidence that the merger and takeover boom resulted in better productivity performance or selectively eliminated bad managers, rather than promoting a race for survival throughfirmsbecoming too big to be vulnerable to takeover (Cowling et al.y 1980; Meeks, 1977; Singh, 1975). The role of government policy in the UK's growth failure during 1950-73 is not easy to quantify, but it was probably important in terms of both acts of commission and acts of omission. Both demand- and supply-side policy deserve to be considered. From the early 1950s, macroeconomic policy embraced Keynesian demand management, particularly aiming at fine tuning through tax changes, based on a strong commitment to full employment, but retaining the rules-based approach of a fixed exchange rate until 1972, with only one devaluation from $2.8 to $2.4 in 1967. Modest budget deficits of around 3-4 per cent of GDP permitted a reduction in the ratio of national debt to GDP to about 87 per cent in the late 1960s (Ward and Neild, 1978). The Conservative governments of the 1950s assiduously pursued the social contract and even appeasement of the trade unions (Smith, 1990). As the locus of bargaining moved to a more decentralized shop steward-based level, governments from 1962 onwards turned to more formal incomes policies, which in various forms were a major plank of counter-inflationary policy until 1979. Dow (1964) delivered a damning indictment offinetuning, arguing that stop-go policy had actually destabilized the economy. This verdict has subsequently been challenged, but the econometric evidence does not suggest any great contribution to stabilizing demand growth (Hatton and Chrystal, 1991). At the same time there is evidence that the UK experienced milder fluctuations than most other OECD economies (NEDO, 1976), and any damage done may have mostly resulted from distracting the Treasury from potentially more fruitful policy initiatives. Supply-side policy evolved very considerably over time, although continuously embracing substantial subsidies to physical investment (Musgrave and Musgrave, 1968). Policy towards investment was generally poorly designed, however, and most commentators are sceptical of any substantial impact on investment volumes (Morris and Stout, 1985: 868). The system of investment allowances was subject to frequent alteration, targeted physical while neglecting human capital, where the externalities were probably much larger, and was highly distortionary in its effects. The 1960s saw a series of innovations to policy, including the establishment of the National Economic Development Council in 1962, the Industrial Training Act (1964) and the setting up of the Industrial Reorganization Corporation (1966), which marked a shift towards a more selective approach to sponsorship of industry. The tradition of voluntarism and absence of reform prevailed in industrial relations until the unsuccessful Industrial Relations Act (1971). There is no evidence, however, to indicate that these initiatives had any substantial positive effects. Probably more helpful was the strengthening of pressures on managers from competition and trade policy, including notably the Restrictive Practices Act of 1956 and during the 1960s participation in tariff reductions under GATT. On the other hand, the nationalizations of the 1940s were not reversed, and a combination of principal-agent problems and lack of competition delivered poor performance in this sector, which accounted for nearly afifthof all investment (Vickers and Yarrow, 1988).
British economic growth since 1945
147
The biggest failings probably lay in the neglected areas of policy - the unwillingness to tackle industrial relations problems, the failure to recognize fully or to give priority to the need to strengthen workforce training, and the inability of successive governments to face up to inefficient use of labour and to create an environment which put pressure on managers to eliminate X-inefficiency. All of these problems were known to informed commentators of the day, although they have certainly been put in sharper focus by the experience of the 1980s. 5
Shocks and stagflation
In common with the rest of Europe, the UK faced a turbulent macroeconomic environment after 1973, entailing severe adverse shocks to both aggregate supply and aggregate demand. Extremely pessimistic prognoses of future growth prospects were common, and great stress was placed both on the balance of payments position and a declining share of profits as long-term obstacles to a revival of growth. In fact, the long-term threat to growth from the stagflationary shocks was generally overestimated. Some of the inhibitions to faster growth in the Golden Age do, however, emerge more clearly when the disappointing years after 1973 are considered. In particular, government supply-side policy once again was developed in the context of an attempt to sustain a social contract with the trade unions. Two questions worth addressing are as follows: • Was growth performance as bad as Table 6.1 suggests? • How successful were industrial policies in remedying market failures which constrained growth? Table 6.1 reported TFP growth of only 0.6 per cent between 1973 and 1979, with the growth rate of real GDP/head falling to 1.5 per cent. This would have been smaller still had it not been for North Sea Oil, which accounted for growth of about 0.5 per cent per year. Recent work suggests that two mitigating factors need to be considered: namely, errors in the estimates of capital inputs and the implications of incorrect expectations in the new situation after the first oil price shock. Muellbauer (1991) discusses the first of these factors, noting that standard series do not adequately adjust for changes in capacity utilization and that the oil shock probably led to 'premature scrapping': that is, asset lifetimes lower than assumed by the CSO statisticians. Allowing for utilization, in manufacturing at least, leaves the productivity slowdown little altered. The 'premature scrapping' bias may play a significant role, but it cannot at present be estimated reliably. Darby and Wren-Lewis (1991) review the second hypothesis. They conclude that there is good survey evidence thatfirmsfailed to anticipate that sales growth would not recover quickly towards its pre-1973 level and hoarded labour during the late 1970s, only to shake this out in the early 1980s when their mistakes became both obvious and very costly. This gives a flavour of disequilibrium to the observed productivity performance of the 1973-9 period. Both arguments point to the possibility of exaggerating the impact of the changes in policy after 1979 on productivity performance. Indeed, Darby and Wren-Lewis claim that underlying labour productivity growth in manufacturing tended to a constant trend rate of about 2.7 per cent over 1967-87.
148 Charles Bean and Nicholas Crafts
The growth outcome took place in the context not only of the oil price shock, but also of a strong domestic tendency to inflation in the context of a new macroeconomic policy regime. After 1972 the pound was afloatingcurrency, allowing greater policy discretion, while at the same time a combination of fears of unemployment, technical changes in methods of monetary control and a hope that a dash for growth might promote productivity advances combined to produce a very loose policy stance (Blackaby, 1978; Dimsdale, 1991). M3 grew at over 27 per cent per year in both 1972 and 1973, while the budget, which had been in balance in 1970, showed a deficit of 9 per cent of GDP by 1975. The incoming minority Labour government of 1974 inherited massive inflationary pressures, but was politically much too weak initially to act to contain them (Dell, 1991). When disinflation was eventually introduced in 1975, incomes policy was at the centre of the stage in the attempt to regain price stability without jeopardizing employment levels, and again the cooperation of the trade unions was seen by the government to be the key to its success. This placed major constraints on policy, and indeed one recent commentator concluded that The recurring theme of this review of the 1974-9 government has been that of the damage done by its anxiety not to offend the trade union movement' (Brown, 1991: 226). Supply-side policy for growth by the mid-1970s had become strongly linked to a more active and selective set of industrial policies, while reforms of taxation, training and industrial relations were neglected (Artis and Cobham, 1991). Further nationalizations included British Aerospace, British Leyland and British Shipbuilders, while the Industrial Reorganization Corporation was succeeded in 1975 by the National Enterprise Board. These interventionist moves were generally supported by trade unions. Subsequent comment by economists has been much less enthusiastic. For example, Hindley and Richardson (1983) show how relatively low the returns were on government-backed mergers where soft loans were involved, while Morris and Stout (1985: 873) concluded that 'through these institutions, it was losers like Rolls-Royce, British Leyland and Alfred Herbert who picked Ministers'. Table 6.14 demonstrates a very mixed productivity performance in the nationalized industries, although airways, gas and telecommunications all show fast growth of output per worker during 1970-80. Reports by the Monopolies and Mergers Commission on coal (1983) and electricity (1981) were severely critical of long records of misuse of both manpower and capital. It is certainly difficult to conclude that nationalization was a plausible way to accelerate catch-up growth, taking into account the data on comparative productivity levels. Vickers and Yarrow, who highlight the very low returns made on nationalized industry investment, stress deficiencies in government control and, associated with this, excessive investment and internal inefficiency as characteristic of nationalized industries in the 1970s (1988: 143-7). In sum, although the raw data probably exaggerate somewhat the weakness of productivity performance in the later 1970s, this was a period in which growth was weak and supply-side policy was badly orientated. As in the 1930s and the early postwar period, the exigencies of coping with a difficult short-term macroeconomic situation produced a policy package which precluded reforms that might have improved medium-term growth performance. The disruptions of the 1970s probably had a worse impact in reducing growth in
British economic growth since 1945
149
Table 6.14. Productivity outcomes in UK nationalized/privatized sector, selected dates
Coal Railways Steel Electricity Postal Gas
Telecom Airways
Labour productivity growth (% per year) 1970-80 -2.4 -2.0 -1.7 3.7
-0.1 4.9 4.3 8.1
1980-90 8.1 3.2
13.7 2.5 3.4 4.9 7.2 6.0
Relative US/UK labour productivity level Late 1930s Mid-1970s 3.80 2.97 1.97 1.93 1.63
7.60 3.95 2.57 3.47 2.28 2.79 2.69 1.52
Sources: Bishop and Thompson (1992); Rostas (1948); Smith et al. (1982). the short term than in further weakening the UK supply side or bringing new binding constraints on long-term growth. At any rate, as theory would predict (Layard et al., 1991:100-90), the profits squeeze turned out to be temporary, and the dismal record of the 1970s promoted an eventual reaction against the supply-side policies of the 1960s and 1970s. 6
Recovery in the 1980s and its legacy
The Conservative government elected in 1979, with Margaret Thatcher as its leader, offered an escape from the disappointments of the 1970s through institutional reform on the microeconomic front, and the espousal of monetarism in macroeconomic policy. Among the key elements of this new policy stance were the following: • Supply-side policy moved towards increasing pressure for cost reductions to eliminate X-inefficiencies, and away from the pursuit of allocative efficiency. Government regulatory rather than market failures were targeted for policy action, subsidies were reduced and eventually privatization was given a high priority. This amounted to a rejection of the central thrust of postwar industrial policies. • The government finally abandoned counter-cyclical policy in the form of demand management, as practised by the Keynesian policy-makers of the 1950s, 1960s and early 1970s. The Thatcherites, if not the Wets, accepted rapidly rising unemployment as the consequence both of seeking to establish a credible counter-inflationary policy and of reducing the UK's productivity gap, although they did not foresee the full implications of the switdi in policy stance. • There was a heightened emphasis on efficiency relative to equity in the resolution of policy conflicts. Thus, priority was given to cutting personal taxes rather than
150 Charles Bean and Nicholas Crafts
expanding welfare spending, and to faster growth and restructuring rather than income redistribution or support for lame ducks. In all these respects the government was proceeding contrary to the wishes of trade union leaders and saw itself as seeking to escape from the trade unions' veto on economic reform (Holmes, 1985). The poor performance of the 1970s and fortuitous political circumstances - notably the Falklands War, which may have raised the Conservative vote by 25 per cent in 1983 (Price and Sanders, 1994) - gave a radical British Prime Minister an extended window of opportunity to undertake economic reform. Many of the policy developments of the 1980s would indeed have been regarded as inconceivable by informed opinion in the 1960s and 1970s - for example, the reduction of the top marginal income tax rate to 40 per cent, the privatization of the major public utilities and many council houses, the indexing of welfare benefits to prices rather than wages, the outlawing of the closed shop and the decimation of the National Union of Mineworkers. Similarly, the government's re-election in 1983 and 1987, despite the very high levels of unemployment in those years, flew in the face of the conventional wisdom previously accepted by all postwar governments. Tables 6.1 and 6.2 give prima-facie support to the suggestion that the new policy regime brought an improvement in productivity performance. This view is strengthened by regression studies allowing for catch-up and factor accumulation. Crafts (1992)findsthat growth is now better than predicted, whereas in the Golden Age it was worse, while looking at the growth component unexplained by Levine and Renelt's (1992) regression along the lines of Table 6.4 suggests that the previous growth shortfall turns into a superior UK performance. The estimated turnarounds are about 1.5 per cent and 0.8 per cent per year respectively. Two important and closely related questions arise from this: • Was the Thatcher Experiment worth it? • What might be the long-term as opposed to the short-term effects of the policy switch? Everyone would agree that there were losses as well as gains from the new policies, and a decisive argument in their favour probably requires a positive impact via a long-run rather than a transitory growth rate improvement. This may seem more plausible than it might once have, since new growth theory suggests ways in which this result could occur. The new policy stance affected the framework in which investment decisions and bargains between management and their workers were made, as well as having direct effects on the use of capital and labour. The thrust of policy was to attack restrictive practices and overmanning, rather than to aim for government sponsorship of investment-led growth. Changes in taxation were relied upon to promote capital accumulation. Table 6.2 showed a large reduction of the manufacturing productivity gap with Germany between 1979 and 1989, much of which occurred in the first half of the 1980s. Yet it is widely agreed that, at least until 1985, this recovery owed little or nothing to investments in plant, people or technology, but came from more efficient use of existing factors of production. Econometric studies found that this initial
British economic growth since 1945
151
change was linked to the conduct rather than the structure of industrial relations, with changed working practices and manning levels contingent on reduced trade union bargaining power, which resulted from unemployment shocks, anti-union legislation and greater competition in product markets (Bean and Symons, 1989; Haskel, 1991; Machin and Wadhwani, 1989). Notably, it has been argued that it is undeniable that the most dramatic changes in working practices have been achieved in industries with product market crises' (Brown and Wadhwani, 1990: 68). Similarly, as is reflected in Table 6.14, the government had some success in improving productivity in the public enterprise sector, notably after defeating major strikes in steel (1980) and coal (1984-5). It appears, then, that in the 1980s the UK performed a somewhat belated catching-up exercise; this seems to confirm the diagnosis that slow growth in the Golden Age owed a good deal to institutional factors which reduced productivity growth, and which were politically difficult to reform once entrenched. Obviously job shedding in manufacturing tended to raise broad capital per worker for the remaining labour force, but O'Mahony and Wagner (1994:8,24-5) conclude on the basis of a growth accounting exercise that about half the improvement in labour productivity in British relative to German manufacturing came through greater efficiency in factor usage (TFP), and they also note that regressions at the sectoral level suggest a doubling of the speed of catch-up of its German counterpart by the average British industry in 1979-89 compared with 1960-73. By contrast, British weaknesses in technology and capital accumulation, both human and physical, continued throughout the 1980s. The immediate route to faster growth was not apparently the one prescribed by new growth economics, as the data on investment, R & D, patenting and vocational training in Tables 6.5,6.11 and 6.12 indicate. An assessment of technological capabilities in the late 1980s concluded that the UK remained a weak performer (Patel and Pavitt, 1988), and spending on R & D by industry lagged behind that in competitor countries - 1.6 per cent of industrial GDP in 1988 compared with 2.2 per cent in Germany and 2.1 per cent in Japan (Stoneman, 1991: 12). O'Mahony (1992) provides estimates of the sources of the German productivity lead over the UK in the late 1980s based on a cross-section production function analysis of productivity gaps by sector. The essence of her results is reproduced as Table 6.15. The key implications are that the pressures of the 1980s did not eliminate the German advantage in accumulation of factors of production, but they did succeed in forcing British management to make better use of the factors of production that were available than their German counterparts. This is an interesting reflection on the relative strengths and weaknesses of German and British institutions, as pointed out by Carlin (1996). So, were these policy reforms worthwhile? The answer obviously depends on what social welfare function is assumed, and on assumptions about the counterfactual path of both policy and the economy. The main costs of the policy reforms can be expected to have arisen in terms of additional unemployment and increased inequality of incomes. Thus, on a standard definition adopted by Eurostat that 'poverty' is an income less than 50 per cent of average equivalent income, the UK experienced a sharp change dissimilar to the rest of the EU. From 1975 to 1985 the percentage of persons in poverty rose from 6.7 to 12.0 per cent in the UK compared
152 Charles Bean and Nicholas Crafts Table 6.15. Productivity of labour in manufacturing, 1987: sources of German lead over UK
Human capital Physical capital R&D Efficiency Total
13.4 10.1 5.8 -7.5 22.2
Source: O'Mahony (1992). with an (unweighted) average increase from 13.4 to 14.1 per cent in the other eleven EU countries (O'Higgins and Jenkins, 1990). Johnson and Webb (1993) estimate that about 80 per cent of the rise in UK inequality was policy induced. The policy stance also risked a large rise in structural (long-term) unemployment as manufacturing shed jobs, particularly among unskilled males. Layard et al. (1991:446) estimate that 'mismatch' raised the NAIRU by 1.54 percentage points in the 1980s relative to the 1970s, and Table 6.6 confirms that the UK NAIRU rose by more than the European average. An illustrative calculation might go as follows. The additional mismatch component of the UK NAIRU of 1.5 percentage points per year might be taken to indicate the cost of abandoning conventional 1970s policies. Evaluating this at average labour productivity (presumably a slight overestimate) implies an annual loss of the same fraction of GDP per head. Losses from rising inequality will depend on the investigator's degree of inequality aversion (e). Crafts (1993) suggests that welfare may have been reduced by about 5.6 per cent of GDP by rising inequality not experienced elsewhere in Europe, if the well-known Atkinson (1970) approach is adopted with e = 1.0, or about 2.9 per cent with e = 0.5. Annual welfare losses from the Thatcher Experiment on this approach can be put at somewhere between 4.4 and 7.1 per cent of GDP. Against this needs to be set the value of the productivity gains it produced. If a 20 per cent improvement in productivity in manufacturing relative to the average of France and Germany between 1979 and 1989, as shown in Table 6.2, is regarded as a plausible lower-bound guess, then these gains were worth at least 6 per cent of GDP by 1989, with at least half coming by 1984. A more refined calculation might seek to express these flows in present-value terms, but would require still more assumptions. On balance, then, it appears that, if the productivity gains are a once-and-for-all catch-up (levels effect), the assessment is finely balanced and will be very sensitive to assumptions and value judgements. If, on the other hand, the changes of the early 1980s were to lead to permanently faster growth, then virtually everyone would be likely to conclude that the present value of this welfare improvement is worth the short-run unemployment and inequality costs. It is exactly such possibilities that are contemplated by endogenous growth theory. An early return to the policy approach and weak performance of the 1970s seems unlikely, despite doubts about the continued political appeal of Thatcherism (Crafts, 1991b). The framework in which firms and unions operate will be
British economic growth since 1945
153
characterized by the competitive environment of the single European market, and by the end of the 1980s, changes in the structure of industrial relations consolidating the loss of trade union power were evident (Gregg et aL, 1991; Purcell, 1991). Also the gains from future technological improvements are less likely to be diluted than in the past by workers' resistance to new working arrangements, so there should be some permanent improvement to the rate of productivity growth (Muellbauer, 1991: 106). Any large long-run effect probably depends on greater broad capital accumulation. Here the evidence is hard to read, although the data of Tables 6.5,6.11 and 6.12 are not particularly encouraging. A more optimistic view might be entertained, if late 1980s education and training reforms bear fruit, if the changes in industrial relations enhance investment and innovation activities offirms,and if lower marginal direct tax rates do eventually have the effect predicted by new growth theory, perhaps reflected in the higher private sector investment of the late 1980s. By 1989 this amounted to 17.1 per cent of GDP compared with 13.1 per cent in 1979 and 13.3 per cent in 1985. On the downside, there are some negative long-term effects to contemplate, mostly reflecting the deindustrialization which accompanied the Thatcher period, and which is considered more fully in section 9. The reduced size of the tradables sector implies future terms of trade losses from a need over time to devalue the real exchange rate in order to combine internal and external balance. The estimates in Church (1992) indicate a possible real income loss from this effect of the order of 0.4 per cent per year. More speculatively, deindustrialization may reduce the opportunity to benefit from externalities to investment and learning effects. Fixed investment in manufacturing fell from 3.6 per cent of GDP in 1979 to 2.8 per cent in 1989. In sum, the long-run implications of the recovery and reforms in the UK during the 1980s are uncertain, and accordingly a full assessment of the Thatcher Experiment is still premature. The next three sections take up these issues in more detail. 7
Bargaining models and productivity change
The surge in manufacturing productivity of the 1980s gave rise to a substantial modelling and empirical effort by economists. The most fruitful approach so far has explored the implications of bilateral bargaining between unions and management over wages and effort (manning levels). In essence, the predictions of this model are that bargained productivity levels will tend to rise when workers' bargaining power is reduced and when the rents available tofirmsin the product market decrease. This leads to the expectation that higher unemployment (which will tend to weaken labour's bargaining position) and greater product market competition will tend to raise output per head, while unionization, particularly of a fragmented and non-encompassing kind, will tend to lower it. Government industrial relations, competition and macroeconomic policies will all, of course, potentially affect the outcome of the bargain. The approach is set out clearly in Haskel (1991), which also contains references to the earlier literature. The emphasis of this literature on bargaining and productivity has thus been on the comparative statics of shifts in the balance of bargaining power between labour and capital that induce changes fn effort levels and productivity. According to this
154
Charles Bean and Nicholas Crafts
view, the effect of the reduction in the relative bargaining strength of labour brought about by the Thatcher government's industrial relations reforms would be to produce a temporary surge in productivity growth during the transition to a higher level of output per worker, but no permanent effect on the growth rate. This contrasts with the rhetoric of Conservative politicians, who argued that the reforms would not only raise productivity immediately through ending overmanning and permitting a more efficient use of the workforce, but also foster a more entrepreneurial spirit in the economy and thus generate a permanently higher growth rate. The 'new' growth theory identifies a number of potential 'engines' of growth in total factor productivity. These include: externalities from investment in physical capital (Romer, 1986); externalities from investment in human capital (Lucas, 1988); and purposeful investment by firms in R&D and new product development (Aghion and Howitt, 1992; Grossman and Helpman, 1991; Romer 1987,1990). All of these approaches suggest ways in which industrial relations and other policies can affect the long-term growth rate of total factor productivity and thus also of per-capita output. In particular, in all of these models total factor productivity is the result, either unintentional or intentional, of an investment decision. Anything that reduces the return to that investment will end up reducing the growth rate. In an important contribution, Grout (1984) showed how bargaining between management and unions can lead to underinvestment in the absence of binding contracts, because ex post the workforce is able to expropriate some of the quasi-rents that are properly the return to capital. In Grout's model, this reduces the amount of capital invested and thus the level of productivity. Furthermore, the more powerful is the union, the bigger is the adverse effect on output. However, with an engine of growth of the kind considered by the new growth theory, this translates into an effect on the long-term growth rate. In an appendix, we integrate Grout's insight into an endogenous growth model in which the engine of growth is purposeful investment by firms in R & D in order to develop new products-specifically Romer's (1987) expansion-in-varietiesmodel. In this model,firmsinvest resources in order to develop new products. The reward for innovation is a period of monopoly profits. (New products are protected by patent. If this were not so, and new goods could be costlessly imitated, then nobody would ever find it profitable to invest in R&D.) When there is bargaining between management and the workforce, the latter is able ex post to extract some of these monopoly profits for itself. This in turn reduces the incentive to undertake new product development in thefirstplace. The result is an inefficiently low growth rate. Since a particular feature of the British industrial relations scene during the postwar period has been the presence of craft-based rather than industrial unions, the analysis also explores the implications of bargaining with multiple unions. Because each union can 'hold the firm to ransom' in turn the inefficiency is worse, and the growth rate lower, than with a single union representing all the workforce. This inefficiency due to bargaining would disappear either if the union, or unions, could write a binding contract ex ante that ties its hands and ensures that investors reap the full return from R & D, or if a reputational equilibrium could be established in which the workforce eschews exploiting its bargaining power today, knowing that in return the firm will undertake more R&D. While the former may be infeasible, the latter is not impossible (see the paper by Eichengreen (1996) for
British economic growth since 1945 155 Table 6.16. Panel regressions of total factor productivity growth in British industry (dependent variable: total factor productivity growth) With Order dummies 0.153 (1.49) Union recognition -0.086 (0.25) Multiple unions -0.754 (2.06) Multiple union dummy 0.689 (1.33) Employment shock -0.130 (3.36) Concentration ratio 0.703 (1.22) -0.301 Import share (0.47) 2.391 Standard error R2 0.210 2.042 Durbin-Watson No. of observations 794 Capital growth
No Order dummies 0.141 (2.11) 0.055 (0.17) -1.112 (3.38) 0.668 (1.28) -0.134 (3.62) 0.388 (0.96) 0.561 (1.07) 2.689 0.196 1.959 794
Notes: f-statistics in parentheses; coefficients on time and Order dummies omitted for brevity.
further discussion of such co-operative equilibria). However, it is far less likely when multiple unions are involved, because of the difficulty of preventing individual groups of workers from free riding on the restraint of others. Thus the model points to the presence of multiple unions as a particular depressing factor on the growth rate. Table 6.16 reports empirical evidence on this, and other, explanations for the UK's relatively poor postwar growth performance from a panel study of British industrial performance. Ideally, one would like to test the hypothesis using firm-level data, but unfortunately at present the best we can do, if we wish to examine both time-series and cross-section variation, is the three-digit industrial level. The dependent variable in these regressions is thus the average annual growth rate of total factor productivity in 137 industries (Orders III-XIX of the 1968 Standard Industrial Classification) over eight periods (1954—8, 1958-63, 1963-8, 1968-73, 1973-6, 1976-9, 1979-82 and 1982-6; data are not available for all industries over all periods, so the panel is unbalanced). These data are taken from table H.ll of Oulton and O'Mahony (1994), which contains more details about its construction, including the linking of data from the 1968 and 1980 SICs. The independent variables include the following: the average annual rate of growth of the capital stock in the industry (from table H.3 of Oulton and O'Mahony); the fraction of workplaces within the industry in which there is at least one recognized union for manual labour (from the 1980 Workplace Industrial Relations Survey); the fraction of workplaces within the industry in which there is
156 Charles Bean and Nicholas Crafts
more than one manual union (also from the 1980 WIRS); a multiple union dummy that takes the value zero before 1979 and is the same as the previous variable for 1979-82 and 1982-6; an employment 'shock' dummy which takes the value zero before 1979 and the proportionate fall in employment during 1979-82 for 1979-82 and 1982-6 (from table H.4 of Oulton and O'Mahony); thefive-firmconcentration ratio in the industry (from the Census of Production); the ratio of imports to home demand (from Wells and Imber (1977) and Business Monitor); and a set of common time dummies. The rationale behind the inclusion of these variables is as follows. The growth rate of the capital stock is included to pick up externalities from physical capital formation on to knowledge creation of the sort discussed by Romer (1986). Since the variable in question relates to capital formation within the industry, this variable should be thought of as relating to 'local', rather than economy-wide, externalities. The latter will be related to the total rate of capital accumulation in the economy, and so will be captured by the time dummies. However, it is surely reasonable to believe that the spillovers of the sort identified by Romer should be stronger on to firms producing similar products than on to those in relatively unrelated industries. Because the rate of capital accumulation may be correlated with the rate of total factor productivity growth, we treat it as endogenous and instrument it with lagged capital growth. The next four variables capture the impact of the industrial relations environment - in particular, the impact of multiple unions. The multiple union dummy permits the adverse effects of multiple unions to be ameliorated during the 1980s as a result of the reduction in union power brought about by the twin effects of the 1980-2 recession and the Thatcher government's industrial relations reforms. The employment shock dummy is intended to capture the idea that the transformation of industrial relations was greatest in those industries subject to the greatest shakeout of employment in the 1980-2 recession; this variable has been widely employed in other work on the 1980s productivity resurgence (Bean and Symons, 1989; Layard and Nickell, 1989; Metcalf, 1989). The concentration ratio and the level of import penetration are intended to capture the intensity of domestic and foreign competition. Theory suggests that the impact of these variables could go either way. On the one hand, a protected market position allows managers to opt for a quiet life and rest on their laurels. On the other hand, it allows larger and more durable quasi-rents from innovation, so promoting growth. All of the regressions include a full set of common time dummies; these are highly significant, with a marginal significance level close to zero. Unfortunately, there is no time-series dimension available for the industrial relations variables, the 1980 WIRS being thefirstof its kind. Consequently, it is not possible to include a full set of three-digit industry dummies and still identify the industrial relations effects. However, we have tried including a set of fifteen Order-level dummies. The first column of Table 6.16 includes these, while the second column omits them (they are just significant at the 1 per cent level). Fortunately, the results do not seem to be too sensitive to their exclusion. In addition, the Durbin-Watson statistics in both regressions are very close to 2, while if a lagged dependent variable is included in the regression, it invariably attracts a small, statistically insignificant coefficient.
British economic growth since 1945
157
Consequently, we are led to conclude that industry fixed effects are relatively unimportant, and their omission should not bias the results unduly. Both sets of empirical results suggest that the presence of multiple unions significantly depresses total factor productivity growth, with multiple union workplaces having exhibited an annual rate of total factor productivity growth some 0.75-1 percentage points lower than that achieved by single union workplaces over the 1954-79 period. Once the presence of multiple unions is controlled for, there appears to be no residual effect from union recognition. This suggests that cooperative outcomes that avoid the Grout-style inefficiencies may have been relatively easy to establish in single union workplaces. The WIRS data set also allows us to distinguish between multiple unions which bargain separately with management, and those which bargain jointly. Machin et al. (1993), using the 1984 WIRS, find that pay tends to be higher, and company financial performance worse, in workplaces where multiple unions bargain separately compared with single union workplaces; however, where they bargain jointly there is no noticeable difference. Thisfindingis in line with the theoretical model of the appendix, which suggests that the adverse effect of joint bargaining by multiple unions should be less than where there is separate bargaining. When we decompose the multiple union variable into the fraction who bargain jointly and the fraction who bargain separately, the point estimates of the coefficients are quite similar, and a formal test of equality of the two coefficients fails to reject the hypothesis that they are in fact the same (the marginal significance levels are 98 per cent with the Order-level industry dummies and 74 per cent without them). However, we cannot reject with much confidence the hypothesis that it is only where multiple unions bargain separately that there is an adverse effect on total factor productivity growth (the marginal significance levels are 30 and 7 per cent respectively). We conclude that there is insufficient information in this data set to indicate conclusively how the precise structure of bargaining has affected the growth rate. Turning next to the impact of the 1980-2 recession and the industrial relations reforms of the Thatcher government, we find that the coefficient on the multiple union dummy is positively signed and of the same order of magnitude as the coefficient on the multiple union variable itself. This suggests that the adverse effect of multi-unionism has been significantly ameliorated during the 1980s; indeed the implication of the estimates containing the Order-level industry dummies is that the effect had been almost completely eradicated. The results are therefore in line with those of Machin and Stewart (1994), who, using the 1990 WIRS and a subjective measure of productivity growth,findno effect from multi-unionism during the late 1980s. The employment shock variable is also highly significant, in line with most earlier work. The rate of capital accumulation is only rather marginally significant. In addition, the size of the coefficient is rather small, being about a quarter of the size required to generate endogenous growth under the Romer (1986) model. Thus the evidence for externalities from capital formation is rather thin, although it should be remembered that the economy-wide externalities are picked up by the time dummies. In addition, a curious feature of the results is that the bias under Ordinary Least Squares turns out to be negative rather than positive as expected. This could reflect the impact of measurement error, perhaps because the effect of capital externalities takes some
158
Charles Bean and Nicholas Crafts
time to spill over to other firms. Neither of the variables intended to capture the impact of thefirm'scompetitive environment seems to play any major role. If anything, the results seem to suggest that a high degree of competition, at least from domesticfirms,may tend to depress total factor productivity growth, but clearly there is insufficient information in the data to draw any firm conclusions. These estimates are, of course, only preliminary and may be subject to reappraisal in the light of further work. They do, however, confirm the view that the UK's poor postwar growth performance has been due in at least some measure to the industrial relations environment, and in particular to its fragmentary, craft-based union structure. 8
Human capital formation
During the 1980s a wide consensus developed that education and, particularly, vocational training were seriously in need of reform, and that successful implementation of new arrangements could contribute to better growth and productivity performance. Research at NIESR was fundamental to this heightened awareness of weakness in human capital formation in the UK. Policy initiatives were seen as necessary to success in this area by virtually all economists, given the prevalence of market failure in training activities and the likelihood of suboptimal skills formation in an environment characterized by multiple equilibria (Finegold and Soskice, 1988). While early government policies seemed to be dominated by massaging the youth unemploymentfigures,the later 1980s were characterized by several new initiatives, notably the development of a new system of National Vocational Qualifications from 1986 and the introduction of the National Curriculum in the Education Reform Act of 1988. The government has announced itself pleased with its new policies, which it sees as delivering record levels of both participation and achievement in education and training (Cm 2563, 1994: 49). Some of the numbers certainly look encouraging (Robinson, 1994). For example, in 1978-9 only 23.6 per cent of school-leavers hadfiveor more A-C grade passes at O level, compared with 41.2 per cent in 1992-3 at GCSE. In 1993, 32 per cent of eighteen-year-olds went into higher education, compared with only 15 per cent in 1988 and 12 per cent in 1981. The Labour Force Survey showed a rise in the proportion of employees receiving job-related training in the last four weeks from 8.2 per cent in 1984 to 15.4 per cent in 1990, with only a small dip in the ensuing recession. The fraction of employees in manufacturing with no qualifications fell from 71 per cent in 1979 to 56.8 per cent in 1989 (O'Mahony and Wagner, 1994:15). Nevertheless, criticisms of both the quantity and quality of the training of the British labour force continue to be made in an extremely vociferous way. For example, Layard et al. (1994: 11) conclude that 'our training system remains a formula for national decline', while Ekinsmyth and Binner (1994: 15) find that 'literary and numeracy problems are just as prevalent among young adults in 1992 as they were in 1981'. Wagner (1991) emphasizes the continued higher content and volume of foreman (Mdster)-level training in Germany, while Prais (1993) deplores a dilution of standards in curricula and tests for those acquiring vocational qualifications. At present, it is quite unclear what has been achieved in terms of human capital
British economic growth since 1945
159
formation. Undeniably, there have been failings in policy design, and incentives to politicians and educational providers to exaggerate the quality of the outputs of the training system have undoubtedly been important. On the other hand, real expenditures by employers on training appear to have roughly trebled between 1970 and 1989 (Crafts, 1991b: 94), and it is difficult to believe that this, together with additional participation in education after sixteen, does not indicate a higher real accumulation of human capital. Given that replication of a German-style system in the UK is not feasible, the focus on higher education in the UK may be a better alternative than many critics imply (Soskice, 1993). It is possible, therefore, that the UK has improved its medium-term growth prospects somewhat as a result of improved human capital accumulation, but this remains to be demonstrated. The quantification of the impact on growth of reforms in this area is highly desirable, but it is not likely to be an easy research project. 9
Deindustrialization
It is widely believed that manufacturing is of crucial and special importance to the achievement of rapid and sustained economic growth in the UK. In 1973,42.6 per cent of employment and 38.4 per cent of value added was in industry, while by 1989 the percentages were 29.4 and 29.5 respectively. Closely linked to this are suggestions that the UK has in the past and once more now faces a balance of payments constraint on growth, and that the experience of deindustrialization and shrinkage of the manufacturing base in the 1980s has adversely affected future growth prospects. Neoclassical economics certainly recognizes that weak export performance and/or a voracious demand for imports may mean that growth implies adjustment to this through adverse moves in the terms of trade: that is, a depreciation of the real exchange rate. The implication would be that growth of real income would fall below that of real GDP, but this would not mean that the latter would be reduced. Thirlwall (1979) argued that, for economies like the UK, real wage rigidity prevented this adjustment mechanism from functioning, and this would lead to persistent deflation and failure to realize fully the growth of productive potential. This fear seems not to be justified, however; OECD (1989: 35) presents data which show that neither the UK nor other OECD economies were characterized by long-run real wage rigidity in the 1970s and 1980s. More recently, the notion that manufacturing matters for growth arises from new growth models of the type proposed by Lucas (1988), building on earlier work by Krugman (1987). Here human capital is accumulated through work experience, but at rates which differ across sectors. Thus an economy's initial comparative advantage, or shocks which temporarily shift equilibrium specialization, can determine future growth rates. If specialization is in activities which have high learning effects, living standards improve relatively quickly and learning effects will also tend to lock in the initial pattern of specialization. High-technology industries have been highlighted in the literature as characterized by strong dynamic learning effects (Mowery and Rosenberg, 1989) and spillovers of R & D (Jaffe, 1986). Krugman (1987) saw the Thatcher-shock as running a risk of lowering long-term growth through switching specialization away from manufacturing. The implications
160 Charles Bean and Nicholas Crafts Table 6.17. Value added in high-technology industry, 1979 and 1988
Germany Japan UK USA
% of manufacturing
% of GDP
1979
1988
1979
1988
16.5 16.6 15.0 20.8
20.3 22.9 19.7 24.4
5.6 4.9 3.7 4.8
6.3 6.6 3.9 4.7
Source: OECD (1992b, 1984); high technology as defined in OECD (1992a). of large invisible earnings from rentier foreign investments might have had similar implications before World War I (reflected in Table 6.9), and the model obviously could be deployed as a theory of why Britain's early start was a disadvantage to later generations. Despite its intuitive appeal, this model is probably not a very good approximation to the way the world works, or a good reason for inferring disastrous long-run growth consequences arising from the decline of the British manufacturing base. There would be greater reason to worry if international trade led to complete specialization, there were no spillover effects of learning and R & D internationally, and productivity growth potential varied enormously between sectors. Even in the interwar period when these conditions were more nearly fulfilled (Nelson and Wright, 1992), the model would be highly overstated - the direct investment of foreign multinationals like Ford, General Motors, GEC and Hoover sustained many of the fast-growing, hi-tech industries (Jones, 1988). The increasing importance subsequently of multinationals and intra-industry trade weaken its validity further. In any event, it must be doubted whether differences or changes in the structure of activity have had or are likely to have a large effect on overall growth. The UK has had a higher than average deindustrialization since 1973, but an arithmetic calculation based on differences in sectoral productivity growth rates shows that, if it had deindustrialized at the average-rate for Maddison's (1991) sample of countries, the impact would haVe been enough only to raise UK growth over 1973-89 by 0.15 per cent per year. Moreover, as Table 6.17 shows, although high-technology industry is a smaller proportion of British manufacturing than in Germany or Japan, it is not negligible and its share rose in the 1980s. It must be doubtful whether the differences in structural changes shown in Table 6.17 are really big enough to carry serious implications for future growth rates.
10
Conclusions
Our aim in this chapter has been to explore the reasons for the UK's relative economic decline during the postwar period, and to assess the evidence that recent changes in the economy have promoted a permanent improvement in growth potential. At the outset we distinguished between three levels of explanation, to which we now return in reviewing our findings - proximate sources of growth, institutional and policy influences on investment in broad capital/productivity
British economic growth since 1945
161
improvement, and fundamental factors sustaining the equilibrium growth path. Various forms of growth accounting provide information on the proximate sources of growth, but as we have stressed, they are subject to quite serious possibilities of measurement error. If human capital formation is proxied by a schooling measure, then this work suggests that the UK's relatively low growth during the Golden Age was due both to low investment and, particularly, to relatively weak total factor productivity growth. The latter probably reflects, in part, a substantial component attributable to less initial scope for catch-up and reconstruction. Better measurement of broad capital formation would probably highlight shortfalls in workforce training and in production of commercially valuable new products and processes (reflected, for example, in patenting) as weaknesses on the British supply side. This would reduce but not eliminate a residual gap in TFP growth stemming from inefficiency in factor usage. During the 1980s, there was a substantial shakeout of inefficiencies in the British economy and, by now, the UK was no longer an economy with relatively low scope for catch-up. The implication was that, compared with the Golden Age, TFP growth in the UK slowed by much less than the European average. The UK remained an economy with relatively low physical investment, but the gap with other countries narrowed somewhat. Despite some successes, notably in pharmaceuticals, British technological performance continued to look relatively weak, while the net effect of changes in education and training is at present impossible to assess. There surely was a significant transitory (catch-up) component in the productivity improvement during the Thatcher years. At the second level of explanation, that of identifying the impact of institutional arrangements and policy frameworks on growth, research is less comprehensive. This is unfortunate because modern growth theory suggests that these effects are central to understanding relative growth outcomes. Our own results, described in the chapter, suggest (1) that the structure of industrial relations in the UK, with its unfortunate emphasis on fragmented bargaining and multiple unionism, operated to reduce TFP growth prior to the 1980s, and (2) that the changed industrial relations scene of the recent past has not only allowed a once-and-for-all productivity gain, but also improved future growth potential. We do not wish to exaggerate the importance of these findings, or to give disproportionate emphasis to this aspect of the British institutional and policy scene. Our discussion has also identified a number of other features which require detailed analysis. These include, in particular, the incentive structures facing agents contemplating investment decisions, and the extent to which the UK differed from its European peer group. Certainly one aspect which is not well understood is the precise impact of the tax system, but more generally it would be interesting to know how ex ante hurdle rates on investment projects compared, and both how and why they may have differed. At the third level of explanation, that of trying to understand why failures might persist and why government policy interventions seemed to perpetuate rather than eradicate relative decline, at least through the 1970s, we have highlighted the short-termism promoted by the pressures of macroeconomic management and the high weighting given to inflation and unemployment rather than growth. In particular, we have suggested that the pursuit of social contracts in the British case
162
Charles Bean and Nicholas Crafts
worked to inhibit rather than improve necessary supply-side reforms, including, of course, effective reform of industrial relations. Again, we do not wish this argument to be overplayed. We see a need for a great deal of further research into 'government failure', and recognize that poor choice and design of policy played a part alongside vote seeking. We also anticipate that much more will eventually be written on the implications for growth of the interplay of government and producer interests. Finally, how optimistic should we be on future growth prospects and the present administration's claim that the period of Conservative government has ended the long period of relative economic decline? As we noted at the end of section 6, any answer to this question can only be provisional, and that remains our view after the more detailed assessment of the following three sections. To some extent, the answer depends on the relative weight given to the success of supply-side reform compared with negative implications that might stem from the deindustrialization of the economy. We take the changes in industrial relations to be good news and unlikely now to be reversed; we regard the new world of education and training as unproven; and we also wait to see whether evidence is forthcoming to support the argument that the slimming down of British manufacturing will have deleterious effects on future TFP growth. To some extent, however, the answer also turns on future governments' ability to commit themselves to a framework more conducive to effective supply management and, in particular, to taking a long-term rather than a short-term view of supply-side reform. Here too we remain to be convinced; if, for example, the debates over privatization policy are anything to go by, short-termism among politicians of all persuasions still holds sway. Appendix We start by setting out the behaviour of the economy under competitive conditions. Output of final goods, Y, is produced using a fixed number, m, of different types of labour, Lt(i = 1... m), and a composite capital good, K. We assume a Cobb-Douglas structure with
The total supply of each variety of labour is inelastic and normalized to unity, while the composite capital good is a CES aggregate of a variable number, nt, of different varieties of capital good:
Xt = (ir = i xr a ) 1 / ( 1 " a )
(2)
The engine of growth will be an expansion in the number of varieties of capital input. This final goods industry is competitive. It would be nice to be able to introduce imperfect competition, since there is abundant evidence to suggest that unions in imperfectly competitive industries tend to be more successful at rent extraction. It would also be nice to allow the number of required varieties of labour to increase as production becomes more complex; this might permit us to capture demarcation disputes between unions, which the evidence suggests have been important. However, both these extensions of the basic model complicate matters significantly and are left for development elsewhere. Turning now to the capital goods industry, a capital goods producer can develop a new variety of capital good by expending p units offinaloutput. Thereafter it can
British economic growth since 1945
163
produce as many units as it likes of the new capital good at a constant marginal cost of r\ units of final output. It can also acquire a patent giving it an infinitely lived monopoly in the supply of the new good. (If there were no patents, and new designs could be costlessly imitated, then nobody would ever develop new varieties of good.) However, entry into the research activity is free. Finally, the demand side of the economy assumes a representative consumer with an additively separable isoelastic utility function (C 1 " 1 ^ — 1)/(1 — I/a) and time preference rate 0. Assuming also perfect capital markets, this generates the usual intertemporal optimality condition for consumption, Ct: g=Ct/Ct = (j(rt-d)
(3)
where rt is the market interest rate. We start by deriving the growth rate under competitive conditions, in the labour market. In that case, all varieties of labour are fully employed and the market demand for the ith variety of capital good is just the usual marginal productivity condition: (1 - oi)AK7t* = Rit
(4)
where Rir is the user cost of capital (rental rate) for the ith capital good. Now it turns out that, under the assumed structure for the economy, Kin Rit and rt are all independent of time (see Romer (1987) for proof). In that case, it will be optimal for the capital goods supplier to produce all the capital it intends to lease out as soon as it has developed a new design. Dropping redundant time subscripts, we may then write the present discounted value of profits, 7tf, from the innovation of a new variety of capital as: 7i* = R.KJr ~(P + rjK;)
(5)
where the demand for the ith variety of capital is given by (4). Thefirstterm on the right-hand side is the present value of the infinitely livedflowof revenues, while the second term is the combined cost of innovation and production. Maximizing (5) subject to (4) gives the profit-maximizing rental rate as Rt = ri7/(l - a)
(6)
If there is free entry into research, we also know that 7r* = 0 and hence from (5) that Kt = 0(1 - a)/oiy
(7)
Combining (4), (6) and (7) then gives the interest rate as r= (l-o# 1
(8) l
where ij/ = A{\ — a) ~\<xl$frf~ . Equation (8) incidentally confirms the conjecture that the interest rate is time invariant. Substituting (8) into (3) then gives the competitive growth rate as 3comp = «x[(l - o # - 9] (9) We also note in passing that the socially efficient growth rate that would be chosen by a benevolent social planner is (see Romer (1987) for proof) 0SocPlan = < # " 0)
(10)
Thus growth in the competitive equilibrium is too low, essentially because of the monopolistic pricing of capital goods.
164 Charles Bean and Nicholas Crafts Now let us turn to the main topic of interest: namely, the impact on this outcome of the bargaining institutions in the final goods industry. The sequencing of decisions is as follows. First, firms decide how much of each variety of capital to install; this decision is irreversible within the period. Second, bargaining over wages and employment occurs. We assume so-called 'efficient bargains' over wages and employment (McDonald and Solow, 1981) in order to isolate the Grout-style inefficiencies that arise due to the inability of workers to precommit themselves not to extract some of the quasi-rents associated with capital. However, similar results can be obtained if one assumes a 'right-to-manage' set-up in which management retains the right to determine employment unilaterally. Let us start by assuming m( > 2) separate 'craft' unions in each firm that bargain independently with management over pay and employment on behalf of each of the m varieties of worker. For simplicity, each union is an expected income maximizer with V{ = w^i + vv/L, — Lt), where wt is the outside wage (e.g. unemployment benefit) and £, is the number of members of the union (unity in aggregate). In the event that there is disagreement, no production occurs (because Lt is an essential input) and no wages are paid, although the firm still has to finance the leasing payments on the installed capital; the workers are free to claim w-. The Nash bargain then solves
Max(w,,,L,i}fi,, = (n, - nunvu - P,,)1"' = (ATIf. , V " * , 1 - - ^ " L . w ^ i K - wJZ,,]1-*
(11)
where an overbar denotes the 'status quo' points and y is an index of relative bargaining strength. The first-order conditions for this problem are M
- OLYJm) = (Yt-^=1wjtLjt)(l
~ y)
yLit(wit - wit) = (Yt- Z 7 = ^ ^ X l - y)
(12) (13)
These imply bilateral efficiency in employment determination: *Yt/mLit = wit
(14)
and also by summing (12) across labour varieties that the wage share is given by [y + (1 - y)m]i; = ,wjtLjt = [ay + (1 - y) m] Yt
(15)
To close the model we need an explanation for the outside wage, w. Here we make the simplest possible assumption: namely, that w represents unemployment benefit paid by the government and financed by lump-sum taxes. Furthermore, the government is assumed to vary w so as to achieve full employment, although unions and firms in the final goods sector will treat it as parametric when making their own wage and employment decisions. While the assumption of continuous full employment is obviously not realistic, it does help to focus attention on the key issues involved. By virtue of the symmetry of the problem, wages (and unemployment benefits) for all labour types are identical. Using (15), we may then write ex ante profits in the final goods industry as 71, = Yt - i ; = ^jtLjt
i Kjt
- L? = 1 RjtKjt a
- £J'= 1 RjtKjt
(16)
where 4>(m) = 5(\ — a)/(d + m) and d = y/(l — y) is an index of the firm's relative
British economic growth since 1945
165
bargaining strength. Note that <j> < 1 and .27 1.63 :>.01 1.97 1.05 L.75 L.27 1L.67 :>.48 :184 :>.14 1L57 :>.56 1L.82
Note: For the very long-run growth comparison between Sweden and other OECD countries, we use data from Maddison (1982) instead of Maddison (1991). The reason for this is that the figures for Sweden for 1870-1950 are based on a provisional series not intended for publication. The only full series for Swedish G D P for this period is the one published in Krantz and Nilsson (1975), which is used in Maddison (1982). To date, this is also the series used by all scholars doing analyses on long-run Swedish economic growth. Since the figures for Sweden presented in Maddison (1991) greatly differ from those in Maddison (1982), we find it is inappropriate to use this series until definite data exist. Source: Maddison (1982: 212).
study of ultimate causality is facilitated by our focus on Sweden's growth performance relative to other countries. This makes it suitable to search for circumstances where Sweden differs from other industrialized countries in important respects.
2
Aggregate performance
2.1
The growth
record
In the middle of the nineteenth century, Sweden was among the poorest countries in Europe. 1 Approximately 80 per cent of the population was engaged in the agricultural sector. A take-off began in the 1850s, and in the early 1870s industrialization based on raw materials, notably iron ore and lumber, provided a base for sustained economic growth which continued largely uninterrupted for one hundred years. As Table 9.1 demonstrates, Swedish productivity growth was exceptional in the period 1870-1950 compared to other rich countries. An analysis of the comparative success
Economic growth and the Swedish model 243 Table 9.2. Growth in GDP per man-hour in 16 OECD countries, 1950-70 1950-70
1950-60
1960-70
Sweden
TTo
lib
497
Unweighted average Unweighted average excluding Japan and Germany
4.46
3.89
5.04
4.08
3.55
4.62
Source: Maddison (1982: 212). Table 9.3. Average annual growth rate of GDP, GDP per person employed and GDP per capita, 1950-70 (%) GDP per person employed GDP per 1950-60 1960-70 1950-60 1960-70 1950-60 2.8 Sweden 3.4 4.6 3.8 2.7 OECD 3.6 5.0 3.9 OECD Europe 3.4 4.9 4.5 GDP
capita 1960-70 3.9 3.9 3.8
Sources: National Accounts from Statistics Sweden for Sweden; Maddison (1991); OECD, National Accounts 1950-1968; OECD, Historical Statistics. of this period is beyond the scope of the present paper, but reasons that have been emphasized include a sizeable initial stock of human capital followed by rapid human capital formation, a resource-based industry that developed into a technologically advanced investment goods industry, partly as a result of sharp exposure to international competition, a liberal policy environment, and the good luck to have avoided participating in wars during the period. Fully comparable data for the 1950s for all OECD countries are not available. Thus, in order to assess Sweden's relative growth performance in this period, we have to rely on several sources. In Table 9.2, Sweden's growth rates in GDP per man-hour are compared to the averages for Maddison's sixteen countries. The growth in GDP per hour worked was very close to the average for the sixteen countries in 1950-70. But if we exclude the extremely war-torn countries Germany and Japan, which disproportionately benefited from a positive catching-up effect, the Swedish growth rate is above the average for the period 1950-70. OECD data for the 1950s and 1960s confirm the view that emerges from Maddison's data (Table 9.3). Overall growth rate for Swedish GDP, GDP per employed and GDP per capita is practically on a par with developments in the rest of the OECD during both the 1950s and 1960s. The economy performed well during the 1960s, despite the fact that it was clearly disfavoured from a catching-up perspective by having a very high income level compared to the OECD average. Sweden's lagging growth did not manifest itself in the data until around 1970, although signs of an underlying weakness in the economy had shown up a few years earlier. One of the first manifestations of a deterioration in economic performance was that when the Swedish economy grew at a rate comparable to other OECD
244 Magnus Henrekson, Lars Jonung and Joakim Stymne Table 9.4. Average annual growth rate of GDP, GDP per person employed and GDP per capita, 1970-92 (%) GDP
GDP per person employed
GDP per capita
Sweden
~L7
T~2
T~2
OECD OECD Europe
2.9 2.4
1.9 2.0
2.0 1.8
Note: No data exist for the development of GDP per employed before 1977 for the two aggregates. Instead we have used unweighted averages for these years. Sources: OECD, Economic Outlook, June 1993 for GDP and GDP per employed; OECD, National Accounts, Main Aggregates, vol. 1, 1982 and 1994 for GDP per capita. countries, a current account deficit tended to emerge. In 1970-1 a sizeable current account deficit arose, prompting the government to respond with drastic austerity measures. Moreover, in 1969, the real product wage began to exceed the level consistent with long-run equilibrium (Wissen, 1982). For these and other reasons, it is proper to regard 1970 as a watershed year, rather than 1973, which is customarily used in economic growth studies. From Table 9.4 it is clear that the growth rate of GDP in Sweden has been only slightly more than half that of the OECD. The same pattern is apparent for GDP per person employed and GDP per capita. Sweden's relative economic performance appears more favourable in terms of GDP per capita than in terms of GDP per employed. This reflects the fact that the growth of employment has been much faster than the OECD average, particularly in the latter half of the 1980s. Thus, in terms of both overall growth and the simplest productivity measures, the performance of the Swedish economy has lagged behind^ since 1970. More sophisticated productivity comparisons across countries are not readily available. In Table 9.5 the results from one study of TFP in fourteen countries during 1970-85 is presented. Swedish TFP growth is the lowest of all fourteen countries. In Table 9.6 a simple growth accounting decomposition of growth of value added in the non-government sector during the period 1950-90 is presented. The decomposition is done using the conventional formula Y A L - = - +(l_a)- +
a
K -
where K is the capital stock, L is hours worked, a denotes the actual income share of capital averaged over the relevant period, and A is the level of TFP. A dot above a variable indicates rate of change. This simple exercise shows that, until the mid-1980s, growth in value added in the non-government sector can be predominantly ascribed to growth in TFP, although increases in the capital stock in some subperiods have been of great importance. During the 1960s and 1970s, the contribution from labour was invariably negative. In the last period, 1987-90, the pattern is dramatically different: the growth rate of
Economic growth and the Swedish model
245
Table 9.5. Growth rate of total factor productivity (TFP) in the private sector in 14 OECD countries, 1970-85 (% p.a.)
TFP growth Japan Australia Belgium Italy Canada France USA Finland Denmark West Germany Netherlands Norway UK
3.29 2.54 2.53 1.95 1.77 1.72 1.66 1.65 1.53 1.21 0.89 0.74 0.67
Sweden
0.61
Note: TFP growth in a country is estimated as the average output growth in the private sector in each country minus the growth rate accounted for by growth of labour, capital and catching-up potential where the latter is measured by the log of the ratio of labour productivity between the productivity leader and the respective country (productivity leadership is measured at the industry level, i.e. different countries are taken to be the technological leader in different industries). Source: Hansson and Lundberg (1991b).
TFP fell to a fraction of earlier levels, whereas the strong growth of employment and capital stock contributed substantially to output growth. The slow economic growth rate in Sweden since 1970 has had a highly significant impact on the Swedish income level vis-a-vis that of other countries. It is well known that comparing income levels is more difficult than comparing growth rates across countries. The most suitable method is probably to use the OECD's purchasing power parity adjusted measures of GDP per capita. Sweden together with Luxembourg had the third highest GDP per capita in the OECD area in 1970. By 1990, Sweden had fallen below the OECD average for the first time. In 1991, Sweden fell to rank 14, and the GDP level was 8 per cent below the OECD average. In 1993, Sweden was ranked seventeenth with a GDP per capita 13 per cent below the OECD average. Sweden is not the only country that has fared relatively badly; the Netherlands, Australia and New Zealand also lag behind, but no other country has regressed to the same extent. On the other hand, there are a number of countries that have performed extremely well, notably Japan, Iceland, Norway and Austria. In summation, the analysis in this section shows that the rate of economic growth in Sweden was comparable to the average of other industrialized countries until the late 1960s. But the data on growth and productivity indicate clearly that since 1970 Sweden's economic performance has been well below the average of other OECD
246
Magnus Henrekson, Lars Jonung and Joakim Stymne
Table 9.6. Proximate sources of economic growth in the non-government sector, 1950-90: contributions from growth of TFP, labour (hours worked) and capital
Decomposition of growth
Period
1950-60 1961-5 1966-70 1971-7 1978-86 1987-90 1961-90
Y ~Y
A ~A
3.3 5.4 3.9 1.9 2.3 2.5 2.8
2.3 4.7 4.4 2.2 2.0 0.6 2.6
Relative contribution from
L
.l 1.2 1.9 L.7 L2
Source: Prados de la Escosura (1995). 1861-90. b 1920-35. c 1861-1913. d 1861-1993.
a
head fell at —0.4 per cent annually between 1929 and 1935. * Finally, the new evidence points to a fall in the level of economic activity per person, as a result of the Civil War (1936-9), at a compound annual rate of — 3.4 per cent between 1935 and 1940. The reconstruction period was longer in Spain than in Western Europe and lasted until the 1960s, resulting in the level of income per head for 1929 (the highest for the pre-Civil War ear) only being reached again in 1954.2 Autarky presided after the end of the Civil War for a period of twenty years. After the economy practically stagnated in per-capita terms during the 1940s (0.2 per cent), the 1950s witnessed a substantial growth per head as isolation was gradually relaxed.3 However, a remarkable acceleration took place after the major policy reform of 1959, and this lasted until 1975.4 The post-Franco years were of faltering growth, with a yearly rate of 0.9 per cent per head between 1975 and 1985. Then, following the admission of Spain into the EEC, a return to the trend of the Golden Age took place up to 1990 (with per-capita income growing at 4 per cent), only to slacken again at the beginning of the 1990s. In short, in early twentieth-century Spain, sustained growth took place up to the Great Depression, when it was abruptly interrupted by the Civil War, from which recovery was slow under Franco's economic autarky. Fast growth in the Golden Age, resulting to a large extent from the post-bellum reconstruction, ended abruptly after Franco's death (1975), which coincided with the first international oil crisis. With Spain's admission into the EEC (1986), an intense, short-lived recovery took place, before a return to slackening growth in the early 1990s. When Spain's economic performance is placed within the international context,
Growth and macroeconomic performance in Spain, 1939-93
357
& uuu 20 00015 00010 000 -
" y~\
5000 -
«»'%M, rf
y
' y
France;'
Spain \ t ;
/
y ^
. . — • • •
r \"
"Zs 1850 1860 1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 Source: Prados de la Escosura (1995).
Figure 12.1 Spain's real GDP per head: comparative performance, 1850-1993 (1990 Geary-Khamis dollars)
several distinctive features emerge (Table 12.1). The remarkable increase in Spain's real per-capita income, over tenfold since the mid-nineteenth century, represents only a moderate pace of growth when it is compared to continental Europe's industrial nations. Spain's economic growth departed from a lower stand-point in terms of output per person, having stagnated over much of the early decades of the nineteenth century. By contrast, Western European nations industrialized, leading to a deterioration in Spain's international position. Thus, it intuitively appears that the (unconditional) convergence hypothesis, according to which growth rates correlate inversely to departing levels, does not seem to apply to Spain's historical experience prior to 1960. In the search for differentials in Spanish economic performance, several significant periods emerge. From the mid-nineteenth century up to the Spanish Civil War (1936), only the moderately free-trading decades of the 1860s to the 1880s and the 1920s represent a mild attempt to catch up with Western European industrial nations. In the late twentieth century, the period 1960-75 is a major step in closing up the gap. Conversely, three periods appear to be responsible for the widening gap between Spain and the advanced Western European nations: 1890-1913,1939-59 and 1975-85. The turn of the century seems to have been a lost opportunity for closing the gap, as comparison with other late comers such as Giolittian Italy, Sweden and Russia suggests (Prados de la Escosura, forthcoming). Economists and historians usually stress Spain's poor economic performance under autarky (1939-59), in particular during the 1940s. It appears, however, that the 1950s, a decade of generalized growth in Western Europe, could be associated in Spain with
358
Leandro Prados de la Escosura and Jorge C. Sanz
incomplete catching up, as the comparison with Germany and Italy tends to suggest. In the 1950s, as forces making for growth and convergence were stronger, those countries like Spain that failed wholly to catch up seem to have paid a heavier penalty than would have been the case in phases of slowing down. Countries such as Spain which remained closed and did not compete in international markets could not share the productivity growth benefits deriving from the leading nations (Baumol, 1986). Nevertheless, when the Civil War is excluded, the largest loss in relative levels of income per head (but not in productivity, due to high unemployment) emerges from the years between Franco's death (1975) and Spain's admission to the European Community (1986). Systematic historical research on the period is lacking and only superficial explanatory hypotheses relating poor performance to the difficult transition to a democratic regime have been proposed. However, deeper institutional reasons seem to have been at work, such as an overregulated, heavily protected economy which was cut off from the international market (see sections 7 and 8 below). Resource endowments in the postwar period provide some clues about Spain's growth. Physical and human capital accumulation was significant in the post-1950 era, while labour expanded only slowly. Employment grew at 0.8 per cent annually between 1955 and 1974,5 to decline, over the following two decades, at —0.5 per cent, while the labour force expanded at 0.7 per cent. Educational attainment, as measured by the average years of schooling, grew at an average annual rate of 2.2 per cent in the period 1950-75, and at 1.7 per cent per year from 1975 to 1985 (Barro and Lee, 1993). In turn, gross fixed capital stock grew faster than output from 1950 to 1973, and the differential growth rates widened thereafter (Hofman, 1993a).6 In the period 1950-73, capital deepening (i.e. increase in capital stock relative to population) took place at an annual growth rate of 5.7 per cent for non-residential capital stock, and even more rapidly for machinery and equipment (8.6 per cent). Although such a process continued over 1973-89, it slowed down for non-residential capital (to 4.5 per cent) and, especially, for machinery and equipment (down to 3.7 per cent).7 A rapid catching up in terms of non-residential capital endowment per head occurred up to 1973, decelerating thereafter (Hofman, 1993b; Maddison, 1993).8 Given the role attributed to capital in machinery and equipment, as carriers of new technology (De Long and Summers, 1991), its slackening after 1973 could be associated with lower joint factor productivity growth.9 A look at the sources of growth can help us to assess the proximate determinants of postwar Spanish performance. For the period 1965-90, an estimate has been carried out by Suarez (1992), who relaxed usual assumptions about constant returns to scale, perfect competition and exogenous technical progress.10 At the economy's aggregate level, however, Suarez found evidence of non-increasing returns to scale. Table 12.2 summarizes his results.11 Total factor productivity dominates Spanish economic growth in the long run, complemented by capital's contribution. Only after Spain's admission into the EEC (1986) did labour make a significant contribution to growth, offsetting total factor productivity's sluggish growth after 1975. Labour destruction in Spain represented a significant brake to growth over the decade after Franco's death. Deep changes in the composition of output and employment, together with a dramatic increase in R & D activities, help to explain total factor productivity growth up to 1975 (and TFP deceleration thereafter, when
Growth and macroeconomic performance in Spain, 1939-93 359 Table 12.2. Sources of Spain's economic growth, J965-90 1965-74 6.48 1 Real GDP growth 0.72 2 Labour 3 Contribution to GDP growth 0.42 9.82 4 Capital 5 Contribution to GDP growth 2.25 3.81 6 Total factor productivity 6.5 7 (3) as % of GDP growth 34.7 8 (5) as % of GDP growth 58.8 9 (6) as % of GDP growth
1975-85
1986-90
1965-90
1.37 -1.62 -0.90 3.46 0.66 1.61 -65.7 48.2 117.5
4.41 3.00 1.69 5.28 1.40 1.32 38.3 31.7 30.0
3.89 0.15 0.10 6.21 1.41 2.38 2.6 36.2 61.2
Source: Suarez (1992). Table 12.3. Relative growth of Spain's real GDP per head, 1950-85 (Spain's deviations from OECD growth)
Actual growth deviation Less Cyclical bias Catch-up = Adjusted growth deviation Less Employment deepening Capital deepening = Unexplained growth deviation
1950-60
1960-73
1973-85
0.89 -0.15 1.13 -0.09 0.21 -0.25 -0.06
2.15 0.45 0.90 0.80 -0.11 -0.27 1.18
-0.31 0.22 0.42 -0.95 -1.27 -0.24 0.56
Source: Dowrick and Nguyen (1989). structural transformation slowed down until 1985). Empirical research within the on-going debate on convergence and catching up has provided evidence on the case of Spain. Thus, Spain's economic performance can be analysed using a convergence equation. Dowrick and Nguyen (1989) have tested whether post-1950 convergence within OECD countries is explained by total factor productivity catching up, or just by the growth of factor intensities, with their results supporting the former. Their individual country data set includes evidence for Spain which is presented in Table 12.3. Each country's deviations from OECD trend growth, after being adjusted for catching up, are decomposed into capital and employment deepening (i.e. relative to population growth) and unexplained growth. Dowrick and Nguyen's unexplained residual accounts for more than half of Spain's differential growth in the 1960s and early 1970s, and exhibits the opposite sign (doubling the size of the actual growth deviation) for the period 1973-85. Such an outcome could provide enough grounds to dismiss the exercise on the basis of its low explanatory power. However, the story that emerges from it, when the unexplained growth deviation is associated with joint factor productivity, seems to be a plausible one, as it is consistent with the evidence already present in Table
360 Leandro Prados de la Escosura and Jorge C. Sanz Table 12.4 Differential growth of Spain s real GDP per head, 1965-90 (with respect to the OECD average) 1965-75 1.19 A Growth differential due to: 0.28 B Initial income per head -0.07 C Real investment/GDP -0.12 D Human capital E Population growth 0.00 F Solow factors (B + C + D + E) 0.09 G Inflation -0.15 H Money growth 0.59 I Variance inflation 0.08 0.44 J Exports growth K Lagged public deficit/GDP 0.00 L Macroeconomic performance (G + H + I + J + K) 0.96 M Unexplained residual 0.15
1975-85
1985-90
-1.31
1.69
0.23 -0.16 -0.02 -0.01 0.04 -0.25 0.00 0.26 0.15 -0.08
0.30 -0.10 0.02 0.02 0.24 0.07 0.50 0.37 -0.30 0.05
0.07 -1.43
0.68 0.77
Note: Average GDP per head growth in the OECD, 1965-75, 3.35 per cent annually; 1975-85, 2.05 per cent; 1985-90, 2.45 per cent. Source: Andres et ai (1994). 12.2, stressing the role of TFP in Spain's growth. Thus, according to Dowrick and Nguyen, catch-up accounts for the differential trend growth in the 1950s, and for half of it in the 1960s and early 1970s and, in turn, prevented a poorer performance from the 1973 oil shock onwards. When compared to OECD adjusted giov/th, Spain performed slightly below the average in the 1950s, with the increase in employment (relative to population) offsetting the low investment ratio per head. Since 1960, total factor productivity was the only positive contribution to adjusted growth as a result of poor capital and employment deepening.12 In fact, TFP growth (including a reallocation of resources away from agriculture) was not enough to maintain convergence towards the advanced nations after 1973, particularly as massive employment destruction took place from the late 1970s. However, despite the usual association between the unexplained residual and total factor productivity, economists remain dissatisfied and try to take a closer look at the Solow residual. In a recent paper, Andres et al. (1994) have differentiated between Solow elements (derived from an augmented Solow model with human capital) and macroeconomic performance in OECD countries' growth since 1965.13 Table 12.4 presents their results for Spain, showing positive deviations in OECD growth up to 1975 and from 1985 onwards, a negative deviation for the 'transition to democracy' decade, 1975-85, and in all periods a positive catch-up term and negative contributions of factor endowment growth. These results tend to support findings from a previous exercise (Table 12.3). As can be gathered from Table 12.4, macroeconomic performance,-and the unexplained residual that can be associated with TFP growth, account for most of Spain's differential growth.14 Prior to 1975,
Growth and macroeconomic performance in Spain, 1939-93
361
macroeconomic behaviour seems to explain most (80 per cent) of Spain's positive deviation from the OECD average growth, with money and exports growth as its main factors. However, in the transition years from dictatorship to democracy, macroeconomic instability (high inflation, unbalanced budget) led Spain to underperform. Finally, after Spain became a member of the EEC (1986), macroeconomic performance (lower inflation and money, but not export, growth) contributed to a positive growth differential (up to 40 per cent). It may be concluded, then, that the analysis of proximate determinants of growth largely confirms, and extends, the more descriptive evaluation carried out earlier in the paper. In the rest of it, a closer look at the main phases of macroeconomic performance can illuminate the search for ultimate causes of growth, in which the role of incentives provided by the institutional framework appears to be determinant. 3
The legacy of the 1930s and the Civil War (1936-9)
The impact of the Great Depression on Spain is not only a major economic but also a political and social issue, as it coincided with the instauration of Spain's Second Republic (1931), after more than half a century of constitutional monarchy that ended in a bloody civil war. Was the depression a significant cause of the political and social unrest? Did economic policies exacerbate the political and social climate, leading to civil strife and, eventually, to a military uprising? Historical studies have tended to confirm contemporary perceptions of a relatively mild impact of the depression, as the relatively moderate decline in factor returns and prices would suggest (Comin, 1987). It has been argued, however, that a shift towards a restrictive budget policy and the interruption of public works, together with uncertainty about the new political regime, were major causes of the crisis of the 1930s in Spain (Palafox, 1986, 1991). Legal changes favourable to the trade unions, such as the reduction in the number of hours worked (Soto, 1989), might have reduced incentives forfirmsto invest under the Republic (Comin, 1994). Another view points to a tighter link to the international depression through the external sector, aggravated by inadequatefinancialand monetary policies (Hernandez Andreu, 1980, 1986). A look at the latest empirical evidence tends to confirm the idea of a milder impact of the depression on Spain, given the small decline in the level of real GDP per head (Prados de la Escosura, 1995), while outwardly oriented sectors suffered a deeper impact (Comin, 1987). Moreover, recent research suggests a less important role for the government in the economy, given its size, the policy instruments available and the dominant ideologies in Spain at the time. Policies, both monetary and fiscal, were not restrictive but the opposite, as the budget was used as an anti-cyclical instrument offsetting the decline in private investment and exports (Martin Acena and Comin, 1984; Garcia Santos and Martin Acena, 1990). What caused the Civil War of 1936-9 then? A consensus appears to exist pointing to non-economic causes. However, expectations after the collapse of the monarchy were not fulfilled, as proposals for land reform, industrial relations legislation and welfare improvements were not completed or enforced, leading to social unrest and a military coup d'etat (Martin-Acena, 1994; Palafox, 1991). Finally, given the vigorous growth during the interwar period, it could be boldly suggested that the
362
Leandro Prados de la Escosura and Jorge C. Sanz
Spanish Civil War might be represented as a redistributional conflict resulting from the social and political tensions of an earlier rapid growth period, rather than from poverty and economic stagnation as has been advocated by historians. The impact of the Civil War on production factors and institutions was not negligible. War affected the endowment and proportion of production factors, although, when compared to the wider European post-World War II experience, Spain's physical capital was probably damaged to a much lesser extent than its human capital. Industrial areas were hardly affected by war destruction, which concentrated on residential construction, the transportation network and livestock (Malefakis, 1987; Barciela, 1986; Catalan, 1992). In turn, postwar exile and internal repression removed a large proportion of human capital (Lopez Garcia, 1991). The postwar increase in the human-physical capital ratio, apparently a feature common to most war-damaged countries and one which might have made a significant contribution to reconstruction, through an increase in the productivity of capital (Dumke, 1990), hardly took place in the case of Spain, and helps explain the country's poorer performance during the 1950s.15 4
Reconstruction: Spain under autarky, 1939-59
A major difference between Spain and the rest of post-1945 Western Europe (if Salazar's Portugal is excluded) was Franco's dictatorship, which emerged from the Civil War and lasted until 1975. The dictatorship presided over a four-decade period in which a clear distinction can be made between autarky and outward-looking development, with the 1959 stabilization plan as a turning point (although clear differences are noticeable between the 1940s and 1950s). Franco's political regime may be defined as an authoritarian system based upon a limited pluralism of political groups around the dictator, who behaved as a maximizer of power (time and quantity) under constraints derived from internal political and economic conditions and the international context (Gonzalez, 1979). Although not without a background,16 Franco's regime was a distinctive one, in which fascism and traditional authoritarianism blended from its very beginning. However, the autocratic regime evolved enough to adjust from isolation and self-sufficiency, in which the economy grew at below potential - even during the recovery of the 1950s (Table 12.3) - to a cautiously outward-looking economy reaping opportunities to reduce the technological gap with advanced European nations (Donges, 1971). Thus, the extent to which fast growth under Franco after 1960 may be attributed to the dictatorship's economic policies and social stability is a still debated issue. Had it been the case, a painfully achieved, stable institutional framework might have provided permissive elements for economic development, such as an improvement in the definition and enforcement of property rights, a reduction in transaction costs, and a paternalistic attitude on the part of the state, gradually moving away from absolute interventionism during the 1950s (Gonzalez, 1979, 1989/90; Martin Acena, 1994). To elucidate whether there was a causal link between Francoism and growth and catching up, macroeconomic performance under the dictatorship will be surveyed and its main policies discussed below. However, some caveats about the beginnings of Franco's dictatorship are required.17 Uncertainty about its viability after World War II led Franco's regime to give priority to immediate political
Growth and macroeconomic performance in Spain, 1939-93
363
stability over any other competing goal, even though such a choice would condition subsequent growth through a misallocation of resources. This strong constraint provided, in turn, a specially advantageous position to those (already powerful) small groups and coalitions which, in exchange for support to the dictatorship, would derive rents from the public sector and even control the state's economic decisions (Fraile, 1991, 1993). After the Civil War, the new economic authorities decided to follow a new scheme of autarkic development. Such a choice, favoured by the new authorities' ideology, was conditioned by the international isolation that followed the Axis powers' defeat in World War II. A widely accepted perception of Spain as a backward, non-industrialized society, in which both private firms (because of the lack of entrepreneurship) and the government (through free trade policies) had failed to promote industrialization, led to a view of the new totalitarian state as the only institution able to achieve Spain's sustained growth and catching up through economic intervention (Martin Acefia and Comin, 1991). In postwar Spain, as in the rest of Europe after 1945, government intervention was perceived as a crucial instrument for the post-bellum recovery. World War II and the United Nations' boycott (1946), plus the exclusion from reconstruction plans and international organizations such as Marshall Aid and Bretton Woods, reinforced nationalistic tendencies towards self-sufficiency, although the Cold War gradually relaxed Spain's isolation throughout the 1950s.18 Eagerness to embrace regulation and state intervention was a common feature of postwar Europe, counterweighted by the Marshall Plan, which in turn provided the environment for a pro-market economic policy (De Long and Eichengreen, 1993). Although American aid started in 1951 and led to influential economic and military agreements in 1953, Spain's case provides support for the counterfactual proposition that, had the Marshall Plan not been enforced, product and factor market controls, quotas and foreign exchange rationing would have dominated economic policies. In fact, government intervention and planning in Spain aimed at reaching economic self-sufficiency, including technological independence - once the expectations of technological transfers from Nazi Germany vanished (Lopez Garcia, 1991) - regardless of the opportunity cost involved. Nevertheless, Spain obtained $1.1 billion from the US government over 1951-9, which amounted to 18 per cent of the goods and services imported during these years (or 1.6 per cent of GDP). Had Spain received the $676 million aid in a single year expected from the Marshall Plan by the Francoist authorities, it would have represented over 20 per cent of GDP in 1949 (Donges, 1976; Prados de la Escosura, 1995). The need to rely on a coalition of nationalistic and fascist elements during the initial uncertain stages of the new regime conditioned the dictatorship's industrial policy, aggravated by the lack of managerial and engineering skills needed to carry it forward. Industrialists, who had opposed the Republic's welfare and redistributive policies, adopted a hesitant attitude towards Franco's regime in its early stages. Thus, the lack of human capital, on the one hand, and the new regime's urge to industrialize, on the other, made strategic factors of asymmetrical information and the flow of personnel from private industries to interventory agencies, eventually leading to the capture of these agencies by interest groups (Fraile, 1993). The autarkic model of development was built around a policy of protectionism
364 Leandro Prados de la Escosura and Jorge C. Sanz 0.25-. 0.20-
-0.05 -
-0.10
i i i i i i i i i i i i i i i
i i i i i i i i i i i i i i i i i i • » i • ' • '
i • • • •
i •
1930 1935 1940 1945 1950 1955 1960 1965 1970 1975 1980 1985 1990 Source: Prados de la Escosura (1995).
Figure 12.2 Rate of inflation: Spain, 1929-93 (calculated from the GDP deflator)
which persistently aimed at import substitution, using two basic tools: quantitative restrictions and exchange controls. Indiscriminate protectionist policies had deleterious effects on Spanish economic performance, according to most economists and historians. The exchange control policy implied the imposition of an official fixed exchange rate within a context of persistent inflation differentials between Spain and its commercial partners, which constituted a permanent strain on the Spanish currency towards revaluation.19 The appreciation of the real exchange rate of the peseta implied a drag on exports and a stimulus to imports, and tended to exhaust foreign reserves. Such a tendency could not be compensated in the 1950s either by the multiple exchange rate system,20 or by the 'special accounts'.21 Bilateral trade agreements (quotas by country and product) and a trade licensing system were an alternative that resulted in high costs, in terms of efficiency, by providing the appropriate environment for rent seeking and corruption. A foreign exchange black market developed and the Spanish currency experienced substantial depreciation. The imposition of a top limit for foreign investments in Spanish industry (25 per cent of the capital of thefirm)was a complementary procedure to prevent national firms from falling under the control of foreign capital. A proximate measure of the import substitution bias induced by trade controls can be gathered from the ratio of domestic to c.i.f. prices for imported goods. According to Donges (1971), such a ratio multiplied by 2.5 between 1948/9 and 1958/9. Import substitution took place mainly among consunier goods, and to a lesser extent among intermediate and capital goods. The contribution of import substitution to output growth represented, for consumer goods, 75 per cent in the 1940s and 60 per cent in the 1950s, while for intermediate and capital goods it
Growth and macroeconomic performance in Spain, 1939-93
365
1.30 -i
1.25 -
1.20 -
1.15 -
1.10 -
1.05 -
1.00 1950
1955
1960
1965
1970
1975
1980
1985
1990
Source: Summers and Heston (1993: PWT5.5).
Figure 12.3 Relative price of capital goods: Spain, 1950-90 (ratio of investment to PPP price level of GDP)
provided 70 and 49 per cent of its increase in the 1940s, and 24 and 106 per cent in the 1950s respectively (Donges, 1976: 155). Inward-looking policies created serious bottlenecks for industrialization, as traditional export earnings collapsed and the scarcity of foreign exchange put a brake on industrial growth by restricting the supply of raw materials and capital goods. Firms, usually small and relying on obsolete techniques, acted below full productive capacity due to shortage of inputs (or the uncertainty in obtaining them) and did not take advantage of economies of scale. In short, import-substituting industrialization (ISI) resulted in low organizational levels and low technical efficiency. Another major feature of Spanish autarky was a systematic policy of government intervention and regulation of the economy. The state controlled every step in economic activity through a licensing system for starting and enlarging industries. The authorities also regulated prices for commodities considered vital for industrialization, and controlled the evolution of nominal wages in the labour market. A rationing system was introduced and remained in use until the early 1950s, while at the same time government agencies took control of food distribution (Barciela, 1986), and price regulations for agricultural and industrial goods were established (although the latter were relaxed in the 1950s) (de la Dehesa et ai, 1991). Distortion in intersectoral terms of trade contributed to diverting resources from agriculture to industry in an attempt to foster industrialization. In addition, protection and internal regulation forced relative prices of capital goods upwards, resulting in a lower rate of investment for a given level of savings and, consequently, reducing the rate of growth of the Spanish economy.22
366
Leandro Prados de la Escosura and Jorge C. Sanz 2.5 i
0.0-
-2.5 -
-5.0 -
-7.5 -
-10.0 1940
1945
1950
1955
1960
1965
1970
1975
1980
1985
1990
a
Including central government. Sources: Comin (1993); MEH, MOISSES model; Prados de la Escosura (1995).
Figure 12.4 Budget balance of central government and public administrations: Spain, 1940-93 (% of GDP)
A strategic instrument for government intervention was the Instituto Nacional de Industria (INI), founded in 1941 and replicated from the Italian IRI, which organized and directed public investment in the 1940s and 1950s, and still today provides a powerful instrument of industrial policy. Both the failure of post-Civil War attempts to promote private investment through government incentives and the political uncertainty of the Second World War led Franco's pro-Axis government to shift to direct intervention in economic activity. INI was established to coordinate a fast, self-sufficient industrialization through a programme of investments in public infrastructure and the creation of strategic industries (including defence), involving high capital requirements, high risk or short-run low profits. INFs intervention included a great variety of activities, ranging from basic services to manufacturing and mining, but under autarky it specialized in the production of energy and intermediate goods, responding to the economy's bottlenecks and contributing to national defence. INFs role in postwar industrialization has been critically evaluated. In short, it has been argued that, while INI eliminated bottlenecks with its heavy investment, it was an instrument of inward-looking strategy that led to a misallocation of resources (Martin Acefia and Comin, 1991). A differential feature in Spain with respect to the economic measures followed in most postwar Western Europe countries was the maintenance of a pre-Keynesian fiscal policy. In fact, Franco's cabinets kept orthodox ideas about a balanced budget and the control of public expenditure (which was achieved in the 1950s), while
Growth and macroeconomic performance in Spain, 1939-93
367
-5 I 1940
1945
1950
1955
1960
1965
1970
1975
1980
1985
1990
Sources: Chamorro et al (1975); MEH, MOISSES model; Prados de la Escosura (1995).
Figure 12.5 Current account balance: Spain, 1940-93 (% of GDP at market prices)
sticking to economic nationalism and intervention. In fact, neither full employment nor income redistribution policies were enforced in the post-Civil War era. Under autarky, the government budget did not contribute to raising effective demand by establishing unemployment benefits, no progressive income tax was applied, and automatic stabilizers were not introduced until the 1950s. Government budget expenditure on public works, education, health, housing and social insurance was of little significance in the post-Civil War years. In fact, social transfers and education remained on average below 10 per cent of total central government expenditure until 1950, and below 15 per cent until the mid-1960s (Comin, 1994). A full quantitative evaluation of the consequences of such a policy, as opposed to a Keynesian approach, is still awaited, but it is considered to have reduced economic growth (Comin, 1992). A constant disequilibrium in the market between supply and demand, which provoked a persistently high rate of inflation, emerged as a consequence of price regulation. In an attempt to soften this continuous price strain, the authorities decided to allow gradual import increases, which worsened the current account balance (de la Dehesa et al, 1991). In the 1950s, imports of goods represented 11.7 per cent of GDP, while exports were only 6.5 per cent (Prados de la Escosura, 1994).23 Growing disequilibria in the external accounts in the late 1950s (Chamorro et a/., 1975) would lead to reconsideration of self-sufficiency and, eventually, to the end of autarky.24 The economics of early Francoism can be summarized, therefore, as an attempt to achieve rapid industrialization based upon indiscriminate import substitution, with severe restrictions on imports and capital inflows, a complex exchange rate
368
Leandro Prados de la Escosura and Jorge C. Sanz
Table 12.5. GDP and employment composition: Spain, 1929-93
Output Agriculture Industry Construction Services Expenditure Private consumption Investment Government consumption Net exports Employment Agriculture Industry Construction Services
1929
1940
1950
1960
1975
29.2 25.2 7.8 37.8
32.7 19.3 4.0 44.0
30.7 23.5 3.4 42.4
23.6 30.9 3.9 41.6
10.1 30.3 8.1 51.5
6.0 29.4 6.3 58.3
3.5 25.6 8.3 62.6
76.7 16.6
78.4 7.1
72.9 14.7
69.8 19.4
65.1 28.5
65.0 18.8
64.7 21.2
10.3 -3.5
17.6 -3.1
11.0 1.4
8.7 2.1
10.1 -3.8
14.0 2.2
16.0 -1.9
47.7 28.3 3.4 20.6
52.5 2.3 4.0 23.2
50.0 20.2 5.5 24.3
42.3 21.8 6.7 29.2
23.8 27.3 9.7 39.2
18.2 24.5 7.3 50.0
9.9 22.4 9.4 58.2
1985
1993
Source: Prados de la Escosura (1994). structure, internal regulation and direct intervention. This set of policies led to a highly overvalued currency, a current account deficit, low reserves of hard currency, inflation (consumer prices increased by an average of 13 per cent per year in the 1940s, and 10 per cent in the 1950s), and a small and inefficient industrial sector. A clear distinction should be made, however, between the 1940s and the 1950s, as the latter witnessed fast G D P growth per head together with a significant transformation in productive structure (Tables 12.1 and 12.5). In the 1950s, substantial changes took place in agriculture, where labour productivity increased while its relative size shrank in terms of both output (at — 2.6 per cent annually) and employment (at —1.7 per cent), leading to a fall in its relative labour productivity. The absolute reduction of the labour force in the agricultural sector represented 0.5 million workers, while it has been estimated that around one million workers emigrated from agriculture in the 1950s (Leal et al.y 1975). Meanwhile, total investment doubled its 1940s average share in G D P (10 per cent). It has been argued that the first industrializing push in the 1950s was a prerequisite for the fast growth of the 1960s, since human and physical capital accumulated and the domestic market expanded under protection (Gonzalez, 1979). In fact, capital deepening took place, with the stock of physical capital per head growing at 3.2 per cent yearly (or increasing by more than one-third), while non-residential capital per person grew at 4 per cent per year, an increase of 50 per cent (Hofman, 1993b).25 Meanwhile, the average years of schooling for people over 25 years old rose from 2.7 years in 1950 to 3.4 in 1960 (Barro and Lee, 1993). Industrialization, however, depended exclusively on internal demand, and the ability of Spanish firms to reach external markets was very low. Actually, commodity exports, after the postwar boom (1947-53) in which they rose to 8 per
Growth and macroeconomic performance in Spain, 1939-93
369
cent of GDP, fell to 5 per cent for the rest of the decade. By 1960, manufactured exports had hardly reached one-quarter of total exports and represented 7 per cent of industrial output. Moreover, volatility in import capacity due to scarcity of export earnings made investment risky, leading to a lower rate of capital accumulation (Donges, 1976). In turn, restrictions on foreign capital inflows made investment dependent almost exclusively on domestic savings. In addition, investment was penalized by high relative prices for capital goods, slowing down GDP growth. Lacking the capacity to import the necessary inputs and technology, firms suffering from insufficient capital, obsolete equipment and old vintage technology were unable to compete in international markets. In the 1950s, Spain did not have access to innovation and technology through international trade in competing products, and as a result it was prevented from fully sharing the benefits of fast productivity growth in Western Europe. 5
The Golden Age: years of accelerated growth, 1959-75
In the late 1950s, the autarkic system imposed by the Francoist authorities after the Civil War collapsed. The need to put an end to Spain's isolation at the time of the Treaty of Rome (1957), and to restore internal and external balances, led to a drastic change in economic policy at the turn of the decade. Although Spain started a mild opening in the early 1950s, associated with the bilateral agreements with the USA, the pressure to open the economy increased as Spain joined the International Monetary Fund and the World Bank (1958) and the OEEC (1959). The creation of the Common Market (1957) and the introduction of external convertibility in Western European industrial countries (1958) provided a favourable atmosphere for growth. This helped a new cabinet in Spain to introduce a pro-market policy, but not without a clash within the Franco regime between pro-market and autarkic groups (Anderson, 1970). The policy outcome of establishing links with international economic organizations was a most ambitious project of liberalization that began with a stabilization plan in 1959 (Fuentes Quintana, 1984). A gradual opening and factor mobility (capital inflows and labour migration) were achievements of the new pro-market orientation, while the lack of structural reforms affecting the tax system and labour andfinancialmarkets represented its main shortcomings (Donges, 1971). The economic reform included specific measures aimed at controlling the growth of internal prices and achieving budget balance (Fuentes Quintana, 1989). Among the different measures passed to control inflation, it is worth mentioning the prohibition onfinancingthe government deficit through monetization of the public debt.26 Additional constraints were imposed on the loans drawn by the private sector from the banking system.27 The government also decided to control public expenditure through orthodox fiscal policies centred around a balanced budget, and to raise the prices of certain goods (petrol, tobacco) and services (telephone, public transport) supplied by state monopolies, in an attempt to close the gap between official prices and their real costs of provision (de la Dehesa et ai, 1991; Comin, 1994). A second package of measures aimed at the gradual opening of the Spanish economy to the international markets, while preserving the external equilibrium. The system of multiple exchange rates developed under autarky was abolished, and
370
Leandro Prados de la Escosura and Jorge C. Sanz
the peseta was devalued by 43 per cent in an attempt to improve export competitiveness (Donges, 1976: 59). A gradual process of import liberalization was initiated by making flexible the different trade regimes effective under autarky.28 A new tariff in 1960 eliminated quantitative restrictions on 90 per cent of imports and marked the return to the use of ad valorem duties, as opposed to quotas, as a protectionist instrument.29 The new tariff, still a protectionist one, tended to be biased in favour of consumer goods and had a cascading effect on customs duty rates, so effective protection was usually higher than nominal protection. Reforms of the 1960 tariff during the following years reduced effective protection (and its unequal distribution) sharply, from 77.2 per cent in 1962 to 38.4 per cent in 1968 (Bajo and Torres, 1990).30 These measures were complemented by allowing a controlled inflow of foreign long-term capital that, according to Donges (1976), averaged an annual 7 per cent of capital formation over 1960-8.31 The stabilization plan proved most successful after a first year of adjustment, in which the severe measures adopted had a negative impact on economic activity (real income per head fell by 2.7 per cent in 1959) (Prados de la Escosura, 1995). The success of the stabilization plan was probably helped by the ban on free trade unions, which avoided wages claims after the severe devaluation in 1959 (Donges, 1976: 62). Inflation declined from around 12 per cent in 1957-9 to 2.3 per cent for 1959-61 (Comin, 1993). Meanwhile, the (commodity) trade deficit shrank from 7 per cent of GDP in 1957/8 to 3.2 per cent in 1961, and the current account cast a positive balance for 1960 (3.7 per cent of GDP) (de la Dehesa et ai, 1991). The increasing number of tourists and the inflow of foreign capital, together with the return of Spanish capital from abroad, led to the appreciation of the peseta over the official rate of exchange, the disappearance of the black market, and the consequent declaration of external convertibility for the Spanish currency in 1961. As an outcome of the reforms, exports of goods and services experienced a sustained growth over the 1960s and early 1970s, raising their share of GDP from 5.8 per cent in 1959 to 10.7 per cent in 1974. In turn, a higher degree of openness was achieved as exports plus imports grew from 13.9 per cent of GDP to 29.5 per cent during the period (Chamorro et ai, 1975; Prados de la Escosura, 1994). Meanwhile, foreign investment constituted an average 5.8 per cent of fixed capital formation over 1959-74, of which 44 per cent was direct investment, and 74 per cent of it went to manufacturing (Donges, 1976: 108). Since 1964, government intervention in economic activity was reoriented through successive four-year planes de desarrollo (development programmes). The plans, inspired by post-1945 France's planification indicatif, lasted until 1975. The planes allocated a prominent role to privatefirmsas the engine of economic development, but maintained government intervention in economic activity by means of market regulation, public investment and enterprises, and direct intervention in specific economic sectors. Planification indicatifwas implemented through a wide variety of measures, such as public subsidies, tax reliefs, tariff advantages and preferential access to official credit. The planes have been considered a setback for Spain's liberalization and associated with a deceleration in the rate of economic growth (Gonzalez, 1979). It is still debated why the government proceeded so cautiously in dismantling trade barriers and regulations. One explanation is based on the social costs in terms of output and employment derived from a sudden opening up
Growth and macroeconomic performance in Spain, 1939-93
371
(Spitaller and Gali, 1992), which complements another one based on the pressure exerted by interest groups (Fraile, 1993). A last attempt to liberalize the economy took place under the dictatorship in 1970, linked to the preferential agreement with the EEC. However, it was cut short by the first oil shock (1973) and the death of Franco (1975) (de la Dehesa et ai, 1991). In such a context of cautious outward-looking policies under close government supervision, accelerated growth and structural change proceeded in Spain, as a lagged response to the 1950s Western European experience of reconstruction and catching up, prevented by autarky at the time. Spain's economy grew and diversified between 1959 and 1975 (Tables 12.1 and 12.5), while efficiency improved remarkably. Between 1965 and 1975, total factor productivity rose at an annual rate of 3.8 per cent for the whole economy (Suarez, 1992; Myro, 1983), while TFP in agriculture and industry grew at 2 and 6 5 per cent, respectively (San Juan, 1987; Gandoy, 1987).32 A deep reallocation of resources took place. Agricultural share in GDP fell from almost one-quarter of GDP in 1960 to less than 10 per cent by 1975 (that is, at an annual rate of — 6 per cent), while its contribution to employment shrank from over 40 per cent to below a quarter of the total labour force (at — 3.8 per cent yearly), leading to a significant decline in its relative productivity.33 Meanwhile, services reached one-half of GDP, employing two out of everyfiveworkers by 1975. Labour released by agriculture reached 2 million in 1961-70 (Leal et ai, 1975), which to a large extent was absorbed by the 'urban' sector (industry and services), although emigration to Western Europe also represented a decisive outlet.34 However, a significant percentage of the labour force still remained in agriculture, as a comparatively low output per worker in European terms suggests, and as relative labour productivity indicates (Table 12.5).35 Given the disadvantage in output per hectare derived from poorer soil, a lower endowment of land per worker is behind the Spanish agricultural productivity gap (O'Brien and Prados de la Escosura, 1992; Simpson, 1995). Outward orientation of Spanish industrialization remained limited over 1959-75. By 1972, two-thirds of manufacturing output was still produced byfirmswith fewer than five workers, so size remained a major shortcoming for industry's reaping economies of scale and, thus, having access to foreign markets. Manufacturing performance was impressive, growing at over 10 per cent annually, with an increasing role for producer goods industries (Carreras, 1992). However, it depended mostly on home demand (87.6 per cent over 1962-72), with exports accounting for only 22 per cent of the increase, and a negative contribution of import substitution (Donges, 1976: 158, 168). The increase in manufactured exports was due largely to competitiveness (50 per cent) (Donges, 1976:206). Spain exhibited revealed comparative advantage mostly in labour and natural resource-intensive goods, but product-cycle goods were increasingly competitive.36 The Spanish experience tends to suggest, nevertheless, that industries that had followed ISI strategies in the 1940s and 1950s gradually managed to export after 1959 (Donges, 1976). In the 1960s, labour market rigidity represented a real obstacle for an outward shift of labour demand that could not match its supply. In fact, the low level of unemployment (1.7 per cent over 1960-75) did not increase, due to a steady migratory flow towards Western Europe. Migrant remittances contributed to balance Spain's current account. Underemployment remained present due to strict
372
Leandro Prados de la Escosura and Jorge C. Sanz 16 000
0.25
0.20-
1 0.15Labour force,, '/
0.10 -
8 0.05 - • Employment
10 000
0.00
1955
1960
1965
1970
1975
1980
1985
1990
Sources: Ministerio de Economia y Hacienda (MOISSES model).
Figure 12.6 Employment, labour force and the rate of unemployment: Spain, 1955-93
legislation that inhibited layoffs of redundant workers, together with low female participation in the labour force. The resulting pattern of development in the 1960s was, thus, largely dependent on capital accumulation and efficiency gains. Capital deepening was substantial, and fixed capital stock per head grew at 5.7 per cent annually (that is, more than doubling its size), while non-residential capital and machinery and equipment grew at 7.2 and 8.9 per cent per capita, respectively (Hofman, 1993b). Suarez (1992)findsthat, in the period 1965-75, most of real GDP growth is accounted for by total factor productivity gains (58.8 per cent) and capital accumulation (34.7 per cent), while increased labour input accounts for only 6.5 per cent (Table 12.2). In turn, macroeconomic performance played a decisive role in achieving a positive deviation from OECD average growth, most of which was accounted for by the differential money supply and real exports growth (Table 12.4).37 However, inflationary pressure (7.8 per cent annually in 1960-75), together with occasional current account disequilibria,38 constituted a threat that forced the economic authorities to impose episodic stabilization measures.39 In summary, orthodox fiscal policy, near equilibrium in the external sector, and expanding output, coexisting with unreformed factor and product markets and exchange controls, are the main features of the Spanish economy during 1959-75. Shocks and stagflation: the transition from dictatorship to democracy, 1975-85 The oil shocks (1973,1979) coincided in Spain with the end of Franco's regime and the transition to a democratic society. Structural inefficiencies were inherited from
Growth and macroeconomic performance in Spain, 1939-93
373
the Franco era (1939-75), and the post-1959 liberalization had been progressively curtailed by the pressure of interest groups, resulting in a mixture of market and dirigiste economy, whose negative effects would emerge with the supply shocks of the 1970s.40 Thus, the late 1970s opened an unstable political phase in which economic performance was dominated by disinvestment, inflation and job destruction. The oil crisis was aggravated by the fact that Spain was a heavily energy-dependent country,41 whose productive structure was biased towards heavy industry with weak demand. Furthermore, the situation was worsened by the fact that Spain's authorities perceived the oil shortage of 1973 as a temporary shock. The deterioration of the terms of trade, resulting from rising energy import prices, was not followed by a reduction in real wages. In fact, between 1973 and 1976, wages (nominal and real) boomed. Political unrest, initiated in the early 1970s and reaching a peak after Franco's death, led the government to implement compensating policies, such as keeping low energy prices by direct price intervention and nominal wage indexation, thus enhancing the impact of the oil shock and fuelling the current deficit and inflation, which rose as high as 24.5 per cent in 1977. Only after the first democratic elections (June 1977) were adjustment measures introduced (Garcia Delgado, 1990). The Moncloa Agreements represented a set of structural reforms and economic policy measures supported by the consensus of the main political parties and ratified by Parliament. A non-orthodox mix of incomes policies, plus accommodating fiscal and tight monetary policies, were its main features (Spitaller and Galy, 1992). Among those worth mentioning are the fiscal reform, in which progressive wealth and income taxes were included. An active monetary policy aimed at reducing inflation expectations followed.42 A new exchange rate wasfixedfor the peseta (20 per cent lower) in an attempt to improve the external disequilibrium, from which derived short-run positive effects on the current account.43 Trade liberalization resumed in 1977, but the 1979 oil shock, the unstable political climate and the negotiations with the EEC slowed it down (de la Dehesaer ai, 1991). At the same time, income policy agreements attempted to moderate increases in nominal wages.44 However, when the new, democratically elected authorities took steps to adjust to relative price changes resulting from the oil shocks, the lack of flexibility of input markets to reallocate resources acted as a brake. Until the early 1980s,financialmarkets remained regulated and segmented, and exchange controls on capital outflows remained in place, while the still effective Francoist legislation impeded layoffs andflexibilityof contracts in the labour market. Thus, if the set of measures included in the Pactos de la Moncloa are evaluated against economic performance over 1977-81, a favourable balance appears for (a still high) inflation, the external balance and firms' profitability (after wage moderation), while it was not so favourable for the labour market and GDP growth.45 The arrival of the Socialist Party to government in 1982 extended and deepened the reforms initiated with the Moncloa Agreements (1977). In particular, the reform of the industrial sectors most affected by the crisis (steel, textiles and shipbuilding) and the opening of the economy became the top priorities of the new government.46 Financial markets experienced a major deregulation, as interest rates and capital movements were freed. After the Socialist electoral landslide in 1982, a devaluation of the peseta took place, with short-term positive effects on the current account (1984-6), while monetary policy became more restrictive to check the rise in prices.
374
Leandro Prados de la Escosura and Jorge C. Sanz
In turn, the growth of public expenditure and the increases of fiscal pressure represented the main features offiscalpolicy under all the Socialist cabinets (both up to 1986 and afterwards). The growth of public expenditure from 25 per cent of GDP in 1975 to 36 per cent in 1985 was due to the increase in current transfers and public investment, and to the service of a public debt (which rose from 13 to 45 per cent of GDP in the 'transition decade') associated with the introduction of the welfare state, and the costs of industrial reconversion and regional devolution (Gonzalez-Paramo, 1992). Such a major expansion in expenditure, without an offsetting increase in fiscal revenues, resulted in a sustained public deficit. In short, during thefirstthree years of Socialist government, the reduction in the rate of inflation continued (from an average of 16 per cent in 1978-82 down to 10.8 per cent in 1983-5), and external equilibrium was temporarily achieved, while substantial increases took place in unemployment (from 11.6 per cent of the labour force in 1978-82 to 20 per cent in 1983-5) and in the public deficit (from 3.3 to 5.7 per cent of GDP). When the entire transitional period from Franco's death to Spain's EEC membership is considered, some major features emerge. The 1973 and 1979 oil shocks resulted in a deceleration of economic activity that lasted up to the mid-1980s and implied comparative retardation (Table 12.3 and 12.4). In fact, real GDP grew at an annual average rate of only 1.7 per cent between 1975 and 1985 (Table 12.2). Real investment stagnated, while growth depended mainly on private consumption. Job destruction (at —1.6 per cent annually) becomes noticeable when the labour force increase over 1975-85 (0.6 million) is confronted with the 2 million fall in employment. Agriculture had the larger share of unemployed, 1 million workers, against 0.8 and 0.4 million in industry and construction. In turn, capital deepening continued, though at a substantially slower rate than in the years 1960-75. Thus, the stock of physical capital per head grew at 3.3 per cent annually (4.2 per cent for non-residential capital) (Hofman, 1993b). A look at the growth accounting exercise in Table 12.2 confirms a negative contribution of labour to growth (— 0.9 per cent yearly) that could not be offset by that of capital (0.7 per cent). Thus, between 1975 and 1985, employment destruction represented a severe brake on growth, while capital accumulation and, mostly, TFP gains explain it all. However, empirical evidence (Table 12.3 and 12.4) tends to suggest that capital deepening was not carried forward sufficiently, and TFP alone was not enough to close the gap with advanced countries over the 'transition years'. Macroeconomic factors did not contribute to catching up with OECD growth between 1975 and 1985 (Table 12.4). Average inflation reached 15.4 per cent, doubling its rate for 1960-75, while a substantialfiscaldisequilibrium, derived from the action of automatic stabilizers, raised the budget deficit up to an average 3.6 per cent of GDP. Besides, the slow reaction of the Spanish authorities to OPEC shocks contributed to fuel inflation, reduce competitiveness and provoke a rise in the current account deficit (up to an average of —1.1 per cent of GDP in 1976-85).47 When compared to the OECD average, Spanish exports were more labour than capital intensive, and more intensive in physical capital and natural resources than in human capital, while they were mainly composed of goods with weak demand (Martin, 1992). In brief, a still highly protected economy,48 dominated by small firms relying on obsolete techniques, prevented Spanish industry from achieving
Growth and macroeconomic performance in Spain, 1939-93
375
economies of scale and having access to innovation and latest-vintage technology.49 Moreover, economic growth during the 'democratic transition' (1976-85) was severely constrained by overregulated markets. In the labour market, restrictive industrial rules introduced under Franco, which aimed at offsetting the prohibition of independent trade unions by banning layoffs, constituted a major shortcoming for employment creation (Bentolila and Blanchard, 1990). Among Spanish firms most of them small in size, labour intensive and sheltered from competition bankruptcies increased dramatically, while the rate of unemployment rose, with an average rate of 12.8 per cent over the period 1976-85. In fact, the non-accelerating inflation rate of unemployment (NAIRU) jumped from 3.4 per cent in the late 1970s to 18.4 per cent in the early 1980s, to remain at a high level ever since (Dolado and Malo de Molina, 1985).50 When aggregate demand contracted during the crisis years, demand for labour fell and, given institutional constraints, quantity, and not price (wage), adjustments took place, resulting in high unemployment. Sectorial wages in Spain depended on the consumption aggregate wage, and not on sectorial productivity or the unemployment rate. Therefore, unemployed workers could not exert a downward pressure on wages, and unemployment has become a persistent phenomenon (Andres and Garcia, 1992). 7
Recovery of the late 1980s and its legacy: the integration of Spain into the EEC
Spain's admission into the European Community coincided with an expansionary phase of the international economy, and both contributed to a fast economic recovery in Spain. From 1986 to 1990, real GDP grew at an average rate of 4.4 per cent, the outcome of cumulative efforts to fight basic disequilibria associated with integration into the EEC, and of a closer link to the international markets. Positive short-run expectations opened the door to foreign capital and induced a burst of investment that was helped by the increase in profit margins and political stability.51 In fact, since a large technological gap existed between Spain and Western European industrial countries, direct foreign investment brought with it a technological transfer (Vinals, 1992). The positive effects on the labour market were immediate, and the unemployment rate fell from 21.9 per cent in 1985 to 16.3 per cent in 1990.52 In all, employment increased by 3 per cent annually, which represented more than 1.7 million new jobs between 1985 and 1990 (as much as from 1957 to 1974), and job destruction persisted only in agriculture (0.5 million).53 Suarez's (1992) growth accounting exercise reflects this improvement, showing that, from 1986 to 1990, job creation accounted for 38 per cent of real GDP growth, while capital, growing at over 5 per cent annually, contributed 32 per cent (Table 12.2). When placed within the OECD context, Spain overperformed in the late 1980s, with a major contribution from macroeconomics. The control of inflation and the management of money supply, together with a reduced budget deficit, accounted for most of it (Table 12.4). However, although the public deficit experienced a substantial reduction between 1985 (7 per cent of GDP) and 1990 (4 per cent), it persisted over these fiscal expansionary years.54 Nevertheless, other economic disequilibria worsened. After the peseta joined the European Exchange Rate Mechanism (ERM), the competitiveness of Spanish
376
Leandro Prados de la Escosura and Jorge C. Sanz
products declined (as is reflected in the negative contribution of exports growth in Table 12.4). The loss of competitiveness due to an appreciated exchange rate reached up to 30 per cent from the mid-1980s to 1993 (Camarero and Tamarit, 1994). Despite inflation falling to 6.5 per cent over 1985-90, the price differential between Spain and its main trading partners (EEC countries) widened. It has frequently been argued that Spain's differential inflation reflected the lack of competition in non-tradable goods markets.55 The implication of Spain's affiliation to the ERM coexisting with an increasing price differential was an appreciation of the real exchange rate of the peseta.56 However, part of the appreciation of the peseta was due to real factors, such as a productivity increase translated into higher wages, which means that not all the exchange rate appreciation implied a net loss in competitiveness.5 7 In turn, the current account deficit reached — 2 per cent of GDP in the period 1986-90.58 Currently, Spain's integration into the EEC has had a larger impact on imports than on exports. Over 1986-90, while the integration effect accounted for 53 per cent of the increase in Spain's manufacturing imports from the European Community, it contributed to only 28 per cent of the increase in Spanish manufactured exports to the EEC (Martin, 1992).59 However, the external deficit was largely due to the private capital inflow (mostly direct investment) that was required to renew and expand the productive capital stock. In fact, the currently account deficit was more than offset by long-term capital inflows due to its high returns over 1986-91.60 Successful restrictive monetary policy conflicted, however, with exchange rate management. When domestic interest rates rose, the inflow of foreign capital increased (reinforced by the liberalization of capital markets), leading to an increase in the amount of money in circulation.61 The intervention of the Bank of Spain in the foreign exchange markets, with immediate sterilization measures to reduce domestic credit, raised domestic interest rates even more.62 Fiscal policy, in turn, was clearly expansionary, and its pro-cyclical behaviour pushed the public deficit up from 1990. An international recession has taken place in the early 1990s. In fact, sluggish economic activity in most of the OECD countries has shown particularly severe features in Spain. The strain in foreign exchange markets, and changes in the ERM resulting from the 1992-3 'monetary storm', led to three successive devaluations in Spain (up to 20 per cent altogether), which improved competitiveness. The authorities decided, however, to keep the peseta in the ERM, trying to avoid the loss of credibility that might have followed had the government decided to apply a discretionary economic policy.63 Slackening GDP growth (at an average of 0.7 per cent), a substantial public and external deficit ( — 5.6 and —3.3 per cent of GDP, respectively), 5.5 per cent inflation, and job destruction (0.7 million and an unemployment rate reaching 19 per cent of the labour force) are the main features of the early 1990s. High unemployment (over twice the OECD average) represents today a unique feature of Spain's economy within the European Union. In fact, Spain's labour market suggests a case of hysteresis, since transitory shocks, such as changes in relative factor prices in the 1970s, have had permanent unemployment effects. Lack
Growth and macroeconomic performance in Spain, 1939-93
377
of market flexibility due to low labour mobility and high wage indexation are considered to be the main obstacles to reducing unemployment in today's Spain (Andres and Garcia, 1992).64 The indexation of nominal wages implies that anticipated demand shocks do not have any impact on employment. However, wage indexation amplifies the unfavourable impact of supply shocks on employment. In addition, the high costs of layoffs have impeded labourflexibilityand created a weak relationship between wages and the value of marginal labour productivity. Unemployed workers cannot exert enough downward pressure on wages, and as a result, unemployment becomes a persistent phenomenon. It has been argued that not only legal and institutional restrictions affected the lowflexibilityof wages, but other factors, such as firms' recruiting methods, also account for it. Thus, given large differences in labour qualifications, together with the existence of adverse selection and moral hazard,firmsmay offer wages over the market average in an attempt to recruit the best qualified workers, provoking a medium-term push on the average wage, and eventually increasing unemployment. Finally, severe restrictions on regional and sectorial labour mobility are posed by unemployment benefits that reduce incentives to migrate, and by housing market rigidities that make geographical mobility difficult. Recent approaches to the labour market depart from a disequilibrium perspective (Sneessens and Dreze, 1986; Bean and Dreze, 1990), assuming that the labour market is rationed by the installed capacity of thefirm,which effectively restricts the number of workers that can actually be hired in the short run.65 In addition, a firm may be ready to hire more workers at the given real wage, but a restriction on the size of the labour force prevents thefirmfrom doing so. Finally, even when the firm has enough installed capacity and there is enough labour available, there may not be enough demand to absorb it. In the case of Spain, Ballabriga et al. (1991) estimated the distinctive contribution of the different regimes (potential employment, demand-determined employment and labour supply) to explain the observed variation in employment between different periods during 1969-88. Two additional variables were included, labour utilization (whose increase contributes negatively to employment), and a residual term, structural mismatch, which accounts for market rigidities and information costs. The main results for the Spanish case are that, during recessions, restrictions on installed capacity and demand, plus structural mismatch, accounted for unemployment. Moreover, the utilization of labour tended to fall, offsetting the decline in employment.66 This fact seems to imply job hoarding, a feature of labour markets under high adjustment costs (Sargent, 1978). In recent years (1989-91), the economic recession contracted investment levels as a result of firms' pessimistic demand expectations, while labour costs increased relative to capital, reducing the amount of employment required per unit of productive capacity, and the demand restriction operated again (Andres and Garcia, 1992). From Spain's macroeconomic performance and policies in the late 1980s and early 1990s, some lessons can be drawn. The strategy of the Spanish authorities after joining the EMU was to gain credibility through a competitive deflation (Camarero and Tamarit, 1994). To do so, an overvalued real exchange rate was established for the peseta that allowed the purchase of raw materials and intermediate goods at
378
Leandro Prados de la Escosura and Jorge C. Sanz
relatively low prices, aimed at favouring structural change. In turn, economic agents were forced to exercise self-discipline in order to avoid further losses in competitiveness within an open economy. This strategy worked successfully for the tradable goods sector, but not for the one producing non-tradable goods, which fed the inflation differential with the EEC countries. Moderation of wages and financial costs did not take place because of the high interest rates resulting from monetary and exchange rate policies and the public deficit. Thus, the attempt to mitigate the social costs of restructuring the economy, by combining restrictive monetary and incomes policies with fiscal expansion, faces strict limits as high interest rates and their associated effects (appreciation of the real exchange rate and budget and external deficits) cannot be maintained for a long time without putting the country's competitiveness in jeopardy. In turn, market rigidities, which persisted in the post-Franco era and were reinforced in the labour market to keep social unrest at minimum levels, constitute a major inefficiency in Spain, with unemployment at the highest level in Europe.67 In fact, a competitive deflation strategy, trying to improve competitiveness through wage moderation, cannot succeed given Spain's high unemployment, as is shown by the failure of rising unemployment in the early 1990s to have a downward impact on wages.68 8
Concluding remarks
As Spain eventually succeeded in the transition from a highly regulated economy under an interventionist, authoritarian political regime, to an industrial democracy, some lessons from recent Spanish economic history may be of interest for today's developing ex-Communist countries. 1. The autarkic period (1939-59) resulted in high costs in terms of growth. These costs are not restricted to the 1940s, but include the 1950s, a decade of fast international growth and catching up, in which Spain remained relatively isolated and paid a heavy penalty in the form of growth that was unstable and below potential. International isolation, together with resource allocation aside from the market, are responsible, it is widely accepted today, for delayed postwar reconstruction and catching up. 2. Whether deeper and more rapid reforms in the 1960s increased the already impressive rate of growth and catching up is still a debated issue. It can be reasonably argued that faster growth (Table 12.1) might not have been feasible, but instead a more balanced development path, closer to that followed in Western Europe, would have taken place with lower efficiency and social costs.69 In fact, the Spanish experience of the 1960s and the early 1970s seems to question the extent to which factors other than capital accumulation and access to innovation and technology played a major role in Spain's economic growth, once the main constraints of autarky were eliminated. In other words, the allocative inefficiencies resulting from government overregulation and direct intervention in the late 1960s (particularly after the planes de desarrollo were implemented) do not seem to have put a significant brake on growth, as the institutionally oriented historiography has claimed.70 Nevertheless, cumulative inefficiencies during the Franco era constrained further adjustment to international competition in future years.
Growth and macroeconomic performance in Spain, 1939-93 379 3. Liberalization in post-1959 Spain was gradual, and several stages can be distinguished, linked to international agreements which posed no serious threat to Franco's dictatorship. In fact, both after autarky (post-1959) and during the transition to democracy (post-1977), financial stabilization was radical, but structural reform was gradual. These features make the Spanish case of special relevance for countries on their way to industrial democracy and aiming to open up, while maintaining social and political stability. However, the fact that liberalization took place over a very long period of time makes the Spanish experience of limited use for countries in which the transition to a market economy occurs within a democratic context, and where, consequently, a strong social pressure exists to proceed as fast as possible. 4. Once Franco's dictatorship was over and the return to democracy took place, a restrictive monetary policy and high exchange rate were necessary but not sufficient conditions for resuming growth and structural change in Spain under a different set of institutions. A restrictive fiscal policy and moderate labour costs, together with supply-side policies to improve firms' productivity and increase the flexibility of economic sectors, such as microeconomic measures to increase the productivity and competitiveness of the non-tradables sector, are also prerequisites for sustained growth and catching up.71
NOTES This paper is produced as part of a CEPR research programme on 'Comparative Experience of Economic Growth in Postwar Europe', supported by a grant from the Commission of the European Communities under its SPES Programme (no. SPESCT910072). Research for this paper has been funded by the Direction General de Planificacion, Ministerio de Economia y Hacienda. The authors are grateful to Nick Crafts and Gianni Toniolo for their comments on an earlier draft, and to James Simpson for his careful reading of the paper and useful remarks. We acknowledge the access to Pedro Fraile Balbin's work in progress on the political economy of Franco's regime, and to Francisco Comin's and Pablo Martin Acena's unpublished surveys of twentieth century Spain's economic history. The usual disclaimer applies. 1 Less intense than suggested by earlier GDP estimates (Prados de la Escosura, 1995). 2 Pre-Civil War highest GDP level (1935) was reached by 1951. 3 Annual growth rates are computed through exponential fitting. If, instead, the average rate of variation is chosen, the result for the 1940s improves (1.0 per cent, with a standard deviation of 4.4), while it remains unchanged for the 1950s (3.7 per cent, st. dev. 3.4). 4 The 6.3 per cent rate of growth for 1960-73 (Table 12.1) results from a new linkage of national accounts data (Prados de la Escosura, 1994). Official figures cast a lower rate, 5.5 per cent annually. 5 Employment grew at 0.82 per cent annually and labour force at 0.87 per cent between 1955 and 1974. 6 Non-residential capital grew even faster, doubling its ratio to output between 1950 and 1990 (Hofman, 1993b). 7 Thus, machinery and equipment doubled its share, reaching 0.44 of non-residential capital in 1973, although it contracted to 0.39 by 1989. 8 In particular, when machinery and equipment is considered, and Spain compared to the USA, France or Japan.
380
Leandro Prados de la Escosura and Jorge C. Sanz
9 De Long and Summers (1991) predict a growth rate of 3.1 per cent for GDP per worker in Spain between 1960 and 1985, while the actual growth was 3.7 per cent. 10 An alternative estimate of Spain's sources of growth, using an ad hoc framework, has been carried out by Hofman (1993a) for 1950-89, with similar results. The periodization by Suarez (1992) makes his estimates more adequate for our purposes. 11 Minor amendments have been introduced in Suarez (1992) TFP growth figures to allow inputs' contribution to growth and TFP to add up to the GDP growth rate. 12 It was a persistent phenomenon even though Spain performed well above the average up to 1973, and below it thereafter. 13 They rely on previous research by Andres et al. (1993). We acknowledge their permission to use the unpublished results. 14 However, supply-side policies are not directly accounted for in this exercise and should, therefore, be included in the residual. 15 For instance, in 1950, the average years of schooling in Spain per 1000 US dollars (1985 prices PPP) of non-residential capital were one-third of Japan's (Barro and Lee, 1993; Hofman, 1993b; Maddison, 1993). This hypothesis is explored in Prados de la Escosura (1994). 16 The Primo de Rivera's coup d'etat in 1923 opened a dictatorial era that ended in 1930. 17 As Fraile (1993) reminds us, dictatorships are constrained by specific circumstances, incentives and preferences, within which they maximize their utility, conditioning their economic performance. 18 FAO was joined in 1950 when financial links were established with the USA, leading to an economic agreement between the USA and Spain in 1953. Later, in 1955, Spain was admitted as a member of the UN. 19 Catalan (1993) provides a preliminary attempt to derive a series of the real exchange rate for the 1940s and 1950s that confirms the peseta's tendency to be overvalued. 20 From 1940 to 1947 the exchange rate remained unchanged. In 1948, a multiple exchange rate system for activities related to exports, imports, tourism and capital movements was adopted, lasting until 1959 (Donges, 1971; Gonzalez, 1979). A weighted average exchange rate is provided by Serrano Sanz (1992) for 1949-58, using the shares of imports and exports under each exchange rate as weights. 21 It allowed exporters to keep the foreign exchange they obtained from their sales abroad, as a means of payment for future imports. The introduction of a system of export subsidies aimed at improving the competitiveness of Spanish goods also failed in such a context. 22 This hypothesis is tested in Taylor (1994) and Lee (1994). 23 In fact, commodity imports reached an average of 11.7 per cent of GDP for the period 1947-59. In the early 1940s they remained around 5 per cent. 24 A foreign exchange crisis took place in 1959 as foreign reserves fell to $8 million from $58 million the year before, while Spain's committed payments represented $60 million. 25 Machinery and equipment per head more than doubled in the 1950s, growing at 8.2 per cent annually. 26 This procedure, called pignoracion, was frequently used in the last century, and constituted an indirect mechanism for financing the public deficit that would be used as collateral against credit from the Bank of Spam. In fact, at the end of
Growth and macroeconomic performance in Spain, 1939-93 381
27 28 29 30
31 32 33 34
35
36
37 38 39 40 41 42
1958, the issue of automatically pledgable securities ceased (de la Dehesa et a/., 1991). Specific limits were established to rediscount financial assets in the Bank of Spain, and the rate of rediscount was increased to make these specific loans more expensive. There were four different trade regimes (state, bilateral, globalized and liberalized), depending on the degree of control and on the restrictions imposed over the different varieties of imports (de la Dehesa et al.9 1991). Still 20 per cent of imports were ruled by bilateral trade agreements and government agencies. By weighting sectorial rates of protection with the sector's share in output, an ex-post (weighted average) effective rate of protection was derived for manufacturing by Donges (1976: 72, 76) for 1962 and 1968. Manufacturing (excluding food) was nominally protected by 31.3 per cent, and by 68.4 per cent effectively in 1962, while for 1968, in turn, it fell to 23.9 and 31.2, in nominal and effective terms, respectively. Durable consumer goods received the highest nominal and effective protection, whereas machinery and equipment had the lowest. A Stabilization Fund was also created with resources provided by the IMF, the OEEC and the private US banking system, to face disequilibria in the current account without affecting the rate of exchange (Donges, 1976: 59). Kendrick (1990) estimates a 4.2 per cent growth rate for TFP in the period 1960-73. From 0.56 to 0.41 between 1960 and 1975, taking 1 as the average labour productivity in Spain's economy. Spanish emigration to Western Europe (excluding seasonal) was 0.7 million in the same period (1961-70), according to Spanish official sources, but it was over 1.3 million if receiving countries' sources are consulted. For the entire period 1960-75, official estimates recorded afigureof 1.1 million permanent (non-seasonal) migrants to Europe. In 1975, Spanish final agricultural output per worker represented, in purchasing power parity terms, 29.9 per cent of the UK's, 44.5 per cent of France's, and 16.1 per cent of the USA's (Prasada Rao, 1993), despite the significant catching up experienced over the period 1960-75 (O'Brien and Prados de la Escosura, 1992). Such as leather and wooden goods, footwear, clothing, furniture, pottery and glass, minerals and metal manufactures, on the one hand, and chemical fertilizers, plastic, rubber and paper goods, and transport equipment, on the other (Donges, 1980). According to Andres et al. (1994), only Italian macroeconomic performance has a comparable explanatory power of differential growth. However, for the period 1960-75, the current account deficit only represented on average -0.6 per cent of GDP. In turn, the commodity trade balance and the trade balance (including services) represented — 6.8 and — 5.5 per cent of GDP. Adjustments were introduced in 1967,1970-1 and 1975 (stop-and-go policy) (de la Dehesa et ai, 1991). An idea of misallocation of resources in the post-1959 years is provided by Donges' (1976:225) estimates of the high opportunity cost of domestic resources in Spanish industry. Up to 75 per cent of its energy consumption was imported (de la Dehesa et al, 1991: 158). As opposed to the traditional passive monetary policy followed since the Franco years, which adjusted to the public sector need for resources and led to high rates of inflation.
382 Leandro Prados de la Escosura and Jorge C. Sanz 43 In 1978 and 1979, the negative trade balance fell sharply and the deficit on current account disappeared. 44 Pay increases were defined as a function of expected inflation, instead of last year's inflation. This measure, despite being a great step forward to break the traditional wage bargaining mechanism, would become a big obstacle to fighting inflation in the 1980s. In fact, nominal labour costs grew steadily at a higher rate than inflation. 45 A brief sketch of the delayed Spanish reaction to the oil shocks could be as follows. When the relative price of energy and raw materials rose,firmsreacted by reallocating their inputs so as to increase their marginal productivity. The fall in the demand for energy and raw materials reduced labour marginal productivity, andfirmshad to compensate through a reduction in their number of employees at a given real wage. The contraction of the demand for energy, raw materials and labour also reduced capital marginal productivity, leading firms to reduce their volume of investment. As a consequence, the change in relative prices of energy and raw materials provoked a contraction in aggregate supply, which reduced real output growth and raised the rate of inflation. Workers suffered an important fall in real wages as a result of increases in prices, and reacted by demanding higher nominal wages. As a consequence of higher input costs (raw materials, energy and labour) and the fall in internal demand (as unemployment increased), profitability declined and firms reacted by reducing investment and employment demand once again. Somefirmstried to avoid the economic crisis with new credits (thanks to the negative real interest rates, due to moderate nominal interest rates coexisting with high rates of inflation). Eventually, the measures of monetary restriction and the new attitude of lenders in a context of high inflation provoked an increase in real interest rates which produced a financial crisis in many firms. 46 Which was reinforced by the authorities' decision to keep Spain in the NATO military structure after joining (1981) and to accelerate economic integration into the European Community, achieved in 1986, which was widely viewed in Spain as the only means to reach political stability and to secure democracy (de la Dehesa et a/., 1991). 47 Meanwhile, the commodity trade deficit and the (total) trade deficit represented an average —5.4 and —4.1 per cent of GDP, respectively. 48 The average effective protection in 1980 was 33.7 per cent (Bajo and Torres, 1990). 49 In fact, by 1985, 80 per cent of firms still had fewer than ten workers, and R & D represented 1.1 per cent of manufacturing value added, against 4.6 per cent in the EEC (Martin, 1992). 50 De la Dehesa et al. (1991) provide alternative NAIRU estimates for 1974^9,6.6 per cent, and 1980-4,11.3 per cent. Crafts (1993), in turn, suggests that NAIRU reached 9.7 per cent in 1969-79, and 15 per cent in 1980-8. 51 The yearly increase in real investment reached a peak of 15 per cent in 1988. 52 For the years 1986-90, the unemployment rate represented, however, an average of 19.1 per cent. 53 Employment grew faster than active population (1.1 million). Labour creation in 1985-90 represented 0.3, 0.4 and 1.5 million in industry, construction and services, respectively. 54 An average public deficit of 3.8 per cent of GDP for 1986-90. 55 In fact, given their lower productivity, because of a lower capital/labour ratio and similar wage levels, non-tradables (and services, in particular) reached higher prices.
Growth and macroeconomic performance in Spain, 1939-93
383
56 Spanish relative prices rose both because of structural change and internal imbalances, i.e. the impact of non-tradables on tradable goods, the small size and backward technologies of firms, and labour market rigidities. 57 Factors that favoured an appreciation of the peseta were the inflow of capital and the terms of trade gain derived from a fall in oil prices, while the reduction of tariff barriers with entry into the EC and changes in consumer preferences towards foreign goods tended to cancel it out. 58 Over — 5 per cent of GDP for the trade balance. 59 Martin (1992:156) found that in the years 1986-90 final demand contributed 35 per cent and competitiveness 12 per cent to the increase in Spain's imports of manufactured goods from the EEC. In the case of Spanish manufactured exports to the EEC, Common Market demand provided 72 per cent, while competitiveness made a negative contribution. The overall increase in manufactured exports was only 38 per cent of that for manufactured imports. 60 Long-term capital inflows represented 2.8 per cent of GDP over 1986-91, and direct investment, 1.4 per cent. Meanwhile, the current account deficit reached —1.5 per cent and the trade deficit, —5.2 per cent. 61 Plus an overvaluation of the peseta and negative consequences on the current account. 62 In 1987 the targets for the monetary aggregate were revised, and some years later certain less orthodox measures were implemented. For example, in 1989 the Bank of Spain made explicit recommendations to commercial banks to impose credit rationing, and from 1987 to 1989 the control of exchange became a frequent policy. 63 The decision to stick to the ERM, in contrast to the cases of Italy and the UK, might also be related to the negative experience of the 1890s, when Spain gave up the gold convertibility of the peseta (Martin Acena, 1993). The poorer performance of Spain, as compared to these two other countries, deserves a thorough investigation (cf. Camarero and Tamarit, 1994). 64 This paragraph draws heavily on Andres and Garcia (1992). 65 Thus, potential employment is the number of workers that corresponds to the full utilization of the installed productivity capacity. 66 From the quantitative exercise it emerges that, in the late Franco years (1969-70 to 1971-4), the levels of installed capacity (0.6 per cent), demand (0.3 per cent) and labour availability (1.3 per cent) accounted for all observed employment growth (2.0 per cent), although structural mismatch ( — 0.4 per cent) and a high labour utilization ( — 0.6 per cent) reduced the explained employment growth (1.2 per cent). After the first oil shock and during the 'transition' years (1971-4 to 1975-82) there were negative effects on employment derived from restrictions in the installed capacity (— 2.1 per cent), in demand ( — 4.8 per cent) and in structural mismatch (—1.9 per cent), while positive effects were due to the fall in labour utilization (1.9) and supply (0.1), which together account for all the observed employment destruction ( — 7.7 per cent). A similar picture is obtained for the post-1979 oil shock (1975-82 to 1983-6), in which the negative contribution of capacity utilization ( — 5.6 per cent), demand ( — 5.9 per cent) and structural mismatch ( — 3.3 per cent) increased, hardly offset by the decline in labour utilization (0.6 per cent) and supply (0.1 per cent), amounting to all the observed decline in employment (—12.9 per cent). Finally, over the 1980s (1983-6 to 1987-8) positive contributions to observed employment growth (3.8 per cent) are derived from recovery of the levels of installed capacity (1.5 per cent), demand (2.5 per cent) and supply (0.2
384 Leandro Prados de la Escosura and Jorge C. Sanz
67
68 69 70
71
per cent), while the rise in labour utilization played a negative part ( — 0.9 per cent). There is an ongoing debate about the reliability and feasibility of unemployment figures for Spain, as the informal sector represents a non-negligible share of total economic activity. Adjusting for employment in the informal sector would probably cast a lower unemployment rate that, nevertheless, would have increased significantly since 1976. In fact, Spain's unit labour costs have increased to those in the EEC. The comparison of industrial performance in Spain and Italy seems to reinforce the argument (cf. Carreras, 1987, 1992). As Blanchard (1993: 232) recently put it, 'when one looks at postwar growth rates, the basic impression is that all countries had impressively high growth, no matter what strategy was being pursued'. If the post-1959 Spanish fast growth is viewed as a delayed postwar recovery, that contention would apply to Spain. In particular, the reform of the services sector (transportation, health, education), which is linked to the reform of public administration.
REFERENCES Anderson, C.W. (1970) The Political Economy of Modern Spain: Policy Making in an Authoritarian Regime, Madison, WI: University of Wisconsin Press. Andres, J. and J. Garcia (1992) Trincipales rasgos del mercado de trabajo espanol ante 1992', in J. Vinals (ed.), La economia espanola ante el Mercado Unico europeo: Las claves del proceso de integration, Madrid: Alianza. Andres, J., J.E. Bosca and R. Domenech (1994) 'Main patterns of economic growth in OECD countries', Direccion General de Planificacion Working Paper D-94001, Ministerio de Economia y Hacienda, Madrid. Andres, J., R. Domenech and C. Molinas (1993) 'Growth, convergence and macroeconomic performance in OECD countries: a closer look', Direccion General de Planificacion Working Paper D-93003, Ministerio de Economia y Hacienda, Madrid. Bajo, O. and A. Torres (1990) 'Estructura y caracteristicas de la protection en Espana', Information Comercial Espanola, no. 687, pp. 13-123. Ballabriga, F.C., C. Molinas, M. Sebastian and A. Zabalza (1991) 'Las restricciones de demanda y de capital en la economia espanola: 1964-1988', in C. Molinas, M. Sebastian and A. Zabalza (eds.), La economia espanola: Una perspectiva macroeconomica, Madrid: A. Bosch/IEF. Barciela, C. (1986) 'Introduction a "Los costes del franquismo en el sector agrario"', in R. Garrabou, C. Barciela and J.I. Jimenez Blanco (eds.), Historia agraria de la Espana contempordnea, vol. Ill, Barcelona: Critica. Barro, RJ. and J.W. Lee (1993) 'International comparisons of educational attainment', NBER Working Paper No. 4349. Baumol, W.J. (1986) 'Productivity growth, convergence and welfare: what the long-run data show', American Economic Review, 76 (5), pp. 1072-85. Bean, C. and J. Dreze (1990) 'European unemployment problem: lessons from a multicountry study', introduction to C. Bean and J. Dreze (eds.), Europe's Unemployment Problem, Boston, MA: MIT Press. Bentolila, S. and J. Blanchard (1990) 'Spanish unemployment', Economic Policy, 10, pp. 234-81. Blanchard, O. (1993) 'Panel discussion: lessons for Eastern Europe today', in R. Dornbusch, W. Nolling and R. Layard (eds.), Postwar Economic Reconstruction
Growth and macroeconomic performance in Spain, 1939-93
385
and Lessons for the East Today, Cambridge, MA: MIT Press. Camarero, M. and C. Tamarit (1994) 'La peseta en el SME: de la fortaleza a la crisis', Information Comercial Espanola, no. 728, pp. 145-59. Carreras, A. (1987) 'La industrial atraso y modernization', in J. Nadal, A. Carreras and C. Sudria (eds.), La economia espanola en el siglo XX: Una perspectiva historica, Barcelona: Ariel. (1992) 'La production industrial en el muy largo plazo: una comparacion entre Italia y Espana de 1861 a 1980', in L. Prados de la Escosura and V. Zamagni (eds.), El desarrollo economico en la Europa del sur: Espana e Italia en perspectiva historica, Madrid: Alianza. Catalan, J. (1992) 'Reconstruction, politica economica y desarrollo industrial: tres economias del sur de Europa, 1944-1953', in L. Prados de la Escosura and V. Zamagni (eds.), El desarrollo economico en la Europa del sur: Espana e Italia en perspectiva historica, Madrid: Alianza. (1993) 'Economia e industria: la ruptura de la posguerra en perspectiva comparada', Revista de Historia Industrial, no. 4, pp. 111-43. Chamorro, S., R. Commendador, J.J. Dolado, R. Repullo and J. Rodriguez (1975) 'Las balanzas de pagos de Espana del periodo de la autarquia', Information Comercial Espanola, no. 502, pp. 161-87. Comin, F. (1987) 'La economia espanola en el periodo de entreguerras (1919-1935)', in J. Nadal, A. Carreras and C. Sudria (eds.), La economia espanola en el siglo XX: Una perspectiva historica, Barcelona: Ariel. (1992) 'Una reconstruction economica diferente en la Espana de la postguerra', Estudis d'Historia Economica, no. 2, pp. 63-78. (1993) 'Estado y crecimiento economico en Espana: lecciones de la historia', Papeles de Economia Espanola, no. 57, pp. 32-56. (1994) 'Las etapas de la economia espanola (1923-1993)', unpublished manuscript. Crafts, N.F.R. (1993) 'Was the Thatcher experiment worth it? British economic growth in a European context', in A. Szirmai, B. van Ark and D. Pilat (eds.), Explaining Economic Growth: Essays in Honour of Angus Maddison, Amsterdam: North-Holland. Dehesa, G. de la, J.J. Ruiz and A. Torres (1991) 'Spain', in D. Papageorgiou, M. Michaely and A.M. Choski (eds.), Liberalizing Foreign Trade: The Experience of New Zealand, Spain and Turkey, Cambridge, MA: Blackwell. De Long, J.B. and B. Eichengreen (1993) 'The Marshall Plan: history's most successful structural adjustment program', in R. Dornbusch, W. Nolling and R. Layard, (eds.), Postwar Economic Reconstruction and Lessonsfor the East Today, Cambridge, MA: MIT Press. De Long, B. and L. Summers (1991) 'Equipment investment and economic growth', Quarterly Journal of Economics, 106 (2), pp. 445-502. Dolado, J.J. and J.L. Malo de Molina (1985) 'Desempleo y rigidez del mercado de trabajo en Espana', Boletin Economico del Banco de Espana, September, pp. 26-36. Donges, J.B. (1971) 'From an autarchic towards a cautiously outward-looking industrialization policy: the case of Spain', Weltwirtschaftliches Archiv, 107 (1), pp. 33-75. (1976) La industrialization en Espana: Politicas, logros, perspectivas, Barcelona: Oikos Tau. (1980) 'The Spanish industry in face of its integration into the European community', Economia Internazionale, 33 (4), pp. 399-415. Dowrick, S. and D.-T. Nguyen (1989) 'OECD comparative economic growth 1950-85: catch-up and convergence', American Economic Review, 79 (5), pp. 101O-30.
386
Leandro Prados de la Escosura and Jorge C. Sanz
Dumke, R. (1990) 'Reassessing the Wirtschaftswunder: reconstruction and postwar growth in West Germany in an international context', Oxford Bulletin of Economics and Statistics, 52 (2), pp. 451-91. Fraile, P. Balbin (1991) Industrializacion y grupos de presion: Le economia politica de la proteccion en Espana 1900-1950, Madrid: Alianza. (1993) 'Regulatory capture in peripheral Europe: the Spanish steel industry, 1941-1981 - an institutional analysis', unpublished manuscript, Universidad Carlos III. Fuentes Quintana, E. (1984) 'El Plan de Estabilizacion economica de 1959, veinticinco anos despues', lnformacion Comercial Espanola, nos. 612-13, pp. 25-40. (1989) Tres decenios de la economia espanola en perspectiva', in J.L. Garcia Delgado (ed.), Espana, Economia, Madrid: Espasa Calpe. Gandoy, R. (1987) 'Production y productividad en la industria espanola 1964-1981', Economia Industrial, no. 256, pp. 33-46. Garcia Delgado, J.L. (ed.) (1990) Economia espanola de la transicion y la democracia, Madrid: CIS. Garcia Santos, N. and P. Martin Acefia (1990) 'El comportamiento del gasto publico en Espana durante la Segunda Republica, 1931-1935', Revista de historia Economica, 8 (2), pp. 397-415. Gonzalez, M.J. (1979) La Economia politica del franquismo 1940-1970: Dirigismo, mercado y planificacion, Madrid: Tecnos. (1989/90) 'La autarquia economica bajo el regimen del General Franco: una vision desde la teoria de los derechos de propiedad', lnformacion Comercial Espanola, 676-7, pp. 19-31. Gonzalez-Paramo, J.M. (1992) 'El papel del sector publico espanol en el proceso de integracion economica en Europa', in J. Vinals (ed.), La economia espanola ante el Mercado Unico europeo: Las claves del proceso de integracion, Madrid: Alianza. Hernandez Andreu, J. (1980) Depresion economica en Espana 1925-1934: Crisis mundial antes de le Guerra Civil, Madrid: Instituto de Estudios Fiscales. (1986) Espana y la crisis de 1929, Madrid: Espasa Calpe. Hofman, A. (1993a) 'Long run growth in Spain and smaller Latin American countries: a comparative perspective', Proceedings of the SPES/EHES Conference on Long-run Economic Growth in the European Periphery, Coruna. (1993b) 'The capital stock of Spain in the 20th century', Proceedings of the SPES/EHES Conference on Long-run Economic Growth in the European Periphery, Coruna. Kendrick, J.W. (1990) 'International comparisons of productivity trends and levels', Discussion Paper No. 90-02, George Washington University. Leal, J.L., J. Leguina, J.M. Naredo and L. Tarrafeta (1975) La agricultura en el desarrollo capitalista espanol (1940-1970), Madrid: Siglo XXI. Lee, J.W. (1994) 'Capital goods imports and long-run growth', NBER Working Paper No. 4725. Lopez Garcia S. (1991) 'La organization de la investigation cientifica y tecnica tras la Guerra Civil: contrates y similitudes con los logros de las primeras decadas del siglo XX', Proceedings of Encuentro de Historia Economica, UIMP, Valencia. Maddison, A. (1993) 'Stime standardizzate dello stock di capitale: un confronto fra sei paesi', Innovazione e Materie Prime, no. 1, pp. 20-47. Malefakis, E. (1987) 'La economia espanola y la guerra civil', in J. Nadal, A. Carreras and C. Sudria (eds.), La economia espanola en el siglo XX: Una perspectiva historica, Barcelona: Ariel. Martin, C. (1992) 'El comercio industrial espanol ante el mercado unico europeo', in J. Vinals (ed.), La economia espanola ante el Mercado Unico europeo: Las claves
Growth and macroeconomic performance in Spain, 1939-93
387
del proceso de integration, Madrid: Alianza.
Martin Acena, P. (1993) 'Spain during the classical gold standard years, 1880-1914', in M.D. Bordo and F. Capie (eds.), Monetary Regimes in Transition, Cambridge: Cambridge University Press. (1994) 'La historiografia economica espanola en el siglo XX', in M. Palacios (ed.), Las ciencias sociales en la historiografia de lengua espanola, Bogota: Siglo XXL
Martin Acena, P. and F. Comin (1984) 'La politica monetaria y fiscal durante la Dictadura y la Segunda Republica', Papeles de Economia Espanola, no. 20, pp. 236-61. (1991) INI: 50 ahos de industrialization en Espana, Madrid: Espase Calpe.
Myro, R. (1983) 'La evolution de la productividad global de la economia espanola en el periodo 1965-1981', Information Comercial Espanola, no. 594, pp. 115-27. O'Brien, P.K. and L. Prados de la Escosura (1992) 'Agricultural productivity and European industrialization, 1890-1980', Economic History Review, 45 (4), pp. 514-36. Palafox, J. (1986) 'La politica presupuestaria de la Dictadura de Primo de Rivera: 6 una reconsideration necesaria?', Revista de Historia Economica, 4 (2), pp. 389-410. (1991) Atraso y democracia: La Segunda Republica y la economia espanola,
1892-1936, Barcelona: Critica. Prados de la Escosura, L. (1994) 'Growth and structural change in Spain, 19th-20th centuries', unpublished paper, Universidad Carlos III. (1995) Spain's Gross Domestic Product, 1850-1993: Quantitative Conjectures,
Universidad Carlos III Working Paper No. 95/05. (forthcoming) From Empire to Nation: Growth and Retardation in Spain, 1780-1930,
Cambridge: Cambridge University Press. Prasada Rao, D.S. (1993) Intercountry Comparisons of Agricultural Output and
Productivity, FAO Economic and Social Development Paper 112, Rome. San Juan, C. (1987) Eficiencia y rentabilidad de la agricultura espanola, Ministerio de
Agricultura, Madrid. Sargent, T.J. (1978) 'Estimation of dynamic labour demand schedules under rational expectations', Journal of Political Economy, 86 (6), pp. 463-99. Serrano Sanz, J.M. (1992) 'La apertura exterior de la economia espanola en perspectiva (1901-1980)', in J.L. Garcia Delgado (ed.), Economia espanola, cultura y sociedad: Homehnaje a Juan Velarde Fuertes, vol. I, Madrid: Eudema.
Simpson, J. (1995) 'Spanish agriculture: the long siesta, 1765-1965', unpublished manuscript, Universidad Carlos III. Sneessens, H.R. and J. Dreze (1986) 'A discussion of Belgian unemployment, combining traditional concepts and disequilibrium economies', Economica, 53, no. 210(S), pp. S89-S119. Soto, A. (1989) El trabajo industrial en la Espana contempordnea
(1874-1936),
Barcelona: Anthropos. Spitaller, E. and M. Galy (1992) 'Spain: landmarks in economic development, 1939-1992', IMF Working Paper 92/78. Suarez, F.J. (1992) 'Economias de escala, poder de mercado y externalidades: medicion de las fuentes del crecimiento espanol', Investigaciones Economicas, 2nd series, 16(3), pp. 411-41. Taylor, A.M. (1994) Tres fases del crecimiento economico argentino', Revista de Historia Economia, 12(3), pp. 649-83.
Vinals, J. (1992) 'La economia espanola ante el Mercado Unico: las claves del proceso de integration en la Comunidad Europea', in J. Vinals (ed.), La economia espanola ante el Mercado Unico europeo: Las claves del proceso de
integration, Alianza: Madrid.
13 Irish economic growth, 1945-88 CORMAC 6 GRADA AND KEVIN O'ROURKE
1
Introduction
Economic growth has made a welcome return to the headlines in the economics profession. Technical advances have made it possible to model the growth process rigorously in the presence of externalities and increasing returns, and this has spawned a vast literature formalizing many of the intuitions about growth previously held by applied economists and economic historians. Whether or not there are diminishing returns to capital, it seems clear that there are diminishing returns to theory, in this field at least. Theory urgently needs to be supplemented with empirical work, be it multicountry regressions or case studies, if thefieldis to retain its present vigour. We believe that Irish economic history offers many potential lessons for students of economic growth. While Dowrick and Nguyen (1989), Mankiw et al. (1992) and Freeman (1989) incorporate Ireland1 in their work, Ireland is excluded from consideration in well-known studies such as those by Calmfors and Drifrlll (1988), Crafts (1992) and Barro and Sala-i-Martin (1991). The last-mentioned (p. 151) go so far as to exclude Northern Ireland from their study on the grounds that it is 4a substantial outlier for the United Kingdom'! This seems to us mistaken: many of the theoretical issues explored by growth economists have resonances in the Irish experience. In this century, Ireland has swung from extreme protectionism to extreme openness. Its development strategy has at times favoured the exploitation of its comparative advantage in agriculture, at other times has been based on import substitution, and most recently has relied on the capital inflows and technology transfer associated with (heavily subsidized) foreign multinationals. It is a country that is peripheral in relation to Europe as a whole, but which has long had intimate links with the UK, one of Europe's largest economies. Not only commodities but labour and capital have been extremely mobile between Ireland and the rest of the world. Above all, there is a widespread perception that independent Ireland 'blew it', not least in Ireland. Two influential books recently published, Joseph Lee's Ireland 1912-1985: Politics and Society and The Economic Development of Ireland in the 388
Irish economic growth, 1945-88
389
Twentieth Century by Kieran Kennedy, Thomas Giblin and Deirdre McHugh, have made the point forcefully. Using Louis Cullen's estimate of Irish national income per capita in 1911, they have argued that, while Ireland was a respectably wealthy country on the eve of World War I, its subsequent growth experience was disastrous, with the result that it slipped dramatically down the European income league. Kennedy et al. make their point largely with reference to growth rates in the UK. Over the period 1926-85, GNP grew at the same rate, 2.1 per cent p.a., in both countries, while per-capita product grew at 1.8 per cent p.a. in Ireland, and 1.7 per cent p.a. in the UK. They then point out that this performance is unsatisfactory, for two reasons. First, per-capita incomes in Ireland were lower than in the UK over this period, so Ireland should have been catching up rather than merely keeping pace with the UK. Second, British growth rates were low by European standards, and so keeping pace with the UK ensured long-run decline with respect to the Continent (Kennedy et al, 1988: table 6.1, pp. 118-21). Lee is less concerned with the UK, and more concerned with Europe. 'Ireland recorded the slowest growth of per-capita income between 1910 and 1970 of any European country except the United Kingdom . . . Ireland slid from being a reasonably representative western European economy, in terms of income per head, at the time of independence, to a position far below the western European average in 1970/ Moreover, 'Income per capita is itself an indulgent criterion by which to assess Irish achievement. Precisely because of the unique Irish population performance, a wide gap opened between developments at the individual and the national levels . . . No other European country, east or west, north or south, for which remotely reliable evidence exists, has recorded so slow a rate of growth of national income in the twentieth century' (Lee, 1989: 514-15). To summarize, recent Irish authors have tended to go beyond simple Irish-British comparisons, and made the following points: Ireland kept pace with the UK; given that the UK was the sick man of Europe, this implied that Ireland did not keep pace with Europe; Ireland should have grown more rapidly than the UK, given that it was poorer. This chapter addresses Ireland's growth experience between the 1940s and the present. Its comparative statistical appraisal focuses on the years since 1950, the period for which both the Penn Table Mark 5 and the OECD provide data.2 It does so in the context of growth theory new and old, and in the spirit of the survey of recent literature on productivity growth by Crafts (1992). How bad was Ireland's relative economic performance, and what can explain it? While the chapter may not provide definite answers, we hope to convince our audience that the questions are worth asking and that Ireland is a country worth studying. 2
Irish growth and the economic convergence debate
In the debate about the empirical relevance of the new growth literature, much prominence has been given to the question of convergence. Do countries with initially lower levels of GDP per capita grow more quickly, as the Solow model suggests; or can divergence rather than convergence occur over time? Standard models suggest that convergence can occur in many ways. First, there is the
390
Cormac 0 Grada and Kevin O'Rourke
closed-economy Solow intuition: as capital accumulates and capital-labour ratios rise, the marginal product of capital falls. This on its own implies convergence (although, of course, if countries differ in various ways, they may converge to different steady states). In addition, there is a host of open economy forces which should strengthen this process, irrespective of differences between countries. We suspect that these open economy forces may in practice be more important. Poor, low-wage countries will attract capital and technology inflows, which boost labour productivity and living standards; and they will experience emigration, which improves the lot of those who stay at home. Furthermore, commodity trade may lead to factor prices converging internationally, and probably did in the late nineteenth century (O'Rourke and Williamson, 1992).3 On the other hand, if emigration involves a loss of human capital, or of the best workers within a given skill bracket, this can lead to capital outflows and divergence.4 In the absence of instantaneous technological diffusion, higher levels of technology in rich countries may also imply capital flows from poor to rich countries (Barro and Sala-i-Martin, 1991). Moreover, the original Solow intuition does not apply in the absence of diminishing returns, and the new growth theorists have argued forcefully that constant or increasing returns may be the rule rather than the exception. Empirical investigations have generally found that catch-up is a feature of economic growth, at least among OECD economies (Baumol et ai, 1989; Dowrick and Nguyen, 1989). The most influential studies of convergence may have been Barro and Sala-i-Martin's papers documenting convergence among US states and European regions, showing convergence to be a robust feature of these economies, occurring at the annual rate of about 2 percent (Barro and Sala-i-Martin,1991). However, while some less developed countries have grown rapidly in recent years, whether the entire globe forms a single convergence club is a controversial matter (Mankiw et a/., 1992; Durlauf and Johnson, 1992). This may mirror Williamson's early study of regional convergence within countries, which identifies a pattern of increasing regional inequality early in the growth process, and increasing equality later on (Williamson, 1965). Ireland is an OECD member, and has been an EC member since 1973. It is one of the most open economies in the world, with a very high trade-to-GDP ratio, policies designed to attract direct foreign investment, and strong links with international labour markets. By rights, then, it should have participated in the OECD convergence experience. Has it? First, a word is in order about what measure of economic growth is most appropriate for the issues we want to address. Economists might legitimately be interested in either productivity growth or the growth in living standards. In the former case, GDP per employed worker is the most reasonable measure to look at. In the latter case, GNP per capita is what matters. When the data are available, we examine trends in these two variables; however, for the most part, data availability dictates that we examine GDP per capita. As we will see, the distinction becomes potentially important in the post-1973 period. Previous studies which have examined Ireland's postwar economic performance have beenflawedin a number of respects. First, while they mention the experience of other peripheral European economies, they typically rely solely on average growth
Irish economic growth, 1945-88
391
Table 13.1. Economic growth in Ireland, the UK and Europe, 1950-88
Ireland
n-s (a) Growth 1950-60 1960-73 1973-88 1950-88 (b) Growth 1950-60 1960-73 1973-88 1950-88
UK M
ii^s
Western Europe M
*Ts
M
rates of GDP per capita (average annual percentage growth rates) 2.15 2.19 2.51 2.46 4.73 4.75 4.21 3.71 2.68 2.61 4.34 4.19 0.85 2.68 1.94 1.89 1.91 1.90 2.33 2.90 2.34 2.29 3.48 3.43 rates of GDP (average annual percentage growth rates) 1.67 1.71 2.89 2.87 5.64 5.64 4.87 4.36 3.22 3.17 5.20 5.05 1.87 3.65 2.03 1.99 2.25 2.24 2.83 3.38 2.66 2.62 4.14 4.09
Note: H-S, Heston and Summers; M, Maddison.
rates in Ireland and other countries, which does not give sufficient weight to the recent literature on convergence.5 Second, the choice of incomefiguresmatters; as is well known, difficulties arise in making national income estimates comparable across countries and time. The Heston-Summers and Maddison/OECD data sets represent the state of the art in this regard, and we now summarize the picture given by the two data sets. Unfortunately, it matters which data set you use. Table 13.1 gives evidence from the Penn Table Mark 5, as well as from Maddison's most recent data set. Average annual growth rates in GDP and GDP per capita are given for Ireland, the UK and the rest of Western Europe.6 GDP per capita is what is most relevant to the convergence debate, and on this criterion, Ireland's performance has been poor, especially according to Heston and Summers. Over the period 1950-88, Ireland's annual growth rate was only two-thirds of the Western European average, and a fraction less than the equally anaemic British rate. The period 1960-73 was a relatively good one for Ireland, when it easily outperformed the British economy, but even then Irish growth rates were somewhat lower than their European counterparts. The other two subperiods were disastrous, with Irish growth rates less than half as high as on the Continent, and substantially lower than in the UK. Note in particular the very low growth rate after 1973. While the oil shocks reduced growth worldwide, the relative Irish performance during this period was (according to the Heston-Summersfigures)even worse than during the 1950s. On the face of it, this challenges the conventional wisdom that EC membership has helped Ireland catch up with the rest of Europe during the past twenty years (although Ireland might, of course, have done a lot worse had it not been an EC member during this period). The Maddison data give a much more optimistic story, with Ireland doing better than the UK over the period as a whole (although it still does less well than Western Europe). The big difference between the two data sets lies in the post-1973 period: Maddison has Ireland outperforming Western Europe during this time. The reason for this can be easily seen in Figure 13.1, which plots the Irish GDP/capita indices
392
Cormac 0 Grada and Kevin O'Rourke
1955
1960
1965
1970
1975
1980
1985
Figure 13.1 GDP per capita indices: Ireland, 1950-88 (Maddison and Heston-Summers data sets)
from the two data sets. The two indices tell broadly similar stories until 1978, after which time the Maddison estimates continue growing according to trend, whereas the Heston-Summers index stagnates. We will have more to say about this discrepancy in what follows. The figures on GDP growth rates tell a broadly similar story, the only difference being that over the period 1950-88 as a whole, GDP grew slightly more rapidly in Ireland than in the UK, even according to Heston and Summers. Again, however, the Irish growth performance was well below that in Europe throughout the whole of the period, for both data sets. Only the Irish tendency to use the UK as a yardstick for performance has disguised Ireland's relative failure during this period. Moreover, these tables giving average growth rates conceal important information which makes Ireland's relative performance look much worse. In 1950, according to Heston-Summers data, Irish GDP per capita was 48 per cent lower than in the UK, and 15 per cent lower than in Western Europe. It is the case that convergence has been a feature of the OECD economies: Ireland should therefore have grown more rapidly than Western Europe as a whole. That it did not makes Ireland an important outlier. This outlier status appears most clearly when one examines graphs plotting GDP per capita growth rates against initial GDP per capita levels. Figure 13.2 does this for 1950-88,1950-60,1960-73 and 1973-88, using both the Heston-Summers and Maddison data. Several things emerge at once from these graphs. First, the familiar OECD convergence story emerges clearly, the strictures of Milton Friedman notwithstanding.7 The only subperiod in which a clearly visible negative correlation between growth and initial income is absent is 1973-88 (although it is weak during the 1950s). This could be due to a number of factors. The oil shocks and associated policy responses made this a turbulent time, of course; it could be that the short-run effects of different macroeconomic policies swamped any longer-run forces during this period. Or it could be that the forces which normally lead to convergence are absent during downturns, as was certainly the case during the interwar period
g
«.
1
s o-
d
o
*+
II
DO
.n»
1
|| O. 2L
^11
no rj o 3rl
3 II
S-3°
3 2 EL rp a* : G- o i
§8-
i
Average annual growth rate, 1973-88
3 S 2
Average annual growth rate, 1973-88 3 O O
® §
N)
U
A
Ul K
N)
-'
©
§ § § ! § ! § !
o
ppppppppppp >-O
S
Average annual growth rate, 1950-60
p o p o o p © ©
Average annual growth rate, 1950-60
Average annual growth rate, 1960-73
iililllilli
Average annual growth rate, 1960-73
MI-
I!
p
2 o
p
p
p
p
©
©
©
©
P
©
Average annual growth rate, 1950-88
p
Average annual growth rate, 1950-88
S
f g
t
I—*
I
TO
o'
o
394
Cormac 0 Grada and Kevin O'Rourke
FL NW
IT
UK
SP
^ K sw £ 0.014< 0.012 7000
BRD
FR 9000 11000 13 000 GNP/capita, 1979 (1990 $)
15 000
Figure 13.3 GNP per capita, initial level and growth: Europe, 1979-88 (Maddison data)
(Williamson, 1994). This could be because downturns in economic activity lead to protection and other autarkic policies, or there could be deeper reasons connected with this growth process. Williamson has found strong international wage convergence before 1914 and after 1945, but no convergence between the wars. Perhaps international stability on the macroeconomic, exchange rate and/or trade fronts is necessary for convergence to take place. Second, Ireland is the clearest outlier over the whole period, having a much lower growth rate than its initial income would suggest. This is true with both data sets. During the 1950s and the post-1973 period, Ireland is joined by Spain, Portugal and Greece; the four countries seem to constitute a mini-convergence club, lying on a line to the left of the rest of Europe. During the 1960s they do a lot better, fitting into the general European convergence pattern to a greater extent; but even during this period, Ireland is an underperformer (especially according to the Maddison numbers). Third, the data set you use matters. We saw in Table 13.1 above that it matters a lot for average growth rates. It also matters for telling qualitative convergence stories about Ireland; but only for the post-1973 period. Ireland seems below (or to the left) of the convergence line with both data sets for the period 1973-88, but is far less clearly an underperformer if you use the Maddison data. The difference in Irish growth rates using the two data sets, which we saw in Table 13.1, and which is worryingly large, is sufficient to reverse one's conclusion about Ireland's post-1973 performance based on GDP/capita performance. However, the pessimist case can be resurrected for the post-1973 period, even on the basis of Maddison's numbers, if one considers GNP rather than GDP per capita. GNP is more relevant than GDP per capita for living standards: Heston and Summers give GNP as a multiple of GDP for a smaller sample of countries, after 1979. Applying these correction factors to the optimistic Maddison GDP numbers, we get the data portrayed in Figure 13.3: Ireland is once again a dramatic
Irish economic growth, 1945-88
395
underperformer during this period. This is due to Ireland's excessive borrowing during the post-1973 period, which led to the GNP/GDP ratio shrinking from 96.4 per cent in 1979 to 88.1 per cent in 1988. Fourth, the meaninglessness of comparing Irish with British growth is apparent from these graphs. The standard story has been that Ireland kept up with the UK, but that the British economic performance was one of the worst in Europe. It appears that UK growth was more or less what would have been expected given its initial level of GDP per capita. Ireland should have been growing faster than the UK, but was not; it is Ireland which appears to be the sick man of Europe, not the UK (see also Castles, 1991: 21-2). Fifth, one result of Ireland's relatively slow growth was that between 1950 and 1973 it was overtaken by Spain (according to Maddison), or Spain, Italy and Austria (according to Heston and Summers). When viewed in the context of general European convergence, Ireland's performance thus appears even more disappointing than previous Irish authors have suggested. To summarize the evidence on Ireland's relative performance presented thus far: Ireland was a clear underachiever throughout the post-1950 period, no matter which data are used. There is one exception to this general finding: Maddison's GDP numbers show a satisfactory Irish performance (from the standpoint of the convergence literature) after 1973. However, even for the post-1973 period, Ireland is clearly a bad underachiever from the standpoint of GNP/capita growth, no matter whose data are used. From the perspective of the convergence literature, Ireland emerges as a spectacular outlier. What can explain all this? A first question which one might ask is: how did Irish productivity growth, as measured by GDP per employed worker, behave over the period? If Ireland was not an outlier by this standard, then its failure by the criterion of GDP per capita must be due to a relative deterioration in either Irish employment or Irish labour force participation. Figure 13.4 gives what information we have on GDP per worker. As may be seen by comparing with Figure 13.2, adjusting for labour force participation does indeed move Ireland closer to the 'convergence line'. Yet Ireland still remains somewhat of an outlier in most periods, the exceptions again being the post-1973 period according to the Maddison data, and the 1960s according to Heston and Summers. Figure 13.5 presents the data for GDP per employed worker, for EC countries only, and for the post-1964 period.8 Again, the qualitative picture is similar, although once more the adjustment appears to make Ireland look better (according to the Maddison data, Ireland is now hardly an outlier for the entire period 1964-88). Unemployment and higher dependency rates thus help partly to account for Ireland's poor growth in GDP per capita in the postwar period. On the other hand, productivity growth was not as fast as it should have been either, and some of the unemployment might plausibly be attributed to this (Kennedy, 1993). Moreover, higher employment might have implied lower average labour productivity: if so, better employment performance would have been possible only at the expense of slower productivity growth. In sum, there was a productivity failure as well as a more general standard of living failure over the period as a whole, and after 1973 as well if you believe the Heston-Summers data. Moreover, even if it were the case that Irish productivity growth, as measured by
396
Cormac 0 Grada and Kevin O'Rourke
rr
A SP
0.040-
A SP
0.040BRD FL
0.035-
0.035-
'
BRD FL FR
IRL
0.030-
0.030-
NW 0.025-
0.025DK
NL DK u ^ w
^ 0.0204000 6000 8000 100001200014000160001800020000 GDP/worker, 1950 (1990 $)
CH
0.020
2000
B
4000 6000 8000 10000 12000 14 000 16 000 GDP/worker, 1950 (1985 $)
(b) 1950-88, Maddison data
(a) 1950-88, Heston-Summers data 0.08
0.08-|
0.07-
0.07 AIT
0.060.050.04
P GR
0.06FL
FR
0.05-
NL
0.04-
SP
0.03-
IRL
0.030.02-
0.02-
0.01-
0.01-
2000 4000
4000 6000 8000 1000012000 140001600018000 20000 (d) 1950-60, Maddison data )0-73
GDP/worker, 1950 (1990 $)
(c) 1950-60, Heston-Summers data GR SP
0.070.06-
8000 10000 12 000 14 000 16 000
GDP/worker, 1950 (1985 $)
0.080.07-
l rate,
0.08-
6000
P A
0.05IRL
| ITra ™ BRD DKvw uiCsw IC
0.03-
CH
0.024000 6000 8000 1000012000 14000 1600018000 20000 GDP/worker, 1960 (1985 $)
SP
0.06-
A
0.05-
1 0.04-
B
FL
0.04-
TT IRL
I
0.026000
10000
P
GR
\]K
0.010
SP
0.005 0 5000
th rate,
S
FL IC
0.020
DK
iL
JiRD
1
B
NL
CH
IRL 10000
NL w
14000 18000 22000 GDP/worker, 1960 (1990 $)
CH 26000
(f) 1960-73, Maddison data
ftW
0.025
BRD ^DK
0.03-
§8
0.030
FL
ca
(e) 1960-73, Heston-Summers data
0.015
FR NW
15000 20000 25000 GDP/worker, 1973 (1985 $)
(g) 1973-88, Heston-Summers data
30000
1 |
0.030 0.025
IRL
IpW
IT
0.020 0.015 0.010
SPJKAB|R DK
sw^
p
0.005
NL
^
0 10000 15000 20000 25000 30000 35 000 40000 GDP/worker, 1973 (1990 $) (h) 1973-88, Maddison data
Figure 13.4 GDP per worker, initial level and growth: Europe, 1950-88
Irish economic growth, 1945-88
397
S8 0.0403
0.035-
|
0.030
2 0.035 5 0.030F BRD B UK
•a 0.025-
I 0.020-
» 0.020
I 0.015 81
i 0.015 4000 6000 8000 1000012 00014 00016 00018 00020 000 GDP/employed worker, 1964 (1985 $)
12000 16000 20000 24000 GDP/employed worker, 1964 (1990 $)
P 0.08-1 i
* 0.07-
* 0.07-
1 0.06
1 0.06-
I 0.05-
I 0.05-
0.04
NLL
|
& 0.030.02-1
DK
(b) 1964-88, Maddison data
(a) 1964-88, Heston-Summers data P 0.08-
|
BRD UK
•a 0.025-
0.04& 0.03-
1
1
.
1
1
1
1
r
0.010-
8, 0.010
j 0.005
8000
12 000 16000 20000 24 000 GDP/employed worker, 1973 (1985 $) (e) 1973-88, Heston-Summers data
28 000
0.005 10000
15000 20000 25000 30000 Initial GDP/capita, 1973 (1990 $)
35000
(0 1973-88, Maddison data
Figure 13.5 GDP per employed worker, initial level and growth: Europe, 1964-88
GDP per employed worker, had been satisfactory since 1950, the conclusions which one would be justified in drawing from this would be unclear. The Irish dependency rate is not just a function of birth rates, death rates and attitudes towards female labour force participation. It is significantly affected by emigration, which is in turn produced by the same problem as leads to unemployment: a failure by the economy to create sufficient employment. Some commentators have argued that Irish productivity growth has been satisfactory, and that slowly growing living standards are thus due to Catholic mentalities reducing the proportion of the total population employed: the implicit inference is that no great economic failure is involved. We would argue that sluggish employment generation represents a serious failure of economic policy; the inference which we would draw from Figures 13.4 and 13.5 is that not only Irish productivity growth, but the performance of Irish labour market institutions, has been unsatisfactory ever since the mid-1960s.
398
Cormac 6 Grada and Kevin O'Rourke
A second immediate question arises: how can there be such a large discrepancy post-1978 between the two data sets examined here? After all, they both make use of national accounts data, and the relative price information contained in the UN International Comparison Program. We confess to being puzzled by the discrepancy, and here content ourselves with listing potential causes for it: • The Maddison data pjecented here use 1990 prices, while the Heston-Summers data use 1985 prices; but the same discrepancy emerged when earlier OECD data using 1985 prices were used. • Revisions to the Irish national accounts which were used by one study may have been neglected by the other. • Transfer pricing by Irish multinationals means that the value of output given in the Irish national accounts may differ significantly from the value of output using world prices. Similarly, Irish agricultural output is sold at CAP prices, which are far above world prices. Both trends have become more important since 1978. These two sectors together account for a large share of total Irish output. If the two studies differed in the degree of aggregation used, so that one study failed to pick up these discrepancies between national and world prices in important sectors, that could explain the difference in the two studies.9 Finally, comparing the Maddison data with the Heston-Summers data for other European countries may be of some help in resolving the issue. Table 13.2 gives the percentage by which the Maddison data for GDP per capita (expressed in 1985 dollars) exceed the Heston-Summers data for 1950, 1973 and 1988/9.10 As can be seen, in the richer European countries the Maddison figures are generally around 10 per cent higher than the Heston-Summers data. In the poorer countries - for which, suggestively, agriculture is relatively more imporatnt - the gap is bigger still. The discrepancies between the two data sets tend to be relatively steady over time, except in the cases of Belgium, Norway and (especially) Ireland. In conclusion, it seems clear that some effort should be put into understanding the differences between these two widely used data sets. In the meantime, while these numbers can be relied on to tell reliable stories about broad international patterns, researchers should be wary of telling stories about individual countries. Nevertheless, on the basis of the evidence presented above, we feel that a negative assessment of Irish economic growth since 1950 is supported. 3
Irish economic history, 1945-92
We now offer a brief narrative outline of economic trends since 1945 in that part of Ireland known officially as Saorstat Eireann or the Irish Free State until 1948, and as the Republic of Ireland since then (see also 6 Grada, 1994a). The period since political independence in 1922 had been one of economic experimentation. The economic stance of the Cumann na nGaedheal party, which ruled between 1922 and 1932, had been deliberately cautious. It preached, and indeed practised, comparative advantage and fiscal retrenchment. This meant giving a wide berth to populist nationalist credos and relying on an improved performance from the dominant agricultural sector as the 'engine of growth' (Daniel, 1976). Agriculture then employed half the labour force, and agricultural policy aimed at emulating Danish
Irish economic growth, 1945-88
399
Table 13.2. Maddison and Heston-Swnmers GDP per capita (% by which Maddisonfiguresexceed Heston-Summers)
Austria Belgium Denmark Finland France FRG Italy Netherlands Norway Sweden UK Greece Ireland Portugal Spain
1950
1973
1988/9
11.7 1.8 13.7 9.5 11.6 5.1 10.3 15.0 6.1 12.4 12.0 15.9 0.04 34.7 24.2
94 4.5 10.5 8.5 7.0 6.7 10.9 10.5 7.2 13.1 10.9 17.9 -4.7 31.5 18.3
105 10.7 12.5 11.8 12.6 8.3 9.6 9.5 1.5 12.4 11.4 22.6 24.7 27.9 26.5
success by capturing the high-quality end of the British market for meat, eggs and dairy products. This policy, always associated with Agriculture Minister Patrick Hogan, had yielded few dividends by 1932; perhaps it was given insufficient time to assert itself (O'Brien, 1936). Another feature of Cumann na nGaedheal policy was fiscal and monetary prudence. Both nationalist ideology (which had long maintained that Ireland was overtaxed within the United Kingdom) and the emphasis placed on minimizing the input costs faced by farmers supported low taxation in the 1920s. Ministers also resisted the demands, which were becoming more vocal in the late 1920s, for widespread tariff protection. That the Irish economy, so closely linked to the UK's, grew sluggishly during the 1920s should come as no surprise. Emigration, long blamed on British maladministration, continued at a high rate, and living standards barely rose. The onset of the Great Depression in 1929-30 dented the appeal of Cumann na nGaedheal's liberal orthodoxies. They failed to maintain popular support, and were replaced by Eamon de Valera's Fianna Fail party in 1932. Now ideology and contingency would combine to transform a virtually free-trading economy into one bent on state supports and import substitution. Faced with the challenges of high unemployment and declining markets abroad for agricultural produce, Fianna Fail's alternative was to build an indigenous industrial sector through protection, and to reorient agriculture towards a more labour-intensive product mix. In a context of world depression, low farm prices and a political imperative to reduce unemployment, this strategy had its logic; in the short run, it produced a significant increase in employment in the sheltered industrial sector. The new manufacturing industries, small scale and low productivity, had already captured the home market by the late 1930s. However, neutrality during World War II forced a degree of self-sufficiency on the South's economy that exceeded the demands of even the most ardent economic
400
Cormac 0 Grada and Kevin O'Rourke
nationalist. Imports as a share of national income fell from 26 per cent in 1938 to 11 percent in 1944, and imports of capital equipment and raw materials fell almost in proportion. Thus although the economy had been spared military destruction, near-autarky led to a running down of the capital stock through wear and tear. Gross domestic capital formation fell to a paltry 3.7 percent of national income in 1944-5, and output per worker in the manufacturing sector fell. Agriculture was less affected, and its share of national income rose from 28 per cent in 1938 to 37 per cent in 1944-5. Yet World War II was far from being a bonanza for Irish farmers: prices on the British market were strictly controlled, and an inadequate supply of fertilizers, feedstuffs and machinery depressed crop and milk yields. The result was a badly undercapitalized agricultural sector. In 1945, Ireland remained one of the poorest countries in Western Europe. Ireland experienced a postwar recovery, like the rest of Europe. This was largely predicated on trade: merchandise trade (imports plus exports) as a share of national income rose from 28 percent in 1945 to 64 percent in 1950. The resulting trade pattern could not be sustained for long, however; merchandise imports and exports had been roughly equal during the war years, but in 1946-50 the value of imports was double that of exports. A consumer boom produced a fourfold rise in the value of imported manufactured goods between 1945 and 1950. The multiparty coalition which ruled Ireland in 1948-51 after sixteen years of Fianna Fail government believed that the 'inadequacy of capital investment, particularly in agriculture' was the root cause of continued emigration. This emphasis on the leading role of the farm sector marked a return to the beliefs of the 1920s, although the new government failed to dismantle the tariff regime established in the 1930s. One result was a rise in the Department of Agriculture's share of government spending from 2 per cent in 1943-5 to 16 per cent in 1950-2. Overall, the economy's investment rate recovered, reaching 14 per cent of income in 1948-50; though (as critics would soon point out) perhaps too high a proportion was in 'social' investment, such as housing, hospitals and schools. Fianna Fail's radicalism of the 1930s had not extended into the monetary or budgetary sphere. Monetary experimentation was rejected out of hand, and Ireland's currency readily exchanged into sterling at par throughout. Though the sterling link was not without its tensions, particularly when sterling devalued, it probably boosted investor confidence and ensured the acceptance of the Irish pound. Thus the coalition government regarded the 1948 devaluation of sterling as 'most unwelcome', but nevertheless resigned itself to following the fate of sterling. The Irish banking sector, dominated by a cartel of joint-stock banks, may have lacked dynamism, but it proved very stable. The banks vehemently opposed the creation of an Irish central bank, and the Free State survived without one until 1943, when the Central Bank of Ireland was set up. Ireland stands out for being one of the few economies free of banking panics or failures during the 1930s. The budgetary stance of all administrations in the 1920s and the 1930s had been rather conservative, and even Fianna Fail paid no heed to Keynes's call in 1933 for deficit spending on urban renewal and other worthwhile projects (Keynes, 1933). At the end of the 1930s, Ireland's national debt was still relatively small by contemporary European standards. The decision to partition government spending plans into current and capital spending in 1949 was seen as the first sign of an Irish conversion
Irish economic growth, 1945-88
401
to Keynesian thinking; in the memorable quip of one economist, 'Keynes had come to Kinnegad' (Lynch, 1969: 187). During the postwar recovery phase,fiscalpolicy was quite lax. However, stop-go policies in which an adverse balance of payments signalled drastic fiscal action were to follow, and produced severe and largely unnecessaryfiscalcontractions in 1952 and 1957 (Kennedy and Dowling, 1975: chs. 13-14). Much was expected in Ireland of the Anglo-Irish trade agreement of 1948, which again opened up the British market to Irish livestock exports. The hoped-for growth (stipulated in an appendix to the agreement) was not realized, since the British system of deficiency payments depressed food prices. The terms of trade proved quite unfavourable to Ireland in the early and mid-1950s. The 1950s turned out to be a dismal decade for the Irish economy. GDP grew more slowly than in any other economy in the OEEC, and net emigration reached levels not equalled since the 1880s. Poor economic performance prompted a fundamental reappraisal of the policy package pursued since the 1930s. The reappraisal was articulated in a famous White Paper, Economic Development, and in the Programme for Economic Expansion (later called the First Programme), both published in 1958. These documents were quite vague about both policy instruments and targets, yet they presaged a new consensus on the need for a more outward-looking economic policy, including trade liberalization and reliance on direct foreign investment. The First Programme was Ireland's first tentative exercise in indicative planning; it outlined the strengths and weaknesses of the economy, emphasized the importance of farm investment, and held out hopes for slow economic growth. Its targets were modest in scope, and vague as to the mechanisms for achieving them. An annual growth rate of 2 per cent over afive-yearperiod was anticipated. In its emphasis on agriculture as the engine of growth, the First Programme was strictly traditional (Lee, 1989: 350-1); more important, it was explicit in its commitment to trade liberalization and the reorientation of industry towards foreign markets. The economy soon picked up; historians have generally tended to give the First Programme the credit, and the sharp rise in investor confidence in the late 1950s suggests that the Programme produced a 'euphoria effect'. However, another view posits that the economy, which was emerging from recession in any case, would have performed just as well in the absence of the Programme. The First Programme was succeeded by the Second (1964-70) and Third Programmes (1969-72). These were far more detailed than the first, and contained detailed sectoral projections. In neither case was even the overall aggregate growth rate target of 4 per cent per annum met, and both were abandoned before the 'due date'. The Second Programme, formally abandoned in 1967, was laid to rest because of the widening disparity between the projections for employment growth and reality. The plans were alsoflawedmethodologically (Norton, 1975; Bradley, 1990). Economic planning went out of favour for a few years, although it briefly returned with a vengeance in 1977 when a new Fianna Fail administration created the Department of Economic Planning and Development, and launched National Development 1978-1980. That plan's ambitious projections - including targets of 25 000 new jobs a year and an annual GNP growth rate of 7 per cent - proved to be wildly unrealistic. There followed The Way Forward (1982) and Building on Reality
402
Cormac 0 Grada and Kevin O'Rourke
(1984), but both of these emphasized budgetary constraints rather than overall growth targets. These documents, better characterized as stabilization programmes rather than exercises in French-style indicative planning, marked the end of Ireland's experiment with economic planning. That experiment is a reminder that setting and meeting detailed medium-term growth targets for a small open economy is a difficult, if not downright pointless exercise. By the late 1980s, planning's only legacy was the Programme for Economic and Social Progress, a joint commitment by
the social partners and government to income and social welfare targets. Much more important for the economy's improved performance than the commitment to planning was the renewed commitment to trade liberalization. This brought membership of the Anglo-Irish Free Trade Area (1966) and the European Community (1973). The stage was set in the early 1960s by the 'corporatist' Committee on Industrial Organization, created by the government in 1961 to study the likely impact of trade liberalization on native industry. The Committee's studies of the whole range of Irish manufacturing - twenty-six reports put together at some speed by representatives of the trade unions, government and business - painted a bleak picture of an industrial sector beset by shoddy design, poor marketing and short production runs. Detailed recommendations by the Committee were followed by the introduction of a grants scheme to helpfirmsseeking to adapt to free trade, and the strengthening of export-promoting state agencies such as An Bord Bainne (The Milk Board) and Coras Trachtala (The Export Board). These measures helped to prepare the way politically. The rather leisurely pace of tariff reduction envisaged under the Anglo-Irish Free Trade Agreement speeded up with EC membership, with Ireland undertaking to remove all tariffs against EC member countries by 1978. The static gains from trade liberalization were presumably small, although there may well have been important dynamic gains (see section 6 below). The Irish economy performed better between the late 1950s and the mid-1970s than at any other time since independence. GDP growth averaged about 4 per cent, and emigration had been brought to a virtual halt by the end of the period. The growth was predicated partly on direct investment by multinational firms, and partly on running down the external assets of the banking system. Ireland's initial main policy response to the oil crisis of 1972-3 was a succession of large budget deficits. The public sector borrowing requirement (PSBR) rose from 8.6 per cent of GNP in 1972-3 to 12.9 per cent in 1976^7 (Leddin and Walsh, 1992:122). At the time, this rise was rationalized in Keynesian terms. But budget deficits continued to accumulate in the following few years, shielding the Irish consumer for a time from the effects of the oil crisis, but raising the PSBR and the national debt to clearly unsustainable levels. The huge rise in public spending - the PSBR reached 17.3 percent of GDP in 1980 and 20.3 per cent in 1981 - failed to generate much productive investment: indeed, despite gross investment rates of 30 per cent of GDP in 1978-81, the economy grew at an average rate of only 2.5 percent in thefirsthalf of the 1980s. Economists were quick to criticize thefiscalexpansion of those years; indeed, the tone of some critics had turned apocalyptic by the early 1980s. Politicians were slower to learn than economic commentators, and as a result most of the 1980s were wasted undoing the damage of earlier fiscal recklessness. However the 'delusion' that Ireland could sustain living standards by borrowing in the wake of the second
Irish economic growth, 1945 88
403
oil price hikes had dissipated by 1983. The rhetoric of Budget Day speeches reflected lessons dearly learned. In 1978 the Budget sought 'to give an impetus to economic activity' that would increase the annual growth rate in GDP to an unprecedented 7 percent, as envisaged by National Development 1978-1980. The combination of tax cuts and public spending would, it was hoped, grease the wheels of the private sector. But while the Budget speech of 1981 derided 'loose comment about so-called disorder in the public finances', by 1983 there was a clear recognition that the cost of servicing the debt had 'preempted resources for the future'. 'Populist quick fixes' were ruled out in the 1993 Budget, and the PSBR has now been reduced to below 3 percent. The pervasive gloom of the mid-1980s, reminiscent of the 1950s, has given way to mild confidence about the future. Irish economic growth since 1987 has been among the fastest in the OECD. The Irish decision in December 1978 to participate in the European Monetary System (EMS) was a landmark in recent Irish economic history, since it brought to an end the monetary union between Ireland and Great Britain that had lasted since 1826. The aim of EMS membership was a monetary discipline which would win Ireland investor credibility, low inflation and low interest rates. At first, the inconsistent stance of Irishfiscalpolicy and repeated realignments of the value of the Irish pound (or punt) within the system's Exchange Rate Mechanism sapped investor confidence, and Irish interest rates remained high. The currency realignments ceased in 1986, and by the early 1990s the Irish yield curve was downward sloping and the gap between Irish and German interest rates minimal. By the early 1990s it could be said that the goals of price stability and low interest rates had been reached. This brought a reduction in the cost of servicing the national debt and in the fiscal burden facing the economy. The victory was costly, however, in terms of output and employment forgone, and some economists now argue that the exchange rate regime pursued by Ireland under the EMS was unduly deflationary (Dornbusch, 1989; Leddin and Walsh, 1992: 390-3). The aim of EMS membership had been to 'free' the punt from the shackles of a weak sterling. With a pwnr-Deutschmark exchange rate of £1 = 2.65DM, which has been maintained since mid-1986, the goal seemed to have been met by the early 1990s. The crisis caused by sterling's departure from the Exchange Rate Mechanism in September 1992 had therefore not been widely anticipated. Moreover, most commentators initially believed that Ireland would resist the pressure to devalue in order to maintain the credibility that had been so dearly won. The costly battle to 'save' the punt against the speculators lasted four months. In January 1993 the Irish currency was devalued by 8 per cent within the Exchange Rate Mechanism. With the virtual demise of the EMS in mid-1993, the Irish currency has become 'one of the smallest independent currencies in the world' (Walsh, 1993b). The search is now on for that golden rule-of-thumb which would yield a punt capable of reducing Ireland's savage unemployment rate without endangering exchange rate stability. The Irish combination of fiscal retrenchment, tough exchange rate policy and economic recovery in the 1980s has attracted the attention of outside experts, evoking the admiration of some and confounding others (Dornbusch, 1989; Giavazzi and Pagano, 1991). Giavazzi and Pagano have proposed the Irish recovery as a classic example of 'expansionary fiscal contraction', a claim denied by Barry (Barry, 1991; also Barry and Bradley, 1991). An alternative interpretation of recent
404
Cormac 0 Grada and Kevin O'Rourke
Irish macroeconomic history has buoyant world market conditions and capital inflows outweighing the deflationary effects of fiscal contraction (Leddin and Walsh, 1992: ch. 13; Barry, 1991). 4
Investment
The traditional Solow growth model places much emphasis on investment, even though steady-state growth is not affected by it. The new growth theory has placed even more emphasis on capital accumulation; indeed, it is largely motivated by the fact that traditional Solow-based growth accounting puts too small a weight on capital's contribution to economic growth (see, for example, Crafts, 1992: 391-3). While the Solow model suggested that an increase in savings rates could increase income levels but not growth rates, the new growth literature attributes both level and growth effects to investment. De Long and Summers have gone further and attributed major explanatory power to investment in equipment as a determinant in growth (De Long and Summers, 1991). The literature has tended to find a significant positive relationship between investment shares and growth, when growth is regressed on a list of explanatory variables including investment (Barro, 1991; De L ong and Summers, 1991; Mankiw et aL, 1992; Dowrick and Nguyen, 1989; Crafts, 1992). Moreover, investment is itself an endogenous variable related to, among other things, initial income levels. Traditional growth models would suggest that in poorer countries returns to capital, and hence investment shares, should be higher than in rich countries. Barro (1991) found this to be so, once initial levels of human capital had been controlled for. Theory and aggregate experience thus suggest that Ireland should have had very high investment shares during our period; if it did not, this might help explain its failure to converge. Irish policy-makers were traditionally concerned about a lack of investment in the Irish economy. The 1958 White Paper (1958: 35) highlighted 'the insufficiency of our current savings as a basis for national capital formation on the scale which would be necessary to enable us even to follow at some distance the rising standards in the rest of Europe'. Were these fears justified? Table 13.3, using EC data, gives investment shares for Ireland, the UK and twelve present EC members since 1960. Irish investment rates were consistently above UK levels, but only exceeded average EC levels after 1970. A similar picture emerges when the Penn Table data are used. Figure 13.6, based on that source, gives annual investment shares for Ireland, the UK and the rest of Europe from 1950. Irish investment rates were continually above UK rates, but only exceeded European rates during the 1970s and early 1980s. Investment was modest in Ireland during most of the 1950s, but increased rapidly in the following decade. The 1960s also saw a drop in marginal capital-output ratio; this was partly due to the utilization of infrastructure created in the early and mid-1950s (Kennedy and Dowling, 1975: ch. 11). The gross investment rate exceeded 30 per cent several times during the 1970s, but declined sharply thereafter. Conventional crowding-out stories seem implausible in a small open economy, and the high rates of the 1970s, a time when public expenditure was rising rapidly, suggest that it was not an issue for Ireland (Honohan, 1992a). The general climate of uncertainty and macroeconomic instability associated with the Irish debt crisis, and the severe downturn of the 1980s
Irish economic growth, 1945-88
405
Table 13.3. Gross investment, 1961-90 (as a % of GDP) Ireland
UK
E C 12
1960-9 1970-9
19^5 25.0
lil 19.2
218 22.8
1980-90
21.6
17.5
20.1
Source: European Economy, table 20. 35-| 309
I
• Ireland + UK * Other Europe
25
15-
10 -tf
1950
1955
1960
1965
1970
1975
1980
1985
Year
Figure 13.6 Investment shares: Ireland, UK and the rest of Europe, 1950-88
(largely caused by the earlier accumulation of debt) are much more plausible reasons for the decline. In the event, Irish investment rates only fell below European rates in 1987. In summary, the Irish underinvested if the yardstick used is European investment rates, and if one bears in mind the implication of the standard convergence story that poorer countries should invest more. Underinvestment might thus explain slow Irish growth prior to 1973, as well as the spurt in growth after 1973 which the Maddison data suggest. However, if Heston and Summers are right, the dismal Irish performance from 1973 onwards which they document cannot easily be explained by appealing to Irish underinvestment. But perhaps aggregate investment shares are not what is relevant: maybe the quality of Irish investment was poor and its composition wrong? De Long and Summers (1991) make a strong statistical case for the proposition that it is investment in capital goods that is crucial for growth; were equipment goods prices too high in Ireland, and was the share of GDP devoted to equipment investment too low? Figure 13.7, which uses the data provided by De Long and Summers, suggests that their insight cannot explain slow Irish growth. The normalized equipment price in Ireland was lower than that in most European countries* and the Irish equipment share was relatively high. Perhaps a more disaggregated approach is needed. Table 13.4 summarizes trends
406
Cormac 6 Grada and Kevin O'Rourke
Table 13.4. Investment by use: Ireland, 1953-90 (%)
1953-9 1960-9 1970-9 1980-90
Dwellings Roads
Other buildings
Transport Agricultural Other equipment machinery machinery
17.1 15.9 22.9 22.4
37.0 33.5 26.8 24.2
14.1 14.4 12.4 14.0
5.6 4.3 2.1 3.7
5.5 4.8 4.5 2.5
20.6 27.1 31.3 33.2
Source: Derived from National Income and Expenditure.
0.13 0.12 0.11 0.10 0.09
FL
A BRD
08
°0.07
FR
IRL
DK
GR
NL
vl
B NW
0.06 0.05 SP
0.04 0.03 -0.5
-0.3
-0.1 0.1 Equipment price (normalized)
0.3
Figure 13.7 Equipment share and price: Europe, 1960-85 in the share of gross investment by use. The share of agricultural machinery in total investment has dropped steadily since the 1950s, while that of 'other machinery' has risen steadily, from one-fifth of the total in the 1950s to over one-third today. But the consistently high proportion spent on transport equipment is the most noteworthy feature of Table 13.4. Was this 'unproductive' investment, or was communications a relatively important industry in Ireland for geographical or other reasons? The output of the Irish transport and communications sector - about 6 per cent of GDP, proportionately no greater in Ireland than in other European countries in the period under review - casts doubt on the last explanation. Surely a more plausible explanation for the high share of transport is the loss-making capital grants to concerns such as the national air, rail and sea carriers. Dividing the output of the transport sector by the sum invested in it in a selection of European economies indicates that the return on investment in transport equipment was lower in Ireland than in any of the other European economies examined (Table 13.5). It is also sometimes alleged that too much Irish capital formation has been in the form of public sector investment. The argument is that the public sector feels less pressure and incentive to allocate funds to the most profitable uses. Defining public
Irish economic growth, 1945-88 407 Table 13.5. Returns on investment in transport equipment, selected years
Ireland
Italy
UK
Norway
Year
Ratio
1970 1979 1986 1964 1980 1986 1964 1980 1986 1964 1980 1987
1.7 1.9 2.4 3.3 2.7 2.9 4.1 3.4 4.2 2.2 3.6 4.8
Year 1964 1980 1987 Netherlands 1969 1980 1987 1964 . Belgium 1980 1987 1964 Sweden 1980 1987 Germany
Ratio 3.0 3.2 3.1 3.5 3.7 3.3 3.5 3.5 4.6 2.7 3.3 3.2
capital formation as public capital expenditure minus redemption of securities and payments to the rest of the world (usually a small item), the public share in total gross fixed capital formation in Ireland has indeed been high, usually ranging between 30 and 40 percent between the 1950s and the 1980s, and falling below 20 percent only in 1990 and 1991. n Eurostat provides a comparative perspective: it suggests that the public share in gross fixed capital formation was not particularly high in Ireland in 1970, but that by 1986 it was the highest in the EC. Capital grants to enterprises have typically accounted for only 10-15 percent of Irish public investment; capital formation by government and local agencies accounted for a half. Loans and share capital accounted for 37 percent in 1960-2, and 17 percent in 1988-90. On the face of it, there is scope for waste here, even if Barro finds evidence 'consistent with the hypothesis that the typical country comes close to the quantity of public investment that maximizes the growth rate' (Barro, 1990: 124). However, the share of the public sector in Ireland was twice that of the average (16.4 per cent) in Barro's 76-country sample (Barro, 1991: App. 1, p. 438). In summary, there would seem to be several reasons why Irish investment ratios, low as they were for much of our period, would overstate the investment performance of the Irish economy during this period. This highlights some of the problems with cross-country regressions of the sort that we will be looking at later in the paper: by using explanatory variables that are of necessity crude, they may on occasion hide more than they reveal.
5
Human capital and emigration
Human capital has been much emphasized in recent theoretical and empirical work on economic growth. In theoretical work, it is an important feature of the endogenous growth literature. In applied work, it has largely served as a way of refining the original Solow model, and making that model more consistent with the facts. Mankiw et al. (1992) stress that incorporating human capital boosts capital's share in income to an extent that capital accumulation can explain quite large
408
Cormac 6 Grada and Kevin O'Rourke
differences in income per capita. More relevantly for our purposes, Barro and others have found that convergence does hold in a conditional sense: controlling for human capital, poorer countries do grow faster. The failure of LDCs to converge as a group on the rich countries is largely due to their poor human capital endowment (Barro, 1991). This finding will come as no surprise to economic historians, who were reminded by Richard Easterlin of the importance of education in facilitating the spread of new technology over a decade ago (Easterlin, 1981). Conditional convergence in the sense described above has become a widely accepted stylized fact. It thus seems natural to ask: is Ireland's failure to converge on the rest of Europe attributable to a poor record of human capital accumulation? Surprisingly, Ireland does quite well in international comparisons of educational expenditure and output. Irish expenditure on education has been both high and rising (from 3.1 percent of GNP in 1962 to 6.4 percent in 1989). This is reflected in Ireland's relatively high rate of school attendance in the recent study by Mankiw et al (1992), where Ireland ranks seventh out of 121 countries. Thus Ireland not only failed to grow faster than richer economies during this period; it also underperformed relative to its rate of investment in education. Walsh's recent econometric study of an OECD subset of the Mankiw et al data-set study confirms Ireland's poor record in 1960-85, 'in the sense that the level of GDP per adult reached in 1985 was lower than would have been expected on the basis of its initial level in 1960, the rate of growth of the labour force and the rates of human and physical capital formation'.12 Decline in the quality of education is unlikely: analyses of the contribution of education (measured by years of schooling) to earnings in 1972 and 1987 imply no decline in the interim, and the marginal returns to higher education are at least as high in Ireland as in other countries offering comparable data.13 Nor is the frequently alleged bias against technical and science-oriented subjects in the Irish education system supported by the facts (Sheehan, 1992). Perhaps emigration, by removing human capital, had something to do with the poor Irish performance? Ireland was after all an outlier as regards emigration, as Table 13.6 makes clear. If emigrants are better educated than the population at large, emigration will reduce the growth of income per capita in the sending country (Berry and Soligo, 1969; Dolado et al 1993). However, the selection bias in emigration from Ireland before 1960 was, if anything, towards unskilled workers, and therefore arguably reduced unemployment and upgraded the structure of the remaining labour force (6 Grada and Walsh, 1992). The net outflow declined during the 1960s, turning negative in the 1970s. When it resumed in the 1980s, the highly educated were disproportionately represented among the emigrants. But this is a problem for policy in the 1990s, and hardly explains the poor record hitherto. Emigration could have hurt the Irish economy in another way, however. Williamson (1993) examined the convergence in industrial wages across countries, and found that Ireland was in no sense an outlier according to this criterion after 1950: indeed, Ireland lies precisely on the estimated 'convergence line'. This finding is obviously at sharp variance with what the GDP or GNP statistics tell us. If industrial workers were doing as well as could have been expected, but the country as a whole was not, then some other group in society must have suffered. Was it capitalists? This seems unlikely, since capital is freely mobile between Ireland and the rest of the world. Farmers? This also seems unlikely, at least for the post-1973
Irish economic growth, 1945-88 409 Table 13.6. Net migration in Western Europe, 1950-87 (in thousands) 1950-73 Belgium France Germany Italy Netherlands Norway Sweden Switzerland UK
Ireland Total
287
3 360 7070 2139 47 0 336 755
^605 -534 15133
1974-87 -8 365
1042 544 -18 77 167 6 15 -68
2310
Note: The totals refer to the sums of absolute values. period of EC bonanzas. Williamson's finding may perhaps be reconciled with our own, if the industrial wages used in his analysis are not competitive market wages, but wages earned by privileged 'insiders' holding desirable jobs and protected by strong unions. 'Outsiders' in casual employment and on the dole queues would then be the ones who have lagged behind. The fact that Irish labour markets are well integrated with world markets might mean that insiders, who have the option of emigrating when suitable jobs come up abroad, demand and receive internationally satisfactory wages. Outsiders have the option of receiving substandard domestic wages, going on the dole, or emigrating. The recent experience of East Germany suggests a possible parallel here. Emigration may not by itself have raised East German wages (since labour is clearly not scarce there); but German integration may nevertheless have led to trade unions pushing up insider wages, leading to mass unemployment. We believe that a careful study of labour market institutions, and the ways in which these interact with world labour markets, has to be a priority for Irish researchers in the near future. In summary, then, while human capital accumulation has provided a useful way of reconciling aggregate evidence with the Solow model, and 'explaining' the lack of convergence at the global level, it does not, at first sight at least, explain Ireland's poor postwar economic performance. The role of emigration, however, deserves further study. 6
Trade policies
While the market imperfections at the heart of both new trade theory and new growth theory can potentially damage the case for free trade, the empirical case for free trade has on balance been strengthened by these developments. A well-known survey by Baldwin (1984:586) found that the estimated gains from trade liberalization in the context of standard competitive CGE models were very small: for example,
410
Cormac 0 Grada and Kevin O'Rourke
Whalley and Wigle (1982) found that the gains to a 50 per cent multilateral tariff cut amounted to a mere one-third of 1 per cent of world GDP. Increasing returns have in practice boosted the static effects of trade liberalization: Harris and Cox (1984) found that a free trade agreement with the USA could boost Canadian welfare by almost 9 per cent of GNP. The dynamic effects of trade policy remain more elusive, at both a theoretical and an empirical level. It is true that if trade leads a country to specialize in less technologically progressive goods, this may imply dynamic losses that outweigh the static gains from trade (Young, 1991); and indeed, it is sometimes claimed that this is precisely what happened when Ireland specialized in linen rather than cotton in the nineteenth century. In a similar vein, it has occasionally been argued that only industry is capable of sustaining general productivity growth, and that economies such as Ireland, whose comparative advantage dictates that they should specialize in agriculture, will benefit from protection. On the other hand, if there are increasing returns to research, and trade leads all countries to put their research eggs in one basket, rather than spreading resources across sectors, then trade can boost growth everywhere (Davis, 1992). Trade in both goods and ideas will do the trick; the theory, by appealing to economies of scale in the production of new ideas, is reminiscent of Smith's dictum that the division of labour depends on the extent of the market (Rivera-Batiz and Romer, 1991). What the new theory adds to traditional, static trade theory is the possibility that trade will increase growth rates as well as income levels.14 More important for the debate than the theory, maybe, has been the plethora of empirical studies confirming that trade and growth go hand in hand. This work has taken the form of country studies, studying the effects of import substitution policies in several developing countries, as well as cross-country statistical studies.15 Ireland too has a failed import substitution experiment behind it; trade policy would certainly seem to offer a promising explanation for Irish retardation, at least during the 1950s. During the 1920s, Ireland continued to be largely free trading. After 1932, however, the De Valera government sought to create an import-substituting industrial sector. In an attempt to prevent tariff hopping by foreign companies, the government tried to restrict direct investments by foreign firms in the economy. Protection resulted in a very rapid increase in industrial employment and output in 1932-7, but that rise was not sustained and productivity remained low. Ireland remained protectionist until the late 1950s, much later than other European countries. It then progressively opened itself to the outside world, with tariff reductions, a free trade agreement with the UK in 1965, and eventually EC membership in 1973. The policy of protection brought in by De Valera in the early 1930s was not unique, and may have been a sensible response to the circumstances of the time (Neary and 6 Grada. 1991; O'Rourke, 1991). However, Ireland was slow to jump on the GATT bandwagon, only significantly liberalizing towards the mid-1960s with the Anglo-Irish Free Trade Agreement (AIFTA). It thus seems reasonable to enquire whether Ireland's failure to ride the postwar boom was due to its mistaken adherence to trade policies that had outlived their usefulness. The ratio of customs receipts to merchandise imports is a fallible,first-cutmeans
Irish economic growth, 1945-88 411 Table 13.7. A measure of two-way trade, North and South, 1926-91
South North
1926
1938
1951
1970
1980
1991
-42
-47 -11
-68 -8
-43
-31 n.a.
^23 n.a.
+3
+2
of tracking protection. That ratio rose from 12 percent in 1929 to 24 percent in 1939. In 1960 it was still 20 per cent, but by 1970 it had fall to 13 per cent, and by 1990 to 0.9 per cent. The fall in the effective rate of protection was greater still (McAleese, 1971).16 The ratio of merchandise trade (imports plus exports) to GDP rose in tandem from about 25 percent in 1945 to 55 percent in the 1950s, to 69 percent in 1960-73, and 106 per cent in 1974-89. A small but indeterminate part of the rise was due to transfer pricing, since tax relief on profits derived from exports prompted multinational corporations to declare low prices for imported raw materials and high prices for exports. The result was inflated value added, an artificially high return on capital, and exaggerated export and output levels. The rate of growth over the long haul is probably not much affected, however. How did protection affect the Irish economy? It certainly made it a lot more inward looking. A useful comparison here is with Northern Ireland, which as part of the UK was compelled to maintain free trade with Britain throughout our period. The trade ratio suggested by Kennedy et al. (1988: 234, 239) (defined here as (X — M)l (X + M ) for manufactured goods) effectively captures the contrast. The home market orientation of the new southern industrial sector caused the ratio to drop, while the more recent rise in the ratio captures the transition to a more outward-looking industrial sector. For the North, however, the two-way trade in manufactures was broadly in balance throughout (Table 13.7). The resulting industrial structure- a large number of widely dispersed small firms and a low degree of horizontal integration - reflected both the powers of local pressure groups and the small size of the local market. The following extract from Mary Daly's recent book on the 1930s will give a flavour of the distortions involved (Daly, 1992: 108; see also 6 Grada, 1994b: ch. 15): J.H. Woodington was refused permission to build a tannery adjacent to his Drogheda shoe plant or in an adjoining town. He was informed that the minister preferred a town that did not have an industry and was dispatched to Mountmellick following local representations, despite the lack of a suitable site. When he turned his attentions to Portlaoise because of an offer of local capital, officials proposed Tralee, 'where all the capital necessary would be available'. Few other sizeable towns in Ireland are as far from Dogheda as Tralee! Such a structure was unlikely to support the innovation necessary for sustained economic growth. It was also bound to deprive Irish industry of precisely the sorts of external economy of scale (skilled local labour forces, for example) that have been used to justify industrial protection. Why then did Ireland take so long to liberalize? Although the protectionist policies pursued by successive administrations since the early 1930s always had their critics, evidence of the damage caused was elusive
412
Cormac 6 Grada and Kevin O'Rourke
before the 1950s. As mentioned, the tariffs imposed in the early 1930s had initially produced a sharp rise in industrial employment. Admittedly, it did not take capital long to absorb the sheltered home market, and industrial employment had already peaked before 1939. However, the message that import substitution could not have produced sustained economic growth was blurred by the enforced autarky of 1939-45, and the postwar recovery produced the illusion that protectionism was doing no harm. 1946-51 was the first five-year period since the Famine to experience population growth. Industrial employment rose considerably, industrial profits rose, and the rate of economic growth was respectable by European standards (Hall, 1951: 4-5; Lynch, 1969: 189). It was only when the rest of Europe left the Irish economy standing in the 1950s that the bankruptcy of the old policies became clear to policy-makers. Policy-makers and most opinion-makers in the early 1950s took protection for granted. The Commission on Emigration (1956:161) defended tariffs by noting that without them 'it would be difficult to conceive of industry on any wide scale maintaining itself or developing further', and insisted that future commitments to international agreements should not compromise 'our freedom to develop our industries as we think fit'. To those economists who continued to support free trade, in the mid-1950s it was still 'an unlikely Utopia' (Lynch, 1969). Even Whitaker's landmark Economic Development (1958), the government report which paved the way for the new economic policies of the 1960s, was circumspect about the issue. Noting that 'the coming of free trade in Europe in one form or another must be faced in due course', its main emphasis was on the need to allow in foreign capital rather than on abolishing tariffs (Whitaker, 1958:190; see also Whitaker, 1956). Indeed, the new Industrial Development Authority, established to attract direct foreign investment, sanctioned some tariff increases in its early years (Whitaker, 1958: 13), and neither the Central Bank nor the Department of Finance proposed freer trade as a panacea in the early 1950s. Policy-makers emphasized instead the need for state mobilization of investment funds and demand management. It took time for the argument for freer trade to sink in. Free trade appears to have helped boost Irish growth, as a glance at Figure 13.2(c)-(0 will suggest. The distance between Ireland and the 'convergence lines' implied by experiences of a large number of European economies narrowed after 1960. On the other hand, Ireland continued to underperform. Perhaps this was because free trade did not mean an end to government distortions. For example, it was surely no accident that the new tax incentives and grants introduced to attract foreign investment in export-oriented industries were geared to complement, rather than substitute for, existing inward-looking Irish-owned industries (Lee, 1989: 352). To this extent the Irish campaign to attract foreign investment was constrained from the start by the protectionist legacy. The shift to 'free trade' in Ireland was matched by a big increase in non-tariff distortions in the form of a huge rise in aids to industry. In 1983-6, 3 percent of GNP was being spent on promoting industrial development, equalling about one-eighth of industrial value added (Leddin and Walsh, 1992:482). Bond and Giesinger (1985:96) state that 'Investors receiving the maximum cash grant are being protected, in effect, with the equivalent of a 24% tariff on top of an average 27% tariff afforded by the EC's common external tariff' Such government distortions, falling under the general rubric of 'industrial
Irish economic growth, 1945-88
413
policy', may simply substitute for outright protection; if so, they may lead to a similar misallocation of capital. More importantly, by creating an impression that bail-outs might be available for underperformers, they might reduce innovation, an effect stressed by Michael Porter (1990). Furthermore, direct government subsidies to industry might be even more 'capturable' by interest groups than across-the-board protection; rent seeking might be as much or more of a problem, with all that this can imply for growth. It is to rent seeking, and other political economy factors, that we now turn. 7
Rent seeking and interest groups: the political economy of growth
The rent-seeking approach associated with Mancur Olson (1982) seems to offer a good framework for interpreting Ireland's relatively poor economic performance since 1945. Olson predicts, in the absence of disturbing causes, powerful tendencies for sectional interest groups to become entrenched in an economy. These groups will tend to cause institutional sclerosis and economic stagnation. According to Olson, societies fortunate enough to have been spared military invasion or serious political unrest for a long time, such as the United Kingdom, pay a price in that their very stability gives such interest groups ample scope to plan collective actions which restrict competition and retard growth. Only an external shock such as defeat in war or economic and political integration can destroy the influence of such groups. Aspects of the Irish record invite a reappraisal of this last hypothesis. First, the political convulsions of the nineteenth century may, contrary to Olson, have encouraged rather than hindered rent-seeking behaviour. The Land War, which pitted the peasantry against the land-owning classes, was eventually won by the former, after a series of rent strikes, some violence and much upheaval. Courts were instituted to set rents, and later the landlords were forced to sell their land at favourable prices, with the taxpayer helping to smooth the transaction. Barbara Solow, for one, has argued that this was disastrous for Irish agriculture: Incentives to adjust the economy in the face of new international conditions were to some extent paralyzed. There is no need to take too seriously landlord contentions that everybody rushed to court and neglected his farming, but if tenants could increase income more by litigation than by changing agricultural techniques, they would certainly do so. If valuers were swayed by appearances, a premium was even put on worse farming, and consequent dilapidation... with the tenants of Ireland crowding into court, no one was thinking about agricultural education, credit and marketing programs, improved cropping, selective breeding, and, in general, ways of assisting tenants to adjust to changed economic conditions. (Solow, 1971: 165-6) Could it be that experiences such as this had a long-run impact on the culture, leading entrepreneurial talent to focus on 'the grant' or some other government concession, rather than competing on world markets? A second point to make is that Ireland did undergo a violent revolution and civil war during the period 1919-23. Why then was rent seeking a problem in the new Ireland? Olson himself (1989) has questioned whether independence 'destroyed the organizational structure of Ireland'. Most historians would agree: the Irish
414
Cormac 0 Grada and Kevin O'Rourke
revolution left property, the legal and banking systems, industrial relations and the civil service largely intact. The new regime sought to build up its reputation through institutional continuity wherever possible, so in the sense of Olson the revolution may not have been thoroughgoing enough. The protectionist coalition formed in the 1930s, which would soon command cross-party support and survive into the 1950s, would seem to fit Olson's model quite well. Urban workers, small farmers and manufacturing interests benefited from De Valera's policy of protection, sweetened with the 'farmers' dole', and together formed a powerful alliance (O'Rourke, 1991). Indeed, Olson (1989) pointed to free trade and EC membership as a means of ridding Ireland of the anti-growth coalitions that had become entrenched since the 1930s. If the maintenance of protection was due to a powerful Olsonian coalition, one might have expected to seefierceresistance to the liberalization that did eventually occur. Yet when the inability of the strategy of import substitution to sustain growth and employment became clear in the 1950s, the resistance offered by Olson-type lobbies was surprisingly weak. This supports the interpretation of the delayed liberalization put forward in the previous section: policy-makers had simply been slow to learn that protection was mistaken. In support of the Olson model, however, it could be argued that the scheme of adaptation grants to threatened firms instituted in 1962 and continued until 1967 was part of the price paid for the support or acquiescence of business and trade unions. The grants provided subsidies to firms which modernized their plant and equipment. The scheme followed the reports of the corporatist Committee on Industrial Organization, which revealed an indigenous industrial sector unlikely to cope with free trade (Kennedy et ai, 1988: 68). Initially intended as a temporary measure, the scheme took on a life of its own. Given the poor prospects and poor record of the majority of firms involved, these transfers must have yielded a very poor return. A recent paper by Patrick Honohan (1992b) has listed several of the more blatant examples of rent-seeking behaviour in Ireland. Semi-state companies, granted a monopoly position in the Irish market, have channelled rents into cost increases rather than profits for the Exchequer. A small state-owned enterprise making a worthless petrol additive was supported by making the additive compulsory in all petrol refined in Ireland. During the recent currency crisis, hand-outs were granted only to thosefirmswhich had been careless enough to leave themselves exposed to fluctuations in the Irish pound/sterling exchange rate. Honohan might also have mentioned the fact that allflightsbetween Ireland and the USA were for years forced to stop at Shannon Airport, as a result of some extremely effective lobbying by local interests. One institutional feature of Irish life that has often been blamed for some of the rent seeking that occurs in the country is the multi-seat constituency, which pits party member against party member, and places a premium on 'constituency service'. It also makes for more marginal seats, giving greater weight to regional issues than would be the case under either a 'national list' or a single-seat constituency system. For example, the Shannon stopover was supported even by the laissez-faire Progressive Democrats, presumably because two of its deputies represent the Limerick area.
Irish economic growth, 1945-88
415
Institutions and politics can matter in other ways. Did Ireland's pay bargaining structure retard output and productivity growth? The Olson model would seem to have a definite bearing here. Olson (1982) argues that interest groups such as trade unions and producers' groups inflict most damage when they are big enough to cause widespread disruption, but small enough for the social cost of their actions to remain an externality to themselves. In this view, either very weak or all-powerful lobbies may be preferable to something in-between. Thus it is often argued that corporatism works, both in pay bargaining and strategic policy decision making, because it takes account of macroeconomic constraints and minimizes the risk of inter-union disputes. Most analysts consider the cost in reduced wage dispersion worth paying. In Ireland, the experience of the Committee on Industrial Organization, which united management, unions and public service in analysing the shortcomings of protected industries in the early 1960s, and in proposing rationalization schemes, seems a good case in point. But more recent Irish experience in the pay bargaining sphere provides less cause for cheer. In an economy with high unemployment such as Ireland, the argument in favour of centralized agreements is that they should produce lower wage increases than plant-by-plant collective bargaining. However, a recent assessment of the Irish experience with centralized bargaining concludes that the outcome 'in terms of pay increases was higher than a decentralized wage bargaining situation' (Durkan, 1992: 349).17 Before the 1960s, old-fashioned collective bargaining was the norm in Ireland. The first attempt at a centralized agreement, the National Wage Recommendation of 1964-6, led to trade union suspicion of incomes policy, and decentralized bargaining followed until 1969. A serious industrial dispute prompted the re-emergence of centralized bargaining in 1969, but the rationale was good industrial relations, not full employment or inflation. Subsequent pay agreements stipulated a national norm, but also allowed for local bargaining to obtain further increases. This combination of centralized and decentralized bargaining arguably produced the worst of both worlds. The 'National Understanding for Economic and Social Development' of 1978 seemed to mark a watershed, in that it explicitly recognized the connection between pay and employment levels. However, the ensuing wage increases belied that view. The breakdown of the Second National Understanding in 1982 marked the end of centralized bargaining until 1987, with employers refusing to have any further truck with it. A sharp decline in wage inflation ensued, and there was industrial peace. Centralized bargaining resumed in 1987, according to Durkan 'once again accelerating] the pace of wage inflation'(Durkan 1992: 362). His assessment of the Irish experience with wage bargaining is cast in a very Olsonian mould: 'it is no surprise that once lobby groups get together the result is suboptimal. The result of bargaining will suit lobby groups, but how does it suit those excluded from the negotiation?' (Durkan, 1991; cited in Durkan, 1992: 363). Where does this fit in with the broader European experience with corporatism? Ireland does not feature in Calmfors and DriffiU's recent comparative study of the bargaining structure and economic performance in eighteen OECD countries (Calmfors and Driffill, 1988). Ireland's industrial relations make it an 'intermediate economy' in their framework, and adding it to their sample corroborates their case for a hump-shaped relation between the degree of centralization and economic
416
Cormac 0 Grada and Kevin O'Rourke
Table 13.8. Calmfors-Driffill indicators of economic performance Unemployment Employment Okun index rate L A Centralized 4.0 economies B Intermediate 6.1 economies Ireland 11.0 B' (including Ireland) 6.9 C Decentralized 5.8 economies
Alternative performance index
C
L
2.3
72.5
2.7 13.0
6.1
6.2
3.6
4.8 5.0 4.9
60.9 54.6 59.9
- 3 . 2 14.5 — 5 9 23.7 - 3 . 7 16.0
8.7 9.8 8.9
7.7 20.9 9.9
6.5 10.5 7.1
2.9
65.8
-1.1
15.2
7.7
5.8
3.1
C
L
C
L
C
L = levels (1974-85 average). C = changes (1974-85 average less 1963-73 average). Note: The Okun index is the unemployment rate plus inflation; the alternative index is the unemployment rate plus the current account deficit as a percentage of GDP. Sources: Calmfors and Driffill (1988: 20); European Economy, no. 50. performance. In Table 13.8, rows A, B and C reproduce Calmfors and DriffilFs results for centralized, intermediate and decentralized economies; row B' shows the effect of including Ireland in the intermediate category. However, quite apart from the possible influence of other factors, the difficulty with measuring the degree of centralization precludes an attempt at measuring the importance of this effect. A related study by Freeman (1989:76) shows Ireland in the uncomfortable position of having moderate wage dispersion coupled with the lowest employment rate in a 19-country sample. Analysis of protection and industrial relations thus yields modest support for the Olson model. However, this model is notoriously difficult to test satisfactorily. Magee et a\. (1989: ch. 8) invoke the ratio of lawyers (a measure of unproductive lobbying activity in an economy) to physicians (representing productive white-collar workers) as an index of rent seeking, and find that per capita income in countries with more lawyers per doctor has grown more slowly in the period 1960-80. High-tariff countries also have more lawyers. Ireland is one of thirty-four countries to feature in their analysis. However, cross-country variations in the structure of these professions vitiate the exercise: in Ireland, for example, the legal profession has always restricted entry, while in the United States, the legal profession is relatively open but the American Medical Association operates a monopoly (compare Walsh, 1993a).
8
Some simple cross-section evidence
Finally, for further comparative focus, we subjected an 18-economy data set, based largely on Heston-Summers Mark 5, to statistical analysis. The strategy wasfirstto
Irish economic growth, 1945-88
417
gauge Ireland's status in the European 'convergence club' detected in section 2, by regressing growth in per capita GDP on initial GDP and an Irish dummy variable. We then added further explanatory variables, in an attempt to reduce the size of the (negative) coefficient on the Irish dummy. These variables were: the investment share of GDP (INV, taken from the Penn Table); a scale variable, as in Helliwell (1992); the share of GDP accounted for by agriculture (AG); secondary school enrolment rates (SEC, taken from the UNESCO Survey of Education (vol. 3) and the UNESCO World Education Report (1991)); a trade variable, TRADE, defined as the sum of merchandise imports and exports, and reflecting the well-known fact that trade is more important to small countries than to large; a proxy for corporatism (CORPOR), defined rather crudely by a dummy variable (since theory suggests that both highly centralized and highly decentralized systems of industrial relations work well, economies in those categories are given a value of one, and the rest (including Ireland after 1960) a value of zero); and a dummy variable, PERIPH, which equals one if the economy is geographically peripheral, in the sense of comprising part of Western Europe's external border, and which equals zero otherwise. The results for the period 1973-88 must unfortunately be treated with caution, due to the uncertainty over the Irish GDP data mentioned in section 2, and the steadily evolving GDP/GNP gap over the period. Tables 13.9-13.12 present the results of the analysis. We estimated variants of the following: GROWTH = a0 + ^(SCALE) + + a3|>(INV/GDP) - ln(n + g + 5)~] + a4[ln(SCHOOL) - ln(n + g + ci(*iM)
(2)
where sci is the share of the ith input in terms of the value of total costs However, once the above assumptions are relaxed, their impact on the productivity residual must be explicitly recognized. Returns arising from market power (imperfect competition), from varying utilization of inputs (short-run fixities), as well as from technological characteristics (scale economies) have to be accounted for, and the traditional productivity growth measures amended correspondingly. Therefore, let the reference model be given by a profit-maximizing factor demand model under unrestricted returns to scale and quasi-fixed inputs. Imperfect competition is assumed to prevail in the output market. In this context, the functional characterization of the technology of a representative firm in short-run equilibrium is given by the restricted profit function or also by the restricted cost function. The latter (cv — cv(w9y9k,t), otherwise known as the variable cost function, defines the minimum cost thefirmincurs for a given volume of output (y), for a given vector of variable factor prices (the / x 1 vector w), a given state of technology (t) and a given level of quasi-fixed factors (indicated by the M x 1 vector k). The restricted cost function allows us to distinguish neatly between variable factors of production (utilized in amounts corresponding to minimum variable costs) and quasi-fixed ones (utilized in amounts that do not necessarily correspond to minimum costs, but are typically characterized by some sort of partial adjustment). Morrison (1992a) has shown that, in order to recognize sub-equilibrium and scale economies, dual (cost-based) measures of productivity growth should be adapted as follows:2
[(
) ^
(3)
where (1 - £e c k , = 1 — £(dlncr(w,r,/c,^O/dln/c./)) *s a n index of factor utilization rate j
j
and ecy denotes the long-run elasticity of costs with respect to output and is an obvious indicator of scale economies. The two indexes adjust the traditional dual
432
Nicola Rossi and Gianni Toniolo
measure for short-run fixities and non-constant returns, respectively. The corresponding primal measure of productivity growth is given by
ey, = -ecl + ec,{l - 1(1 + epy)/Ul - ^
JecJJ}
(4)
where one plus the inverse demand elasticity facing thefirmepy is a natural index of market power.3 Equations (3) and (4) summarize the implications of sub-equilibrium models for productivity growth measurement. In order to isolate exogenous technical change, they account for the impact on the cost structure of (1) capacity utilization, (2) returns to scale and (3) market power. The 'economic' index of factor utilization (1 — JX*,) represents the outcome of j
explicit economic optimization processes. For a given 'potential' output level, it expresses the economy's ability to adjust to sudden changes in aggregate demand. For a given aggregate demand, however, it shows the economy's optimizing reaction to changes in input prices,fixedfactor quantities, non-constant returns to scale and adjustment cost. This 'economic' index of capacity utilization provides information on both cyclical and trend factors affecting the economy. In this respect, the importance of allowing for capacity utilization in the assessment of long-term productivity growth should not be overlooked: economic historians have long recognized that undercapitalization and overcapitalization are far from being purely cyclical phenomena. The persistent changes in relative factor prices in the 1970s and 1980s provide a clearcut example. Far more difficult to pinpoint is the content of the indicator of scale economies: that is, of the long-run elasticity of costs with respect to output, ecy. At the aggregate level, returns to scale may be associated with a large number of factors other than internal firm scale economies.4 The possibility arises, for example, if one firm's economic activity or accumulation of knowledge conveys external benefits to other firms. Similarly, externalities linked to market thickness or infrastructure expenditure may induce scale economies. Unfortunately, the indicator generated by the above theoretical framework captures the direct effect of scale on costs, but does not provide any information on the nature of that effect. From the economic historian's point of view this is more unfortunate, since different determinants of scale economies at the aggregate level may well lead to substantially different views of the growth process in specific countries. Therefore, the information contained in the scale economies indicator needs to be supplemented by additional qualitative evidence on factors which could be responsible for aggregate non-convexities.5 The excess of price over marginal cost provides a straightforward indicator of market power. In principle, mark-up levels respond to demand and supply changes facing the firm. Indeed, considerable efforts have been devoted in recent years to assessing the behaviour of the mark-up over the business cycle. In what follows, however, the pattern of aggregate mark-up behaviour will only be allowed to react to a limited set of demand-side variables. The three indicators just mentioned combine into a fourth index measuring economic profitability:
py/c, = ^ 1 - Ee c ^6 c y - 1 j/(l - €py)
(5)
Italy
433
Table 14.5. The sources of growth: Italy, 1938-90 (revenue side, % per annum)
Output Labour inputs Energy inputs Non-energy imports Private capital Public capital Productivity residual €
t
1938-90
1950-90
1950-73
1974-90
4.0 0.4 6.3 7.0 3.8 2.7
5.3 0.9 7.9 8.5 4.6 3.1
6.89 1.1 14.1 10.3 5.3 3.4
3.2 0.7 -0.2 5.9 3.6 2.6
1.8 0.2
2.7 0.4
4.0 1.1
0.9 -0.7
Note: Average annual compound growth rates. Equation (5) indicates the way profitability relates to market power and shows that mark-ups may actually be consistent with zero profits because profits generated by market power may turn out to be counteracted by the effects of short-run and long-run fixities. For example, if the economy is undercapitalized, profits may well be higher than under full long-run optimization. Similarly, if returns to scale are permitted, the profitability arising from market power may be further reduced. If reliable proxies were available for the indexes measuring the deviations from Solow's assumptions, then computing adjusted productivity growth measures would be immediate. Unfortunately, this is not the case for most if not all historical data sets. It is therefore necessary to specify parametrically the representation of the firm's technology, and to revert to econometric methods to obtain coherent estimates of the relevant magnitudes. The specification and estimation of the relevant econometric model as well as the underlying statistical information are briefly described in the appendix.6 The model assumes three variable inputs (labour, energy and non-energy imports) and two quasi-fixed factors (private and public capital), and its parameters have been estimated over the sample 1892-1990 (allowing for the shorter time span for which labour input information is available: that is, 1907-39). The model appears to perform remarkably well. In particular, it is worth noting that, thanks to the highly flexible specification of the technology, no structural break (in a statistical sense) shows up in the period under examination. Nevertheless, the estimated coefficients turn out to imply substantial variations in key magnitudes, such as own- and cross-price elasticities. Tables 14.5 and 14.6 provide a full decomposition of output and cost growth over the period 1938-90.7 Over the period 1950-90, output growth rate exceeds 5 per cent per annum, despite a limited contribution of labour inputs. The overall contribution of factor inputs reaches about 2.5 percentage points per annum. As a result, total factor productivity (TFP) growth, as reflected in the traditional measure of the Solow residual (measured as potential output growth for given factor inputs: that is, by the primal measure eyt) amounts to almost 3 per cent per annum. Increases in the efficiency of production over time would therefore apparently account for more than 50 per cent of total output growth. Evidence of a multifactor productivity
434
Nicola Rossi and Gianni Toniolo
Table 14.6. The sources of growth: Italy, 1938-90 (cost side, % per annum)
Total cost Output price Labour price Energy price Non-energy imports price Private capital user cost Productivity residuals
1938-90
1950-90
1950-73
1974-90
19.5 14.0 18.6 12.1 12.5 15.5
13.4 7.4 12.1 5.0 5.8 8.7
11.0 3.6 9.8 -1.7 2.5 3.7
17.1 13.0 15.4 15.3 10.6 16.1
-3.0 -2.3 -0.9
-3.0 -1.9 -0.5
-4.4 -3.3 -1.4
-1.0 0.0 0.9
Note: Average annual compound growth rates. 1.41.21.00.8 0.6 0.4 Dual TFP unadjusted Dual TFP adjusted for short-run fixities Dual TFP adjusted for short- and long-run fixities
0.2 0
1940
1945
1950
1955
1960
1965
1970
1975
1980
1985
1990
Figure 14.1 TFP levels: Italy, 1938-90 (1985 = 1) slowdown shows up in the early 1970s. Between 1974 and 1990, despite a stable contribution of labour inputs, output growth reaches 3 per cent, while TFP growth falls to less than 1 per cent.8 On the cost side, the evolution of output and factor prices conveys a similar story. Throughout the period, average costs grow at a rate of about 8 per cent per annum, in spite of much larger increases in labour costs (more than 12 per cent per annum over the period 1950-90). As a result, the dual TFP measure (as given by potential cost reduction for given factor input prices, ect) averages about 3 per cent over the whole period. However, it reaches only 1 per cent in the post-1974 period. As it turns out, though, much of this pattern simply reflects returns to scale and adjustment costs (short-run fixities). Once adjusted for the latter, the dual TFP growth measure (€ct) is severely reduced, with the period 1974-90 showing no 'true productivity' gains. Moreover, when the adjustment takes into account both
Italy 435 short-run and long-run fixities (ec,), total multifactor productivity growth is nothing but a tiny fraction of output growth.9 Over the period 1950-90, out of 5 percentage points of average output growth, only one-tenth may be imputed to the 'adjusted' Solow residual: that is, to exogenous shifts of the production frontier. One-third (that is, 1.5 percentage points on average) may be attributed, instead, to the scale economies emerging at the aggregate level. The adjustment of the multifactor productivity residual also provides information on the nature of the productivity slowdown which apparently concerns only the exogenously shifting component of technology. In the late 1960s, fully adjusted productivity levels reach a maximum in 1968 and taper off (Figure 14.1), with the corresponding residual turning from usual negative average values to positive ones. As a result, during the period 1974-90, cost decreases (productivity increases), instead of originating from technical change, can be attributed entirely to the combination of scale economies and short-run adjustment costs.10 Interestingly, the timing implied by unadjusted figures turns out to be highly misleading. Unadjusted productivity levels reach a maximum in 1974. They subsequently stagnate for almost a decade to show remarkable increases after 1982. Quite clearly, unadjusted figures end up by magnifying the role of thefirstoil shock, while fully adjusted figures point to the 1960s as the roots of the malaise. The productivity trends just mentioned originate from structural characteristics of the Italian economy prevailing throughout the period 1950-90, which may be summarized as follows: • During the whole half-century, revenues never fell short of covering costs of production, including appropriate returns to capital (Figure 14.2). Over the period 1950-90, the product price exceeded average costs by 20 per cent on average, thereby giving rise to economic profits. Profit margins reached a high during the war and then rapidly fell to a local minimum in the early 1950s. The following upward trends ended in the early 1980s, when rising real interest rates led to a further unprecedented fall. • The profitability slowdown of the early 1950s can partly be explained by declining market power deriving from profit-maximizing behaviour under monopolistic competition. However, the representative firm's market power, as given by the ratio between output price and marginal costs, (1/(1 4- epy) (Figure 14.3), while changing considerably over time, remained of sizeable magnitude during the whole period under investigation, averaging about I.6.11 • On the cost side, the post-1945 Italian economy appears to have been characterized by substantial long-run returns to scale. The inverse long-run cost elasticity with respect to output (Figure 14.4) averaged 1.3 in the period 1950-90.12 • Finally, in the late 1950s, the Italian economy turns from undercapitalization to excess capacity, the market price of capital substantially exceeding its shadow price on the margin. Since the capacity utilization index, 1 — £e ckj (Figure 14.5), j
reveals whether or not the economy operates at the desired level of capital stock, the capital level observed throughout the 1960s, the 1970s and most of the 1980s turns out to exceed the desired (optimum) level, despite decreasing social-overhead capital formation.13
436
Nicola Rossi and Gianni Toniolo
2.12.01.9 1.81.71.61.51.41.3 1.21.1 1.0 H 0.9
1940
1945
1950
1955
1960
1965
1970
1975
1980
1985
1990
1970
1975
1980
1985
1990
Figure 14.2 Profitability: Italy, 1938-90 1.80 n 1.75-
1.70
1.65
1.60-
1.55-
1.50 1940
1945
1950
1955
1960
1965
Figure 14.3 Market power: Italy, 1938-90
Italy 1.7 1.61.5 1.41.31.21.1 1.0 0.9 0.8
1940 1945 1950 1955 1960 1965 1970 1975 1980 1985
1990
Figure 14.4 Scale economies: Italy, 1938-90 1.41 1.3 1.21.11.00.90.80.7
1940 1945 1950 1955 1960 1965 1970 Figure 14.5 Capacity utilization: Italy, 1938-90
1975 1980 1985 1990
437
438 Nicola Rossi and Gianni Toniolo 4
The legacy of Fascism and the war
Three features of the economy during the 1930s had a long-run impact on the postwar economy. The first relates to a particularly strong 'autarkic' reaction to the difficulties of the 1930s. Not only were trade restrictions such as quotas and bilateral agreements applied to a large extent, but exchange controls became increasingly binding from 1934 onwards. Moreover, the Fascist government set out to introduce a large number of administrative controls over almost every aspect of economic life. The second point to be recalled here concerns the counter-cyclical hiring policy in the public sector, introduced by Fascism during the Great Depression (Piva and Toniolo, 1988) as part of its successful consensusseeking drive aimed particularly at the lower-middle classes. The third event of far-reaching importance was the huge lending of last resort undertaken in 1931-3 in order to avoid the liquidation of the two largest (German-type) banks in the country and thefinancialpanic that was likely to follow. As a result, a state-controlled agency, Istituto per la Ricostruzione Industriale (IRI) was created to own and manage a large number of manufacturing and utility companies previously owned by the banks (Toniolo, 1993). Against this background, it is possible to single out those legacies of the 1930s that are likely to have affected Italy's potential for growth after 1945: • A technology gap developed in most manufacturing industries relative to most advanced countries in the West, as witnessed, for instance, by an inquiry into the engineering sector carried out by a Stanford team in the early 1950s; in this respect, the fact that only a small part of industrial equipment was destroyed during the war turned out to be a mixed blessing. • A large body of legislation was in existence in 1945, granting the government power for a rather minute regulation of the economy. • Public administration came to be thought of and organized as a way to create consensus through patronage rather than primarily as an instrument of cost-effective allocation of public resources. • The size of public ownership in manufacturing and utility companies was by far the largest in the western world; the credit sector was almost entirely under direct or indirect government control. Last but not least, postwar Italy inherited from the past the large gap in production and income that characterized the South relative to the rest of the country. According to one estimate, in 1938 'the total per capita income in Southern Italy was only just over half (55 per cent) that of Northern Italy' (Molinari, 1949:31). One must recall, however, that the origin of this peculiarity of the Italian economy dated back several decades before both the Great Depression and Fascism.14 The effects of the Second World War on the productive potential of Italy have not been the subject of a study as careful and detailed as the importance of the issue would suggest. In general, however, it is safe to say that 'the bulk of Italian industry survived the war largely intact' (De Cecco and Giavazzi, 1992). Social overhead capital, particularly roads, railways and bridges in the Centre-North, was more penalized than industrial equipment; its repair, however, required mostly pick and shovel work, then an abundant factor in Italy, and could easily be carried out.
Italy
439
Therefore, in the aggregate, the war did not result in major physical long-term bottlenecks for growth. 5
Setting the stage: reconstruction and stabilization
At the end of hostilities, Italy's GDP was roughly back to the 1915 level,15 down 39 per cent relative to 1939, the best prewar year (Rossi et al, 1992). Recovery was slightly more rapid than in the case of Germany and Japan. Italy's GDP regained its prewar level sometime between 1949 and 1950; the other two defeated Axis powers, between 1951 and 1952 (Maddison, 1991). Real growth was particularly rapid in 1946 (25 per cent) and in 1947 (14.5 per cent). It slowed down in the following two years due both to the monetary squeeze undertaken to stop inflation and also to diminishing advantages deriving from 'closing the gap' by exploiting aggregate economies of scale. That the latter 'explain' most of the growth realized soon after the war is hardly surprising in the light of the effect of the war on Italy's productive potential. Since industrial equipment was largely intact, its productivity was bound to increase rapidly as repair work on roads, bridges, etc. proceeded, unifying the domestic market and re-establishing communication with the rest of the world (harbours and the like). Between 1945 and 1949 a number of decisions affecting long-term growth were made. In particular, foreign policy was firmly anchored in the western camp, and domestic policy was shaped around the priority given to the isolation of the Communist Party. Both choices had wide-ranging economic implications. After two years of coalition governments, which included both Christian Democrats and Communists, the latter were excluded from the De Gasperi cabinet which took office in May 1947. In the general election held in the following April, the Christian Democratic Party secured almost 50 per cent of the votes and the possibility to rule the country with the support of the small parties of the centre for the following fifteen years. The final decision about long-term foreign policy commitments followed: in March 1949, Parliament authorized the government to sign the North Atlantic Treaty. On the institutional side, it is worth noting that a new constitution was approved in 1947 by a large majority which included the Communist Party. As a reaction to twenty years of Fascism, the Constitutional Assembly strived to make sure that dictatorships or even authoritarian governments could never rule Italy again. Careful attention was therefore devoted to building a well-designed system of checks and balances between the various state powers and individual institutions, while at the same time disregarding the need for efficient government. It is easy to see the long-run economic implications of such an understandable ideological bias of the constitution's fathers. The main economic decisions made at the time appear to be strictly connected with the political ones. A bold pro-market, pro-trade liberalization decision was made as far as international economic relations were concerned. This was, to be sure, a revolutionary departure from a time-honoured protectionist stance dating back to the 1870s and 1880s. In March 1947, Italy applied for membership of the IMF, thus entering into an early commitment to fixed exchange rates and currency convertibility. The Marshall Plan, the OECD and the European Payment Union
440
Nicola Rossi and Gianni Toniolo
followed (on the latter, see Carli, 1993). Quotas and other non-tariff barriers to trade were abandoned. Import duties were lowered. In 1953, Italy joined CECA (European Coal and Steel Community). Monetary policy, after the 1947 price and exchange rate stabilization, was so consistent with the choice of an open economy that it gained sharp criticism from Paul Hoffman, the author of the official Country Study report on Marshall aid presented to the US Congress in February 1949 (ISE, 1948). The Italian authorities were accused of being obsessed with inflation and foreign currency reserves at the expense of growth in domestic demand. Domestic economic policy decisions were nowhere near as clear cut as those on the international front. If'democratic planning' advocated by the left and by some sectors of the Christian Democratic Party was ruled out, both the overall policy stance and day-by-day decision making were often inconsistent with the needs of a developing market economy. As mentioned above, postwar Italy had inherited from Fascism a highly regulated private sector, an inefficient public administration and large state stakes in manufacturing, utilities and banking. Both the left and, on different grounds, relevant segments of the business community argued that the administrative tools for state intervention in the economy should be left in place, to be used by democratic governments in the pursuit of their own goals, be they income distribution or protection of individual enterprises. A small if vociferous patrol of free-traders argued the opposite, but their political weight was negligible. Rather than being privatized, state enterprises came to be seen as a tool for economic development. Their number increased. Government subsidies to the South were made permanent with the establishment, in 1950, of the Cassa del Mezzogiorno. Public administration continued to be assigned the role of provider of income to the lower-middle classes: the number of civil servants rose from 16.7 per thousand inhabitants in 1937 to 23.2 in 1952 (Ragionieri, 1976). At the same time, nothing was done to pass anti-trust legislation, to create a stock exchange security commission or, more generally, to define clearly the rules of the economic game. The credit sector remained almost entirely under public control (either by the state or by local authorities, as in the case of savings banks). As Guido Carli (1993:3) put it, Italy's policy making was, like Faust, torn between two conflicting souls: the first made for a long-term commitment to trade liberalization and fostering of competition in the international markets; the second could not get away from state intervention and protection of vested interests on the domestic front. To be sure, this state of affairs was partly unavoidable and partly beneficial, at least in the medium term. Privatization of IRI companies was out of the question unless banks were allowed to step in, thereby recreating the unstable financial environment that had characterized the heyday of the 'German-type bank'. The latter had ended in such a disaster that the option was rightly seen as most undesirable. An active policy for the South was advocated from virtually every quarter of the political spectrum: besides being acceptable in principle on ethical grounds, it was made necessary to prevent creeping social unrest from becoming open. As for the public administration, the overall stability of the country made it necessary to avoid crushing the lower-middle classes. The fact remains, however, that vested interests were created which in due time were able to establish themselves at the very heart of political parties and trade unions. The same can be said about the reluctance of the main political parties to enhance competition in non-tradable
Italy 441 sectors: on the contrary, national and local governments alike imposed minute regulations on small shopkeepers, craftsmen and owner-occupiers in agriculture, which tended to shelter them from open confrontation in the marketplace. 6
An 'economic miracle'?
The years 1950-63 have often been described as those of the 'economic miracle'. To be sure, this period saw the single most outstanding performance of the Italian economy for a comparable period of time. GDP rose at the very respectable annual rate of 5.9 per cent. Productivity, as measured by our 'adjusted' residual, increased on average by 3.7 per cent per annum when the effect of scale economies is taken into account (ecr), the highest in the subperiods considered in this chapter. It is noteworthy, however, that 'pure' (exogenous) productivity growth was also exceptionally large: more than three times higher than the average for the whole postwar period (1950-90). Efficiency was thus increased both by 'endogenous' factors, mainly a swift reallocation of resources, and by 'exogenous' ones such as imported (unincorporated) technology. That this episode of remarkable growth was termed a miracle can be understood in the light of previous history and of the rapid transformation which Italy underwent in a very short time span. There was, however, nothing to it that economic historians may not try to explain using the tools of their trade. On the supply side, a number of reasons are given for Italy's outstanding economic performance during the 1950s and the early 1960s: • Due to the technology gap accumulated during the 1930s and the war, large productivity gains could be obtained from borrowed technology, particularly given the country's endowment in well-qualified engineers16 and the existence of a fairly disciplined and educated labour force. • Given the low marginal productivity of agricultural labour, the reallocation of manpower to manufacturing resulted in rapid increases in aggregate productivity per worker. • An elastic (Lewis-type) labour supply contributed to a virtuous circle whereby wages increased less than labour productivity, allowing profits, and thereby investments, to expand more rapidly than output (Kindleberger, 1967). • The rapid growth in social overhead capital augmented private sector productivity. • The newly found international and domestic stability (Boltho, 1982) unbound those Promethean 'animal spirits' of which Northern Italy suddenly discovered itself to be particularly well endowed. Favourable conditions prevailed on the demand side as well: • By the first half of the 1950s, Italy's per capita income reached a level characterized by a high (income) elasticity of demand for consumer goods: a large domestic market therefore existed for goods the production of which enjoyed particularly large returns to scale at the prevailing technology. • Aggregate domestic demand was sustained by public works; the construction sector boomed due to both private and public demand of subsidized homes with the so-called Piano Fanfani (Ciocca et al., 1975).
442
Nicola Rossi and Gianni Toniolo
• World-wide, demand for goods in which Italy enjoyed a comparative advantage increased rapidly (Lamfalussy, 1963), feeding into production scale and increasing returns. Finally, both the Bretton Woods setting and domestic fiscal and monetary policy were partly responsible for creating a productivity-enhancing environment: the 1950s and most of the 1960s were characterized by a stable foreign exchange at a slightly undervalued rate, by low and predictable inflation, and by the most moderate public sector borrowing requirement in the country's history (Ganugi and Toniolo, 1992). A policy of artificially low interest rates was enforced throughout the period, thanks to government direction of the banking system and to the existing controls on international capital movements. During 1958-63, 'almost Japanese' growth rates were attained: GDP grew on average by 6.3 per cent between 1958 and 1963, and by 7.6 per cent in 1961 alone. In less thanfifteenyears, labour markets turned from structural unemployment to full employment. In fact, one of the peculiarities of the experience of 1950-63 relative to long-term trends was the very rapid rate of growth of labour inputs (1.34 per cent per annum, as against 0.67 for 1950-90). At the same time, labour participation rates fell (from 66 to 51 per cent between 1958 and 1983) mostly as a (partly statistical) result of labour force migration from country to town. In 1963 the unemployment rate reached its lowest historical level: 2.6 per cent. More generally, changes in the structure of the economy and of society at large, which in the advanced countries of the West took several decades to be completed, were compacted into less than twenty years, during which Italy was abruptly turned from an agrarian economy (and society) into an industrial one. Given the rapidity of the aggregate transformation, it may come as no surprise that some areas and sectors did not keep pace with the overall advance. The most notable and well-known problem in this respect is that of the regional diffusion of industrialization, which moved from the North-West to the North-Eastern and Central areas, while, as shown in Table 14.3, the Southern provinces even saw a widening in their productivity gap with the North. But it was not only the South that lagged behind: a large part of the intendence did not keep up; in particular, the modernization of public administration was further delayed. 7
The roots of the malaise, 1963-73
From the point of view of aggregate performance analysis, the years 1963-73 did not mark a sharp break with the previous period. The slowdown in output and productivity was anything but dramatic, since both remained way above the secular trend. The annual growth rate in GDP was more than respectable, 4.8 per cent; that in cost-based productivity was about 3 per cent. The decline was more pronounced in 'pure' (exogenous) productivity than in that deriving from scale economies (endogenous). When seen in the light of Italy's economic history since the political unification of 1861, full employment was the truly revolutionary result of the 'miracle' years. It had a number of implications both in the short and, what matters more, in the long run. In the medium term, labour shortages in some areas and trades simply fed into
Italy
443
wages and output prices. In 1963-4, a severe monetary squeeze was put in place to cool down inflation and reduce the balance of payments deficit. Unemployment rose to 3.8 per cent in 1965, but a further decline in labour participation rates kept labour markets fairly tight throughout the 1960s. A capital flight of extraordinary magnitude accounted for an important part of the external account deterioration. Italian capitalists, unprepared as they were to face full employment, were slow to adjust to its most immediate political and industrial relations consequences. Trade unions gained wider consensus and stronger bargaining positions with approaching full employment, but they were not culturally and ideologically equipped to use their newly acquired power to foster growth rather than to push for drastic immediate changes in income distribution. The year 1962 saw the second highest number of strike hours in postwar history (up 411 per cent on the previous decade's average). More generally, the equilibria at the centre of the political spectrum, which had guaranteed stability since 1948, gave increasing signs of weariness starting in the late 1950s. In 1962, the first centre-left government was formed; as a price for its participation in the government, the Socialist Party asked for, and was granted, the nationalization of the electricity industry. Both facts were relatively harmless in themselves, except that they were taken by some as a sign that postwar stability was over: many of those who had either legal or illegal ways of taking some of their assets abroad did not hesitate in doing so. In 1964 a half-hearted attempt at preparing a coup by segments of the armed forces (Lanaro, 1992: 325) paved the way for a swift retreat of the reformist army. Despite a decade of centre-left governments committed to 'reforms', setting the rules of the economic game once more proved impossible. In the second part of the 1960s, the economy reacted in a textbook fashion to full employment and the attendant rise in wages: for the first and only time in Italy's modern economic history, labour inputs actually fell, while private investment, after a slump in 1964-5, grew at an annual rate of 8 per cent for the rest of the decade. To counteract the effects on employment of the substitution of labour for capital by the private sector, the government stepped up its own investment, this again being an obvious textbook reaction. On the whole, it may be that agents overreacted to full employment - a largely unanticipated situation - but the expected signs of their reactions were correct. In spite of this, the seeds of a number of future problems were sown during this period. The economic problems of the following decades had largely non-economic causes. Most of them can ultimately be traced back to a less than satisfactory adjustment of Italian society to the economic reality created by largely unanticipated rapid growth. Institutions, politics, public administration, unions of employers and employees, and ideologies remained to some extent those of an underdeveloped economy. This seems to us to be a crucial characteristic of Italian history during the second half of the 1960s and the 1970s. Later on, during the 1980s, it can be plausibly argued that it was the 'political order' alone which lagged behind the 'economic and social order' (Pombeni, 1993). The reasons for society at large not entirely keeping pace with a growing economy are far from being fully investigated. A few tentative explanations may, however, be advanced. Two of them bear directly on the way postwar growth proceeded:
444 Nicola Rossi and Gianni Toniolo • High returns could be drawn simply by shifting resources from one sector to the next: that is, by widening rather than deepening the capital stock. The economy itself was not, therefore, fully stimulating a transformation in social organization, in education or in ideology. • The 'dualistic' character of growth meant that a large portion of the country lagged culturally and socially behind whatever changes were induced by economic development. This was the part of the country where most of the public administration and a fair share of national political leaders came from. Apart from these, one can list a number of non-economic reasons why the whole of society and politics lagged behind. One of them has been already mentioned: the reaction of Fascism had built into the heart of most institutions a system of checks and balances which privileged control - and veto powers - over efficient government. The success of policy making came to depend - or was thought to depend - on the existence of a wide consensus from society at large. This meant that the more far-sighted politicians, and members of the ruling class, found it more difficult than elsewhere to carry out 'modernization' policies (i.e. in public administration) without having to compromise with conservative political and social visions. It is also argued that the Cold War and the existence of a strong Communist Party made true political competition impossible, the Christian Democrats being condemned to govern, the Communists to eternal opposition. While this is certainly true, the worst effects of such a situation on Italian society were mainly felt after the second half of the 1970s; during the earlier years, the economy suffered from the fact that, in a society where ideologies seemed to matter more than elsewhere, both the Catholics and the Marxists were singularly inept at leading the necessary cultural change. Finally, one should mention vested interest. When in order to achieve 'modernization' it is necessary to compromise with the most conservative parts of the society, when market, efficiency and similar concepts are, at best, regarded as dirty words, traditional interests (some of them legitimate and even progressive in an agrarian society) are compacted and strengthened; they are given positive support in becoming vested. It is then easy to see that, given the initial conditions and the character of the Italian 'miracle', the uneven growth of the economy on the one hand and of society and politics on the other can be seen, at least in part, as being endogenous to the process of economic growth itself. In this respect, the events of 1969-71, which help to explain the subsequent slowdown in the economic efficiency of the system, have their deep roots in the stubborn resistance of Italian politics and society to adapt to the ultimate result of previous growth. It has been argued that 'the crisis of the "model" of economic development that had led to the "economic miracle" becomes manifest in 1963 and irreversible in 1969' (Sassoon, 1986). In as much as the 'model' was based on a high elasticity of supply of labour, this is a safe conclusion. It is during the 'hot autumn' of 1969 that the effects of full employment on the Italian economy and society become unmistakably clear. During that year, 302 million hours of strike were recorded, up by 438 per cent on 1968.17 Strikes coincided with a more general upheaval in which the student movement played a non-negligible role (Gigliob*ianco and Salvati, 1980; Sassoon, 1986; Ginsborg, 1989; Lanaro, 1992). As far as workers were concerned, the
Italy
445
'hot autumn' was neither only about wages nor only about working conditions: it was primarily about a larger say in politics and in society, starting from 'control in the use of manpower'. As far as productivity growth is concerned, these developments meant a considerable decrease in the flexibility of the economy. This was unfortunate: exogenous shocks of unprecedented magnitude would soon require exactly the opposite course to be taken. The results of 1969-70 on labour markets can be summarized as follows: • Changes in labour organization became much more difficult to introduce, since they had to be negotiated at shop-floor level. • Rigidity was introduced in the wage system by making individual negotiations almost impossible. • Tight hiring and, particularly, firing regulations were introduced, and overtime was either abolished or severely constrained. • A Workers' Charter was approved by Parliament in 1970, which strengthened the bargaining power of workers and made a number of anti-trade union practices illegal. Moreover, it was during this period that a set of generous social policies were initiated which, in the long run, had an undesirable influence both on the efficiency and efficacy of public expenditure, and on its size. The changes introduced in the social security system set a time bomb into the state budget, with the introduction of generous state-financed pension schemes for private and public employees, and of special plans for self-employed workers, together with the increasing use of disability pensions for welfare purposes. All this was 4in connection with the switch in 1968 from a contributions-based system to an earnings-based system and the far-reaching reform of INPS pensions in 1969, in particular the introduction, and subsequent improvement, of methods of indexing pension benefits to prices and real earnings' (Franco and Morcaldo, 1990: 107). In general, it was from the early 1970s onwards that the expansion of public expenditure became increasingly dependent on vested interest pressures on decisionmakers, who responded as if no budget constraint applied (Giarda, 1988). The use of public expenditure as a tool for easing social conflicts was magnified, and efficient resource allocation was neglected (Brosio and Marchese, 1986). In the long run, however, as our productivityfindingsconfirm, the growing inefficiency of the public sector and public expenditure resulted in low-quality services and perverse distribution effects, in turn affecting social consensus itself (Musu, 1992). 8
Productivity slowdown, 1973-92
8.1
An overall appraisal
Between 1973 and 1989, Italy's GDP growth remained above the G7 average; it was in fact the second highest in the group (Table 14.1). According to our measures (Rossi et al.y 1993), which differ slightly from Maddison's and thus from those in Table 14.1, GDP average annual growth was 3.2 per cent in 1973-80,0.6 per cent in 1980-3, and 2.7 per cent in 1983-90. Capital input growth did not differ significantly from that in the 'high growth decades', while labour inputs increased at about the secular rate. Import growth declined sharply from the previous periods. Both
446
Nicola Rossi and Gianni Toniolo
Solow's and our adjusted residuals nose dive from the late 1960s for about fifteen years, showing a moderate recovery thereafter. Over the whole period 1974-90, Solow (cost-based) residual averages a growth rate of about 1 per cent per annum, whereas our adjusted 'exogenous' (pure) dual residual takes a positive value, indicating a net negative contribution to aggregate performance from 'exogenous' inefficiencies. Most of these we attribute to inadequate institutions and rent seeking in the protected service sector, both private and public (given the safe assumption that the growth rate of both incorporated and unincorporated technical progress remained positive). In order to understand the peculiar characteristics of Italy's productivity slowdown, one must consider (1) inefficiencies in the service sector and (2) the inadequate provision of social overhead capital. The share of the service sector in GDP was 31.9 per cent in 1951 and 51.1 percent in 1990 (43.0 and 65.3 respectively if public administration is included). Barca and Visco (1992) have provided convincing evidence that productivity growth in this largely non-traded sector was low, relative to most other European countries, particularly in those services - such as telecommunications - that are typical of the most advanced phases in capitalist development. Aggregate productivity was affected in two main ways: (1) by a sheer composition effect (i.e. an increasingly large share of GDP originated from a sector characterized by productivity growth lower than the EEC average); and (2) by the fact that, at the time of a major shift from making to buying in the supply of services to manufacturing, such services were characterized by low quality and high prices, due to lack of exposure to foreign competition. In fact, the market power of our representativefirmpicked up again in the mid-1960s after declining for over two decades. Another reason behind Italy's productivity slowdown is to be found in the inadequate supply of social overhead capital. Our estimates show that, at the on-going set of relative prices, agents consistently desired a higher level of public capital as a complement to private capital. This pattern of output and productivity growth is consistent with an interpretation of Italy's economic and political history that runs as follows. As mentioned earlier, the Italian economy was made less rather than more flexible in the years immediately preceding the first oil shock. During the same period, social and political instability increased. Trade unions took a more radical stance in order to cover their 'left flanks' and to compete with each other. In 1968 the 'organic' centreleft governments gave way to a succession of weaker centrist and centre-left coalitions (Sassoon, 1986), not authoritative enough as crisis cabinets. Inflation and low-cost government borrowing could for a while contain social discontent. The same can be said of the agreement signed in 1975 between employers and trade unions, with the highest political blessing, for a new and stronger form of wage indexation. After the second oil shock, the commitment to exchange rate stability provided an anchor for industrial restructuring and curbing inflation. When an antiinflationary policy stance was finally adopted, this resulted (1980-3) in slower growth than in most industrial countries.18 The subsequent recovery and relatively high growth (1984-9) derived in part from (largely state-aided) industrial restructuring and improved labour relations. The political parties in the ruling coalitions,
Italy
447
however, remained relatively weak and in competition with each other: therefore, they resorted to deficit spending in competing for popular consensus. For a while the fiscal stimulus thus fed into the economy resulted in relatively high growth rates. In the longer run, however, it precipitated the economy into the weak and edgy position of the early 1990s, which fed instability in the process of political transition. 8.2
Inflation, unemployment and industrial restructuring, 1973-82
The social and institutional developments of the late 1960s and early 1970s had made the Italian economy more vulnerable to exogenous shock. After the Yom Kippur crisis of 1973, fragile governments under heavy social pressure resorted to inflation and devaluation, as lesser evils then unemployment. In 1975, GDP actually fell for thefirsttime since the Great Depression and war, but average GDP growth for 1974-9 was far from negligible (3.3 per cent per annum), while both unemployment and inflation rose considerably.19 Investment, both private and public, was stimulated by low (at times negative, ex post) real interest rates. In spite of these soft policies, the country was hit by a wave of terrorism, both left and right wing, the answer to which was an experiment with 'national unity' governments (1976-9). While such governments proved again unable to cut inflation and the public sector borrowing requirement, they were capable of taking a far-sighted decision. Joining the EMS in 1978 was, in fact, a bold step. It looks even bolder considering the timing: social and political tension was at its peak. The issue was hotly debated. Thefinaldecision was made by the government amidst the opposition of powerful interest groups and against the advice of some of the most qualified members of the economic profession. The move had one political and two economic objectives: on the one hand, it seemed important to signal both at home and abroad a renewed interest in the politics of European unification; on the other hand, exchange rate stability was seen as a precondition both for a crack-down on inflation and for the creation of an environment that would stimulate industrial restructuring and productivity growth. The second oil shock made anti-inflationary policies less effective; at the same time, industrial restructuring became more urgent. Monetary policy succeeded in getting price increases back to single-digit growth only in 1984-5. The reorganization of a number of large and medium-sized enterprises was helped by the government in at least two ways: by providing generously for early retirement and special unemployment benefits (cassa integrazione), and by giving its political support to agreements aimed at changing the system of industrial relations and wage setting (including wage indexation) inherited from the 1970s. The reaction of Italian industrialists to falling industrial productivity, low profits and high corporate debt was, on aggregate, only half successful. While the large private companies were able to shed excess labour, to start new investment programmes, to redress thefinancialstructure and, therefore, to prepare themselves to harvest some benefits from the upturn of world production and demand, important problems remained unsolved which were to have a negative impact on productivity over the following years. Three stand out as being of paramount importance:
448 Nicola Rossi and Gianni Toniolo • Since 'restructuring' took place with the decisive help of the government, the long-standing unhealthy relationship between big business and politics was strengthened rather than severed. • While there was some restructuring of public enterprise as well, it did not go far enough. • Little attention was paid to such elements affecting productivity as R & D, vocational training, the school system and, in general, efficiency in the service sector, both private and public. In these circumstances, it is not surprising that the productivity upturn we observe from the mid-1980s onwards is extremely moderate. 8.3
Growth, deficit spending and increasing inequality, 1983-92
Italy too had its 'splendid eighties'. From 1984, the growth rate of GDP increased, reaching 4.1 per cent in 1987; private consumption grew at an even higher rate. Inflation fell from a two-digit figure to an average of about 6 per cent. On the political front, the Craxi cabinet, in office from August 1983 to April 1987, seemed to provide both stability and effective government: such qualities were magnified and almost unanimously praised by the media. In spite of all the glitter of extravagant consumption, newly found price stability (of a kind) and the much publicized sorpasso of the UK in GDP per capita, both the Craxi government and those that followed during the decade (led by the Christian Democrats) did little to tackle the causes of low productivity growth. According to a British historian, The 1980s look in many respects similar to the years before 1968; then too the economy was developing, material bases for reforms existed but the centre-left politicians missed that opportunity' (Ginsborg, 1989: 568). Another historian observed: 'Italians returned to the opulence of the early 1960s, without having redressed the fundamental anomalies in the economic fabric' (Lanaro, 1992: 450). The growth in domestic welfare was itself severely flawed by adverse changes in income distribution (Figure 14.6). From the late 1960s, the egalitarian bias of Italian society translated into unprecedented welfare expenditures and absolute (as opposed to proportional) wage indexation. As a result, notwithstanding the evolution of the labour market, economic inequality, as measured by the Atkinson index,20 remained relatively stable until the early 1980s. Similarly, poverty (head count) rates oscillated around 5 per cent until 1982.21 During the 1980s, however, both the poverty and the Atkinson inequality indices deteriorated relative to the previous decade. At the same time, unemployment remained stubbornly around 9 per cent from 1982 to the end of the decade: the highest level in postwar history. The inability or unwillingness of the governments of the 1980s to put their own house in order by reducing the government borrowing requirement is the best synthetic indicator of substantial impotence in dealing with the structural problems of the economy. The ratio of outstanding state debt to GDP rose from 51 per cent in 1982 to 102 per cent in 1990. According to our estimates, productivity growth remained extremely low due to 'exogenous inefficiencies', which we tend to attribute to public administration, to the low quality/high price of services, notably 'network
Italy 25 i
449
Poverty index (head count) Atkinson's inequality index
20-
15-
10-
5-
0
I
1973
1975
1977
,I
1979
1981
1983
1985
1987
1989
1991
Source: Rossi (1993).
Figure 14.6 Inequality and poverty: Italy, 1973-91
services', and to poor R & D. Attempts were made to increase efficiency in labour and capital markets; they were only partially successful in making them more competitive. As mentioned above, integration into the European economy widened the gap between those sectors, such as manufacturing, that were fully exposed to international competition and those that enjoyed natural protection (i.e. construction and most services). Overall, the market power of the 'representative firm' remained exceedingly high by international standards. Lack of competition explains a stubbornly high inflation rate compared to Italy's main competitors (Barca and Visco, 1992). Finally, mention should be made of social overhead capital formation. An inefficient public administration, high construction costs due to poor competition and kickbacks (the two being obviously related) and budget deficits account for public capital remaining below its desired (optimum) level throughout the decade. As private and public capital are complementary, here is an additional reason for growth remaining below its potential level. 9
Concluding remarks
As postwar history draws to an end, this long episode in Italian 'modern economic growth' looks like an undeniable success, in spite of its flaws and structural weaknesses. The significance of Italy's postwar achievement is evident in the face of the past performance of the Italian economy itself and of international comparisons. The question, however, remains: could Italy have grown faster? While no comprehensive answer can be offered given the complexity of the counterfactual involved, our results confirm an opinion widely held by economists and observers at large that the inefficiency of the public sector, lack of competition in sheltered industries, and corruption account for at least part of the productivity slowdown
450
Nicola Rossi and Gianni Toniolo
which is observed after the mid-1960s. An exercise in productivity growth conducted on the UK economy using the model employed in this paper yields the following tentative result: given a service sector as efficient as the British one and the UK's level of corruption, Italy's growth in the 1970s and 1980s would have been on average one yearly percentage point higher than it actually was.
Appendix The empirical implementation of the theoretical model referred to in section 3 requires specification of the unknown restricted cost function (cv = cv(w,y,kfy The description of short-run firm behaviour is based on aflexiblefunctional form known as Generalized Leontief (GL) due to Diewert, modified where necessary to account for the existence of fixed factors. Assuming three variable inputs (labour, energy imports and other imports, denoted by the subscripts /, e and m, respectively) and two fixed factors (private capital, /c, and social overhead capital, z) we let: /
c>,y,/c,z,r) = y[a 0
i
+ Ivv/V + yyyy + yty(ty)0-5)-] 4- yO5\ I f t w ^ 5 + E i
Li
+ yykiky)05)!^05
i
+ Ivv^jzr)0-5 + y y2 (z> 5 )]
i
J
i
where a, /?, e and y are parameters to be estimated and where the scalar t is, in fact, a vector, accounting for: (1) a linear trend, (2) a cyclical indicator (number of bankruptcies), (3) the change in private capital stock, and (4) the change in social overhead capital. The latter two variables obviously take care of adjustment costs. The corresponding system of restricted factor demands is given by: xjy = dcJLwXyJVdWi = X«oK.w/-5
+
pitt05
+ ynt + yyyy + yty(ty)0-5 +yytk0-5 + ykk(k/y)
+ PiWy)0-5 + 7zl(zt/y)0-5 + yyzz0-5 + yjz/y)
+ y,2(fc0-5z°-5)/y
(A2)
while price determination is as follows:
p = [1/(1 + n0 + / / ^ K + I l a ^ w / - 5 + X/W 05 »
j
l.5yly(ty)0-5)
] i
(A3)
i
The simultaneous system of equations (A2) has been estimated with maximum likelihood estimation methods over the sample period 1894-1990.22 Table Al
Italy Table A l . Parameter