Research in Economic History, Volume 27

RESEARCH IN ECONOMIC HISTORY RESEARCH IN ECONOMIC HISTORY Series Editor: Alexander J. Field RESEARCH IN ECONOMIC HI...

Author: Alex J Field

52 downloads 1888 Views 2MB 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

DOWNLOAD PDF

RESEARCH IN ECONOMIC HISTORY

RESEARCH IN ECONOMIC HISTORY Series Editor: Alexander J. Field

RESEARCH IN ECONOMIC HISTORY VOLUME 27

RESEARCH IN ECONOMIC HISTORY EDITED BY

ALEXANDER J. FIELD Department of Economics, Santa Clara University, USA CO-EDITED BY

GREGORY CLARK Department of Economics, University of California, Davis, USA

WILLIAM A. SUNDSTROM Department of Economics, Santa Clara University, USA

United Kingdom – North America – Japan India – Malaysia – China

Emerald Group Publishing Limited Howard House, Wagon Lane, Bingley BD16 1WA, UK First edition 2010 Copyright r 2010 Emerald Group Publishing Limited Reprints and permission service Contact: [email protected] No part of this book may be reproduced, stored in a retrieval system, transmitted in any form or by any means electronic, mechanical, photocopying, recording or otherwise without either the prior written permission of the publisher or a licence permitting restricted copying issued in the UK by The Copyright Licensing Agency and in the USA by The Copyright Clearance Center. No responsibility is accepted for the accuracy of information contained in the text, illustrations or advertisements. The opinions expressed in these chapters are not necessarily those of the Editor or the publisher. British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library ISBN: 978-1-84950-770-7 ISSN: 0363-3268 (Series)

Awarded in recognition of Emerald’s production department’s adherence to quality systems and processes when preparing scholarly journals for print

CONTENTS LIST OF CONTRIBUTORS

vii

EDITOR’S INTRODUCTION

ix

EDITOR’S NOTE ON THE PARKER MANUSCRIPT

xi

ESTATE ACTS, 1600–1830: A NEW SOURCE FOR BRITISH HISTORY Dan Bogart and Gary Richardson THE MACROECONOMIC AGGREGATES FOR ENGLAND, 1209–2008 Gregory Clark

1

51

CAPITAL ACCUMULATION IN THE LONG RUN: THE CASE OF SPAIN, 1850–2000 Leandro Prados de la Escosura and Joan R. Rose´s

141

U.S. TRADE POLICY AND THE PACIFIC RIM, FROM FORDNEY–MCCUMBER TO THE TRADE EXPANSION ACT OF 1962: A POLITICAL–ECONOMIC ANALYSIS Lei (Sandy) Ye

201

A COMPARISON OF FEDERAL FINANCIAL REMEDIATION IN THE GREAT DEPRESSION AND 2008–2009 Barrie A. Wigmore

255

v

vi

CONTENTS

TRENDS IN FOOD CONSUMPTION IN THE UNITED STATES, 1840–1910 AN EXPERIMENT IN ECONOMETRICAL HISTORY William N. Parkerw

305

LIST OF CONTRIBUTORS Dan Bogart

Department of Economics, University of California, Irvine, CA, USA

Gregory Clark

Department of Economics, University of California, Davis, CA, USA

Leandro Prados de la Escosura Departamento de Historia Econo´mica and Instituto Figuerola, Universidad Carlos III de Madrid, Spain William N. Parkerw

Department of Economics, Yale University, New Haven, CT, USA

Gary Richardson

Department of Economics, University of California, Irvine, CA, USA; National Bureau of Economic Research (NBER)

Joan R. Rose´s

Departamento de Historia Econo´mica and Instituto Figuerola, Universidad Carlos III de Madrid, Spain

Barrie A. Wigmore (Retired)

Goldman, Sachs & Co., The Dakota, New York, NY, USA

Lei (Sandy) Ye

Department of Economics, Cornell University, Ithaca, NY, USA

vii

EDITOR’S INTRODUCTION Volume 27 is the last of 12 volumes of Research in Economic History I will edit. It includes six papers, evenly divided, as has been the case in the past, between European and North American topics. The lead paper, by Dan Bogart and Gary Richardson, opens up a new area of research in British economic history. Estate Acts were Parliamentary bills allowing landowners to relax restrictions on the sale, division, lease, or development of real property. These acts, which may have been as important as the more wellknown Enclosure Acts, facilitated the move of real property to more productive uses as economic conditions changed. Bogart and Richardson’s paper is followed by a Greg Clark’s massive synthesis of data on English growth and development between 1209 and 2008. Clark calculates the growth of wages, rent, and various categories of capital income, he explores how sectoral shares have varied over time, and he uses growth accounting methods to investigate trends in total factor productivity. In a second paper focusing on macroeconomic aggregates, Leandro Prados de la Escosura and Joan Rose´s provide data on the Spanish capital stock and input (service flow) over a century and half. They document the increasing importance of producer durables relative to structures, a feature Spain shared with other countries, and they show that although capital productivity fell along with long-term capital deepening, it rose during periods of accelerated growth. Lei (Sandy) Ye explores U.S. tariff policy toward the Pacific Rim. He first shows that U.S. trade policy disproportionately disadvantaged manufactured exports from Pacific Rim countries. Using Congressional roll call votes, he then investigates the political economy underlying this result, in particular the role of import-competing manufactures at the state or county level. Barrie Wigmore’s study attempts to estimate the magnitude of financial remediation – attempts to shore up or restore the financial system – in the 1930s and in 2008–2009. Admitting that the various equity infusions, loans, loan guarantees, and other subsidies were apples and oranges in terms of the dollar for dollar liability they posed for taxpayers, he nevertheless provides a unique overview of policy interventions central to understanding past and contemporary macroeconomic history. ix

x

EDITOR’S INTRODUCTION

Finally, REH is pleased to publish one of Bill Parker’s papers from 1955. For a number of reasons, Parker’s investigation of food consumption trends in the United States did not find its way into print during his lifetime. The history of how the contribution came to be published in REH, as well as the place of the paper in the literature, are discussed in an accompanying editor’s note, to which Paul Rhode also contributed. With the publication of Volume 27, I step down as editor of Research in Economic History and turn over the reins to Professors Chris Hanes and Susan Wolcott of the Department of Economics at SUNY-Binghamton. I would like to thank Associate Editors Greg Clark and Bill Sundstrom, as well as all the authors and referees who have contributed to the success of Volumes 16–27. Submissions should now be sent to Professor Hanes or Wolcott. You can contact them by e-mail at [email protected] (telephone: 607-7775487) or [email protected] (telephone: 607-777-2339). Research in Economic History will continue to welcome innovative contributions to economic history in any area. REH remains a particularly appropriate outlet for longer papers, including those with statistical appendixes, which can not easily be accommodated in such outlets as the Journal of Economic History, Explorations in Economic History, or the Economic History Review. As a review of prior volumes will illustrate, moreover, REH is receptive to papers of varying lengths covering a wide range of topics and approaches. Potential authors may submit their work in hard copy or (this is preferred) as attachments in an e-mail addressed to Professor Hanes or Wolcott. REH is flexible on matters of style and formatting during the first round of submission. If it looks as if the paper has a reasonable likelihood of acceptance, we will ask that you prepare it according to Emerald guidelines, and are happy at any stage to send you a copy of these guidelines as well as a recently prepared manuscript for use as a template. Authors may wish to consult a recent volume of REH for examples of house style. Alexander J. Field Series Editor

EDITOR’S NOTE ON THE PARKER MANUSCRIPT A manuscript copy of the paper by William Parker was given to John Komlos by Robert Gallman in 1985 while they temporarily overlapped at the University of North Carolina. Komlos was in the process of estimating food consumption trends in the United States in the antebellum period for his paper on the ‘‘Height and Weight of West Point Cadets’’ which was eventually published in the Journal of Economic History. Parker’s paper had handwritten annotations for a talk that he must have given in 1956 or 1957. It occurred to Komlos – several decades later – that it would be a pity if the study remained unpublished and Komlos contacted me about the possibility that the paper might appear in Research in Economic History. As Editor, I turned for advice to Paul Rhode, who recommended publication, and suggested that we explore the Archives of the Yale University Library to see whether Komlos’s was the only or most definitive copy. After a little detective work (Tim Guinnane contributed here) a second almost identical copy of the manuscript turned up. Komlos notes that it is a mystery why Parker never published these results that must have taken some considerable effort to weave together. Parker was a professor at the University of North Carolina and listed this affiliation as well as Resources for the Future when he wrote the paper. He was the Phillip G. Bartlett Professor Emeritus of Economics and Professor Emeritus of American Studies at Yale University when he died in April of 2000. Paul Rhode provides additional background information. This paper brings into the public sphere a classic work by William Parker that has entered into the ongoing debates over the ‘‘antebellum food puzzle’’ but has previously circulated only as an unpublished working paper. This version dates to 1955 and is from William Nelson Parker papers, Special Collections, Yale University Library.1 This paper provides estimates of per capita US food consumption in the period before the official USDA statistics. This work predates and pushes the numbers back further than the well-known analysis of M. K. Bennett and R. H. Pierce.2 Over the period he considers, Parker shows American diets tended to shift away from molasses, cornmeal, xi

xii

EDITOR’S NOTE ON THE PARKER MANUSCRIPT

and pork toward refined sugar, citrus, and poultry and dairy products. Consumption of potatoes, wheat flour, and beef held roughly steady. A key issue is the trend in pork consumption. The Census and state sources provide counts of the number of animals. But to generate meat production and consumption levels, one needs weights as well. So in this paper, Parker contributes to another controversy, one Lee Craig has deemed the oldest continuing debate in cliometrics. Namely, how much did hogs in antebellum and early postbellum America weigh? Estimating average slaughter weights quickly becomes entangled with issues such as the dressed-to-live-weight ratios, the regional differences in animal husbandry practices, and the changing seasonality of swine slaughter relative to the date of enumeration in the Census. Accounting for the hog-corn cycle adds to the complexity. I recall a story that Gallman told about Parker and this paper, one which reflects on how the places of economic historians in the history and economics professions have changed. Apparently Parker presented this work before both the Econometrics Society and the American Historical Association in very short succession. In each case, Parker emphasized the material opposite from the audience’s predilections – pushing the narrative historical side to the econometrics crowd and the statistics to the historians. As partial support for this recollection, I note that Parker did present a paper entitled ‘‘Trends in Food Consumption in the United States, 1870–1910: An Experiment in Econometrical History’’ at the Atlantic City meetings of the Econometrics Society in September 1957. According to the abstract: ‘‘The most striking results of the measurements are the indications of a declining per capita consumption of meat (resulting from a fall in pork consumption) and of flour (from a fall in cornmeal consumption). The data and the assumptions used to calculate the trend of pork consumption are fully set forth to illustrate the methods used.’’3 What was old is now new and noteworthy.

NOTES 1. For the paper’s place in the literature, note that Robert E. Gallman, ‘‘Dietary Change in Antebellum America,’’ Journal of Economic History, Vol. 56, No. 1 (March, 1996), pp. 193–201, cites it in response to John Komlos, ‘‘Heights and Weights of West Point Cadets: Dietary Change in Antebellum America,’’ Journal of Economic History. Vol. 47, No. 4 (1987), pp. 897–927. The Yale version has the same text as that Gallman possessed and circulated, but lacks the presentation notes. In the Parker collection finding aid, this item (no. 12) does not bear the warning accompanying item no. 42: ‘‘Preliminary Draft-not to be quoted or used in any way; to be read skimmingly, it [sic] at all.’’

Editor’s Note on the Parker Manuscript

xiii

2. M. K. Bennett and R. H. Pierce, ‘‘Change in the American National Diet, 1879–1959,’’ Food Research Institute Studies 2 (May 1961), pp. 95–119. 3. Econometrica, Vol. 26, No. 2 (April, 1958), p. 320. This account is not definitive because Parker presented another paper, ‘‘A Statistical Framework for American Agricultural History, 1840–1910,’’ at the St. Louis meeting of the Econometrics Society in December 1960 and I (Paul Rhode) have found no indication of presentations before the AHA.

Alexander J. Field Editor with contributions from John Komlos and Paul Rhode

ESTATE ACTS, 1600–1830: A NEW SOURCE FOR BRITISH HISTORY Dan Bogart and Gary Richardson ABSTRACT A new database demonstrates that between 1600 and 1830, Parliament passed thousands of acts restructuring rights to real and equitable estates. These estate acts enabled individuals and families to sell, mortgage, lease, exchange, and improve land previously bound by landholding and inheritance laws. This essay provides a factual foundation for research on this important topic: the law and economics of property rights during the period preceding the Industrial Revolution. Tables present time-series, cross-sectional, and panel data that should serve as a foundation for empirical analysis. Preliminary analysis indicates ways in which this new evidence may shape our understanding of British economic and social history. The data demonstrate that Parliament facilitated the reallocation of resources to new and more productive uses by adapting property rights to modern economic conditions. Reallocation surged in the decades following the Glorious Revolution and was concentrated in areas undergoing urbanization and industrialization. The process was open to landowners of all classes, not just the privileged groups who sat in the Houses of Lords and Commons. Parliament’s rhetoric about improving the realm appears to have been consistent with its actions concerning rights to land and resources.

Research in Economic History, Volume 27, 1–50 Copyright r 2010 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 0363-3268/doi:10.1108/S0363-3268(2010)0000027006

1

2

DAN BOGART AND GARY RICHARDSON

1. INTRODUCTION Between 1600 and 1830, Britain’s Parliament passed more than 3,500 acts – known as estate acts – which freed land from the shackles of a landholding and inheritance system – known as the system of strict settlements. Strict settlements imposed equitable estates in land (enforced by the court of Chancery) on top of real estates in land (enforced by common law courts). These overlapping estates involved numerous individuals, including the landholder, his extended family, additional beneficiaries designated by past landholders, and potential heirs (including those unborn). All of these individuals possessed rights to revenues derived from the land. These complicated bundles of overlapping rights prevented property holders from using resources as they saw fit. Landholders could neither mortgage, nor lease, nor sell much of the land under their control. Estate acts freed property from this rigid system. They permitted previously prohibited actions, reorganized complicated bundles of rights, and conveyed those new rights to new users. Estate acts began as petitions from landholders seeking relief from restrictions that strict settlements imposed on the employment of land and resources. Parliament reviewed these petitions, ensured that they met certain standards for protecting the rights of all interested parties, and then passed legislation establishing new rules regarding the employment and conveyance of property. The volume of estate acts exceeded the volume of all other legislation, with the exception of acts establishing statutory authorities and enclosure commissions. Scholars expend considerable energy analyzing the latter legislation, but few scholars have studied estate acts. As a result, basic questions remain unanswered. When were the majority of acts affecting rights to land and resources passed? In which geographic regions were they concentrated? Who procured the acts? Why did they do so? What were the consequences for the use of land and resources? In this essay, we answer these factual queries using a new dataset of all estate acts passed between 1600 and 1830. We show that estate acts authorized a variety of transactions such as sales, leases, and mortgages by changing legal rights to land and resources. Acts authorizing sales and leases were minimal in the 1600s before increasing in the late seventeenth and early eighteenth centuries. The number of estate acts peaked in the early nineteenth century and were concentrated in counties like Middlesex and Lancashire where urbanization was proceeding rapidly. They were procured by landed families of all ranks – nobles, the gentry, and commoners – throughout the 1700s.

Estate Acts, 1600–1830

3

Answering these factual queries is a necessary first step toward answering deeper questions concerning the causes and consequences of Britain’s economic ascent. Our essay explores several sets of these bigger questions. One set is whose interests did Parliament promote during the seventeenth, eighteenth, and nineteenth centuries? Who had access to Parliamentary legislation? The data introduced in this essay indicate that all landowners, even those in the middling ranks of the social and political hierarchy, had access to the legislature. The proportion of estate acts granted to different social classes mirrored those classes share of total landownership. A second set concerns the Marxist narrative about the decline of the English aristocracy and growing class conflict in early-modern Britain. Did aristocratic families struggle to maintain their social position relative to a rising entrepreneurial and mercantile class, as Marxist historians contend? Did aristocratic families sell valuable properties to pay for debts incurred maintaining expensive lifestyles? Were estate acts a mechanism for facilitating these sales? Does the documentary record prove that landed families incurred large debts in an ultimately futile struggle to outspend their class rivals? While it is true that a large number of estate acts mention ‘‘to pay debts’’ as a rationale for the sale of property, which the Marxists’ advance as the key piece of evidence supporting their hypothesis, only a handful of acts indicate that the debts arose due to extravagant expenditures or profligate lifestyles. Most acts mention debts in the context of investing in the lives (educations, careers, dowries) of children or in improvements to land. Over time, the use of estate acts to pay debts declines (to nearly zero), and the use of acts to facilitate investment rises. This pattern is the opposite of that predicted by Marxist scholars. A third (and probably the biggest) set of big questions concerns Parliament’s actions and England’s economic development. Did Parliament’s actions match its rhetoric about acting in the public’s interest and to increase the realm’s wealth? Did Parliament’s actions foster economic growth? A great deal of evidence indicates that estate acts served constructive purposes. Acts authorizing long-term leases, for example, typically described the projects, such as the opening of mines or construction of residences, that the leases facilitated. Acts authorizing the sale of property (or otherwise releasing property from the strictures of settlement) typically specified that a portion of the proceeds of the sale had to be dedicated to purchasing lands and settling them to the old usage (or taking other actions that would ensure all beneficiaries of the estate remained as well off financially as they had been in the past). The texts of the acts, in other words, reveal Parliament’s intentions. Parliament approved

4


reallocating resources to new and more productive uses, as long as the financial interests of beneficiaries to estates were protected. A fourth and equally important question concerns the effects of political changes. Is there any link between the Glorious Revolution of 1688–1689 and property rights in Britain? If any link exists, did property rights become more secure, more flexible, or both? The data reveal a substantial increase in the number of estate acts authorizing property sales and leases shortly after the Glorious Revolution. In other words, Parliament increased the rate at which it changed rights to property after 1689. Moreover, the data show that the types of property transactions authorized by Parliament were relatively similar before and after 1689. These findings suggest the Glorious Revolution increased the adaptability of property rights in Britain and encouraged dynamic efficiency by relieving constraints on investment. A companion paper (Bogart & Richardson, 2009) outlines our transaction cost interpretation of estate legislation. By reducing transactions costs and increasing the ease and security of conveyance, estate acts facilitated the reallocation of resources to new and more productive uses. Case studies and quantitative evidence indicate that this was the case. Concentrations of estate acts occurred in urbanizing areas, such as the periphery of London, and in industrializing regions, such as the county of Lancaster. Correlations between estate legislation, urbanization, and industrialization suggest a link between the reorganization of rights and Britain’s march toward modernity. Another companion paper advances a broader argument (Bogart & Richardson, 2010). At the end of the seventeenth century, Parliament established forums where rights to land and resources could be reorganized. These forums issued estate, enclosure, and statutory authority acts, which enabled individuals, families, and communities to alter property rights and reallocate resources in response to changing economic conditions. The ensuing institutional adaptability, we argue, fostered Britain’s expansion. We test a key implication of our argument by demonstrating that between 1700 and 1830, when the public’s demand for acts reorganizing property rights increased, Parliament responded by passing more acts. Property rights, in sum, were adaptable. This essay establishes a factual foundation for our argument. It also disseminates data on estate acts, including statistical series characterizing the acts’ economic, legal, social, geographic, and temporal characteristics.1 The series illuminate the connection between property rights and economic development at the microeconomic level. The academic literature is divided on the degree to which property rights evolved during the seventeenth,


5

eighteenth, and nineteenth centuries and whether the evolution of property rights contributed to agricultural, industrial, and financial developments in Britain.2 By publishing these series, we hope to stimulate research into these issues.3 Section 2 discusses information necessary to understand estate acts’ context and content. Section 3 describes the sources of evidence and our procedures for processing the data. Sections 4 and 5 describe estate acts’ legal and economic dimensions, respectively. Section 6 presents information on trends over time. Section 7 provides information on geographic patterns. Section 8 discusses the social ranks and professions of individuals mentioned in estate acts. Section 9 discusses the implications of this evidence. Estate acts shed light on many issues in British economic, social, legal, and political history. The acts also reveal much about the lives of the individuals and communities that expended great efforts passing so much legislation reorganizing property rights at a tipping point in economic history.

2. HISTORICAL BACKGROUND Understanding the rationale for and the impact of estate acts requires historical knowledge of three types. The first concerns the system of landholding and inheritance that generated the need for estate acts. The second concerns the evolution of politics and Parliamentary structure from the seventeenth to the nineteenth centuries. The third concerns general trends in legislation to which estate acts can be compared.

2.1. System of Landholding and Inheritance Estate acts arose from an English system of inheritance that solidified around the Civil War of the 1640s and prevailed for several centuries thereafter.4 During this era, large landowners held most of their land under settlement. Lesser gentry and yeoman families also employed the legal device, even on single-family farms. While estimates vary, at the peak, at least one quarter and as much as three-fourths of land in England was strictly settled.5 A settlement was designed to care for extended family and to keep a family’s estate together for future generations. The current holder of settled land was a life tenant and was not the absolute owner. The land belonged to

6


a trust, for which the current holder was named the beneficiary. The holder, in turn, held the land in trust for other beneficiaries, typically including his wife, children, unborn descendants, all potential future heirs, and members of his extended family, such as his brothers, sisters, nieces, nephews, and other descendants of previous holders of the estate. The life tenant controlled the use of the land possessed by the nested trusts as long as they fulfilled the terms of their stewardship. Once the life tenant died, the settlement dictated that the estate descend intact from one generation to the next. It did this by assigning control of the estate to a single heir, usually the eldest son of the current holder. Settlements had three features which created a need for parliamentary involvement. First, settlements restricted the uses to which resources could be put. The holder of a settled estate (who was just a life tenant) could not grant leases lasting beyond his life and could not grant leases from which he benefited at the expense of his successors (such as leases in which tenants paid lump sums up front in return for concessions). The holder of a settled estate could not sell, exchange, or mortgage the property. If he completed such transactions, he could be held liable for damages to the estate, and the transaction could be voided, because he had no power to transfer title. Similarly, the holder of an estate could not alter the property, even if he considered the alterations to be an improvement. The removal of trees, hedges, and buildings, the opening of new mines, quarries, and peat bogs, and the conversion of arable lands into pasture (or vice versa) could be considered waste. All those who benefited from such actions could be liable for damages, if upon inheriting an estate, the successor claimed to have been harmed by the acts. Sales, exchanges, mortgages, improvements, and longterm leases could only be undertaken if the powers section of a settlement contained specific clauses authorizing such actions. Second, a settlement legally bound the hands of all heirs alive when it was written. A settlement could not be changed until a tenant in tail (i.e., the next in line to inherit) who was born after the date of settlement came of age (i.e., reached age 21). Then, the current life tenant (usually the father) and the future tenant in tail (usually the son) could remove the entail by the legal process of common recovery. These facts meant that a settlement could be changed only infrequently, at intervals of 21 years or longer, as a family waited for an heir to come of age and for the father and son to reach an agreement about restructuring the estate. Third, conducting transactions and enforcing contracts on settled land could be costly, uncertain, and insecure. Settlements were long, complex documents, traditionally unpunctuated, and full of repetition.6 Interpreting


7

settlements required experience, skill, detailed knowledge of the document, and a large library of property laws, precedents, and legal texts estimated at 674 volumes in 1826 (English & Saville, 1983, p. 18). Settlements were not part of the public record. Copies of the deeds were usually held by the settlers, trustees, and lawyers. Settlements had to be consulted before taking out mortgages, drawing up leases, or completing sales, because if the settlement did not specifically authorize a transaction, the transaction could be voided. Ambiguities in settlements often deterred individuals from acting on estates; for fear that the transactions would be disputed by successors. Settlements placed restrictions on the deployment and use of property but they also prevented the holder of the estate from dissipating resources dedicated to the support of future heirs and the extended family. Families did not know how the Chancery court would react to the inclusion of additional powers or whether the wording of a novel clause might provide a life tenant with a loophole enabling him to circumvent all other restrictions. They did not know the personality of the person(s) who would inherit the estates and who would have power over the widow, dependants, and descendants of the individual who established the settlement. Uncertainty about the impact of providing powers to the life tenants and the threat that powers might pose to the interests of the extended family meant that families often favored narrow rather than extensive and/or novel powers in their settlements. Parliament provided a way of overcoming these restrictions through the passage of estate acts. To understand how Parliament emerged in this role we must first review some parliamentary history.

2.2. Political and Parliamentary History Estate legislation formed a key component of the surge in legislative activity that began in the 1690s following some of the most dramatic events in British political history: the Restoration of 1661 and the Glorious Revolution of 1688–1689. In the early 1600s, the Stuart monarchs, King James I and King Charles I, attempted to overturn English political tradition, and impose an absolute monarchy ruling by divine right, rather than a kingship ruling with the consent of the nobility and clergy, as enshrined in the Magna Carta. James I and his son argued with Parliament over taxation, religion, foreign policy, the prerogatives of the monarch, and the limits of royal power. To weaken their Parliamentary opponents, the Stuart monarchs seldom called Parliament into session.

8


The Stuarts’ attempt to impose personal rule ended when Parliamentary forces defeated the royal army during the Civil War. After the conflict, King Charles I lost his head, and a republican government temporarily reigned. Cromwell’s Protectorate lasted for two decades. In 1661, however, the Stuarts returned to power, and Charles II ascended the throne. During the decades that followed, political instability reemerged. The Stuart dynasty ended with the Glorious Revolution of 1688, when William of Orange and his wife Mary invaded England. The last Stuart monarch, James II fled. Parliament accepted William and Mary as the new king and queen. As part of his Parliamentary alliance, William accepted the Declaration of Rights, which gave Parliament the power to convene each year and determine the length of its sessions, and required the monarch to consult Parliament on legislative issues. William and Mary’s reign lasted from 1689 to 1701. Queen Anne’s reign followed from 1702 to 1714. This 25-year period experienced transcendent political developments. Parliament solidified its role as the principal lawmaking body. The House of Commons took control of the authorization of taxes. Central government taxation expanded as Britain waged an expensive war. The Bank of England was formed in 1694 and helped to propel the growth of government debt. The Whig and Tory parties battled for control over the House of Commons and the spoils associated with ministerial posts. In 1714, the monarchy was transferred to George I of Hanover. Over the next 100 years there were numerous shocks and challenges but the political system was not fundamentally altered. Hanoverian monarchs reigned for over 100 years through a succession of kingships (George I, George II, George III, George IV, and William IV). Following the Septennial Act in 1717, Parliament had an election every 7 years. The right to vote remained in the hands of elites. The Whig and Tory parties maintained a presence in Parliament and came to exercise greater control over the ministry. Excise taxes and government debt continued to expand fuelling Britain’s involvement in foreign and colonial wars. The pattern of political fluctuation and stability mirrored the process of procedural innovation and constancy. In the late seventeenth and early eighteenth centuries, legislative procedures evolved rapidly, as legislators strove to standardize and streamline the process for passing legislation. Effective procedures emerged by 1715 and changed relatively little thereafter. Private bills began with a petition from individuals desiring to alter rights. Petitioners hired lawyers to prepare paperwork. Parliamentary committees investigated merits of petitions and issued reports. Petitions

9


were written into bills, read to the public three times, passed through both houses of Parliament, and then sent to the king for royal assent. Public notice ensured those with interests in bills knew about the proposal. Parties could oppose bills by submitting counterproposals to Parliament. Parliamentary committees considered the contending proposals, and then passed one of the bills, modified the original bill to satisfy the opposition, or rejected both proposals. The multiple layers for review and numerous opportunities for opposition ensured that Parliament considered the interests of all concerned before coming to a decision. The acts themselves also became standardized in the late seventeenth and early eighteenth century. Parliament used templates whenever individuals requested legislation enabling them to divorce, naturalize, change their name, or alter their estate. Consistency of form and function enabled Parliament to rapidly process petitions.

2.3. Trends in Legislation Estate acts emerged at a time when Parliament passed increasing quantities of legislation of many types. Fig. 1 illustrates the rising tide of legislation and indicates the dates of political events mentioned in the previous section. Fig. 2 decomposes the total number of acts passed into three categories authorizing changes in rights to specific pieces of property – estate, enclosure, and statutory authority acts – and all other acts. The figures illuminate important patterns. During the eighteenth century, the quantity of legislation surged. In the early nineteenth century, the number of acts passed per year peaked near 350. Between 1690 and 1720, estate acts accounted for much of the growth. Between 1750 and 1830, the number of estate acts remained relatively stable, while the number of statutory authority and enclosure acts soared. Estate, enclosure, and statutory authority legislation shared key features. All altered property rights. Statutory authority acts created new organizations with the authority to provide infrastructure and public services. These organizations contributed to the improvement of roads, rivers, harbors, canals, railways, water supply, waste disposal, and county buildings and the provision of police, poor relief, and disputes resolution services. Enclosure acts eliminated the open fields by ending the collective ownership and management of land and for the old system substituting private ownership and management of resources.

10


400 Annual Number of Acts, 10-Year Moving Average Acts Per Session

Number of Acts

300 STUART RESTORATION DYNASTY GLORIOUS BEGINS REVOLUTION CIVIL WAR BEGINS

200

100

0 1500

1550

1600

1650

1700 Year

1750

1800

1850

1900

Fig. 1. Political Events and Acts Passed by Parliament, 1500–1900. Source: See text. Note: Denote as At the number of acts passed in year t. The series ‘‘acts per is an 11-year moving session’’ equals if AtW0. The series ‘‘annual number of acts’’P average. The formula for the moving average is A t ¼ ð 5i¼5 Atþi Þ=11.

3. DATA DESCRIPTION AND CODING METHODS As acts passed through Parliament, clerks wrote their contents on long pieces of parchment. Clerks stored the documents by rolling them tightly and writing a summary, called a clerical title, on the exterior. These clerical titles summarized the act, usually in a concise paragraph containing enough information for the clerks to identify the act and its principal provisions amidst thousands of similar pieces of parchment, without opening the rolls to read the full text. These clerical summaries form the foundation of our database. There is a large and venerable body of scholarship which uses the clerical titles to categorize and count acts of Parliament. The most recent examples include Hoppit (1996, 1997, 2003), Innes (1998), Tate (1967, 1978), Turner (1980, 1984), and Wordie (1983). All of these studies relied on printed sources, like the Statutes of the Realm, or the conversion of printed sources into an electronic form. This paper uses an existing electronic database of

11

Estate Acts, 1600–1830 2500

Number of Acts

2000

1500

1000

500

16 01 16 11 16 21 16 31 16 41 16 51 16 61 16 71 16 81 16 91 17 01 17 11 17 21 17 31 17 41 17 51 17 61 17 71 17 81 17 91 18 01 18 11 18 21

0

estate

enclosure

statutory authority

other

Year

Fig. 2. Categories of Legislation Pertaining to Property Rights, 1601–1830. Source: See text. Note: The plotted series are 11-year moving averages of the annual data as defined in the note for Fig. 1.

clerical titles instead of the printed sources. The Parliamentary Archives maintains a catalog, Portcullis, which indicates the clerical title, calendar year, regal year, and parliamentary session for all acts passed since the early sixteenth century.7 For acts passed before 1798, Portcullis indicates whether the act was public or private. The distinction referred to the type of documentation required in legal cases. Parties at law had to present to the court authorized copies of private acts in order to have the provisions of the act applied in the case. Public acts were considered to be part of the public record and applied in all cases without the need to submit them as evidence. For acts passed after 1798, Portcullis indicates whether the act was public or local/personal. This distinction pertained to the scope of the legislation. A public act created a law of general application throughout the jurisdiction in which it was proposed. A personal or local act affected only a single person, group, or locale, which was named within the act. We use the clerical titles in Portcullis to identify all estate acts that were passed between 1600 and 1830. We define estate acts as private (and personal) acts which affected individuals and families rights to equitable estates and to real property held within equitable estates. We exclude

12


marriage acts, naturalization acts, name acts, and office acts, which were private acts that affected individuals’ and/or families’ rights but did not affect the structure of equitable estates. We also exclude restitution and restoration acts, which restored individuals’ and families’ rights to property that had been taken from them due to convictions for crimes, typically treason following civil unrest. Identifying estate acts among all private acts is straightforward. Parliament standardized procedures for writing acts. Clerks possessed templates from which they wrote the initial acts and standard procedures for converting final versions into summaries written on the exteriors of rolls. For example, the title for all acts that settled the property of an estate onto an individual began, ‘‘An act for settling the estate y.’’8 The title for all acts that authorized the sale of property from an estate usually began ‘‘An act for the sale of y,’’ ‘‘An act for effecting the sale of y,’’ or ‘‘An act for selling y.’’9 After identifying estate acts, we classify several categories of information contained within the clerical titles. This information includes legal actions; economic transactions; the property rights changed by the former to facilitate the latter; the geographic locations of land involved in economic transactions; and the social rank, profession, and gender of individuals whose estates were altered by estate acts. We created our categorizations by reading all of the clerical titles and then devising algorithms to extract the relevant information. The authors applied these algorithms independently and then compared their results. We also encapsulated these algorithms in computer coding, which scanned the clerical titles for key words, and compared our results to the computer output. Most of the key categories were also checked by one of our research assistants. The consistency of estate acts’ form and function makes extracting the relevant information straightforward. An example should clarify our procedures. Consider an estate act from the mid-eighteenth century. We begin with two pieces of information: (1) A reference number from the Parliamentary archives: ‘‘HL/PO/PB/1/ 1741/15G2n48.’’ (2) A descriptive clerical title: ‘‘An act to empower Henry Earl of Carlisle to make leases of Coal Mines and Coal Works, lying within his settled estates in the county of Northumberland, for any term not exceeding 99 years.’’ The reference number reveals that the document was a private act (PB), originally written by the clerks in the House of Lords (HL), passed by


13

Parliament during the fifteenth year of the reign of George II in 1741. The clerical title determines that someone (Henry Earl of Carlisle) was legally empowered to make leases of a property held in an estate (Coal Mines and Coal Works) in the county of Northumberland. We enter all of this information into a spreadsheet. Each of the spreadsheet’s rows pertains to a single act of parliament. Each column contains the same type of information for each act. Our detailed encoding enables us to construct statistical series that reveal trends of interest to historians and social scientists. These statistical series usually report the annual number of acts that did something – such as authorize the sale of land – over time. Such time series could, of course, conceal as much as they reveal, particularly if the scope, scale, and nature of the legislation changed over time. For this reason, we visited the Parliamentary Archives and examined samples of original acts from the beginning (1610s), middle (1700s, 1740s, and 1780s), and end (1800s) of our sample period. Our examination reveals that estate acts possessed salutary statistical properties. Estate acts were standardized in form and content. Estate acts described the property that was affected, the individuals involved, and the rights that were changed. There was variation in the geographic extent of properties affected. Later we show that the proportion of acts which name properties in multiple counties as opposed to a single county was stable over time. This property ensures that counting the number of acts passed annually reveals broad trends in the amount of property being affected by the acts and the types of rights being created, altered, or annulled because the distribution of the area affected was constant. For the vast majority of acts we are able to determine correctly the year in which it was passed. For most of our sample period, a convention dated all acts passed by a session of Parliament as if they were passed on the opening day of the session. This convention lingered from an earlier period when Parliament met infrequently at royal request and handled a limited volume of business in a short time period. In the eighteenth century, Parliament met annually. Sessions began in the fall, usually in the months of October, November, or December; lasted throughout the winter; and adjourned in the spring. Complications arose, however, in 1699, 1701, 1715, 1752, and 1761 when the monarch died, and/or Parliament opened late. In 1714, for example, Queen Anne died. George I assumed the throne. His ascension delayed the opening of Parliament until January of 1715. This parliament adjourned in the spring and another opened on schedule during the next fall. So, in the year 1715, the conventional dating method assigned the acts

14


passed in two Parliamentary sessions – the winters 1714–1715 and 1715– 1716 – to one calendar year, 1714, yielding a count of zero acts in 1715. The years 1699, 1701, 1715, 1752, and 1761 are also recorded as having zero acts, while their adjacent years have an especially high count of acts. In other works, we propose methods for dealing with this issue in regression analysis (Bogart & Richardson, 2010).

4. LEGAL FUNCTIONS OF ESTATE ACTS Estate acts changed rights to equitable and real estates or clarified the rights and responsibilities created by strict settlements. These modifications enabled land holders and trustees to undertake transactions that they could not undertake given existing arrangements. It is useful to summarize the most common legal changes associated with estate acts. The typical vesting act contained five elements. It granted (i) some property, right, or benefit, which had been a portion of (ii) someone’s settled estate, (iii) to someone else, (iv) for some reason, and in some circumstances (v) in exchange for some property or asset. Some examples illustrate the function of vesting acts. An act from 1702 vested ‘‘certain lands and tenements of Montague Earl of Abingdon, in trustees, to be sold’’ and the proceeds employed in the purchase ‘‘of other lands of equal value’’ to be employed ‘‘to the same uses, as the lands to be sold are limited.’’10 An act in 1759 vested ‘‘the inheritance of certain estates in the County of Northampton, part of the entailed estate of John Freeman Esquire, in him, in fee simple’’ in return for ‘‘settling other estates in the Counties of Wilts and Middlesex, in lieu thereof.’’11 Enabling and empowering acts were similar to vesting acts. They typically enabled or empowered (i) someone, (ii) to do something, (iii) for some reason. For example, an act in 1692 enabled ‘‘Abel Atwood to sell some Lands for payment of debts and to make provision for younger children.’’12 An act in 1725 enabled ‘‘Charles Lowndes Gentleman and the persons in remainder after him to make contracts for getting brick earth in, and grant building leases of the house and ground called Spring Garden.’’13 An act from 1695 empowered ‘‘the Most Noble Anne Duchess of Buckcleugh, and the Right Honourable James Earl of Dalkeith, her son, of the Kingdom of Scotland, to grant leases for improving a piece of ground in the Parish of St. Martin in the Fields.’’14 An act from 1749 empowered ‘‘trustees to cut down and sell timber upon the estate late of John Trevor Esquire, in the Counties of Denbigh and Flint, for discharging his debts, and also to make leases of mines in the said counties.’’15 Acts that enabled, empowered, or


15

vested rights in trustees or life tenants shared a common characteristic: they provided legal protections for individuals. Legal reformers in the early nineteenth century often noted ambiguities in trustees’ powers and the complications that could arise when disputes arose.16 One judge might find that a trustee was liable for taking an action, while another judge may not. Estate acts marked an early attempt to address this problem. A number of estate acts change rights to undertake transactions. Acts authorizing sales affected rights over a specific property. They typically authorized the sale (i) of something, (ii) by someone, (iii) for some reason, and (iv) if certain conditions were met. An example from 1692 authorized the ‘‘sale of lands by Sir Robert Smith’’ under the condition that he ‘‘settle other lands of greater value to the same uses, in lieu thereof.’’17 An act in 1773 authorized the sale of ‘‘certain charity estates’’ and apply the proceeds to ‘‘the building of a town hall and shambles in the Town of Newark upon Trent and in the purchasing of lands and hereditaments for enlarging the Church.’’18 An act in 1702 authorized ‘‘Trustees to make Sale of Part of the Estate of Humphrey Bury, for paying of a Mortgage, and a Portion charged thereupon.’’19 Acts authorizing exchanges were similar but they authorized the transfer of two or more properties. The typical act authorized the exchange (i) of some property, (ii) possessed by someone, (iii) for some other property, (iv) possessed by someone else, (v) for some reason, and (vi) if certain conditions were met. An example from 1692 permitted the exchange ‘‘of several small Parcels of Land in the Parish and Manor of Fulham, belonging to the Bishoprick of Londonyfor other Lands of the like Value, to Charles Earl of Monmouth, and his Heirs.’’ An act in 1739 permitted the exchange of land ‘‘belonging to Thomas Inaven Esquire, in the Parish of Wootton in the County of Bedford, for other lands of equal value in the same Parish, belonging to the master, fellows and scholars of Sidney Sussex College.’’20 An act in 1785 authorized the exchange of ‘‘part of the settled estate of Heneage, Earl of Aylesford, in the County of Kent, for another Estate, of greater Value, in the same county, to be settled in lieu thereof.’’21 Acts for discharging were also specific to a property. They typically discharged (i) something, (ii) from some restriction, (iii) for some reason, and (iv) substituted something else in its place. An example from 1677 discharged the Manor of Winstead in the County of York from a Settlement in Tail, and charging other Manors and Lands in the County of Lincoln of a greater Value with the same Uses.22 An act in 1733 discharged a ‘‘certain Piece of Ground called The Pesthouse Field from certain charitable Trusts,

16


and for settling another Piece of Ground of equal Extent, and in a more convenient Place, upon the same Trusts.’’23 Acts for settling had a structure similar to acts for discharging because they placed property into the confines of a settlement. An example from 1697 settled ‘‘certain lands in Essex on Thomas Burgh Esquire, and his heirs, in lieu of other lands of greater value, conveyed by him, according to the decree and the will of Sir Samuel Jones deceased.’’24 An act from 1735 settled ‘‘the estate of William late Earl Cowper deceased, to the uses and for the purposes mentioned in certain articles of agreement, made between William now Earl Cowper, and his brother, and the issue of Spencer Cowper Esquire deceased.’’25 Some estate acts confirmed a transaction that had already taken place. For example, an act from 1692 confirmed ‘‘the sale of certain wood lands in the County of Southampton, and certain articles of agreement made between Isaac Woollaston and Richard Woollaston Esquires.’’26 An act from 1738 confirmed ‘‘an exchange, agreed to be made between Thomas Holles Duke of Newcastle, and Sir Miles Stapleton Baronet, of their settled estates in the County of York; and for settling the lands given in exchange to each party’’ to the uses of their estates.27 An act from 1790 confirmed a lease ‘‘lately made by Henry Nevill, Earl of Abergavenny, of certain entailed mines, and other hereditaments, in the County of Monmouth, and to enable granting future leases of the said entailed mines and other Hereditaments.’’28 Acts for settling, discharging, exchanging, selling, and confirming shared a common characteristic: they authorized or confirmed a specific transaction. It is worth repeating that many of these transactions could be challenged in court if they were not authorized in a settlement. Estate acts provided immunity from such legal challenges. A last type of estate act remedied mistakes in acts previously passed by Parliament or inadvertently included in families settlements. For example, an act in 1719 was for ‘‘supplying the Defects in, and better Performance of the Will of Edmund Dunch Esquire, deceased.’’29

5. THE ECONOMIC FUNCTION OF ESTATE ACTS The examples above make it clear that landholders used the legal forms described in the preceding section to achieve economic ends, which we describe in this section. The legal forms and economic ends do not map into each other in a one-to-one relationship. Some legal forms could be used for multiple purposes. Some economic ends could be achieved through multiple

17


Table 1. Transactions Authorized by Estate Acts, 1600–1830. Type of Transaction

Sale Lease Exchange Discharge Mortgage Partition Harvest timber Mine ore/coal

Number of Acts (1)

Percent of Acts (2)

1,814 538 273 193 137 93 60 44

51.5 15.3 7.8 5.5 3.9 2.6 1.7 1.4

Source: See text.

means. The sale of property, for example, could be accomplished via an act that directly authorized the sale of a particular piece of property or that empowered a landowner to sell property or that vested in someone the authority to decide what property should be sold. Table 1 summarizes the economic transactions most often authorized by estate acts. Column (1) indicates the total number of estate acts that authorized sales of land, long-term leasing of land, exchanges of property, discharges of property from the restrictions of strict settlements, mortgaging of property, partitioning of property, harvesting of timber, and mining of ores, metal, and coal. Acts authorizing sales and leases allowed particular pieces of property (and at times entire estates) to be put on the market. Exchanges of land involved removing one (or more) specific pieces of property from a settled estate and replacing it with specific pieces of property from another estate. Discharges involved removing restrictions from land without necessarily specifying what would happen to the land and if/whether/ what would replace it. Mortgage acts allowed property within an equitable estate to serve as collateral for a loan. Partitions separated what had been considered a single tract of land into two (or more) units that could be put to different uses or sold separately. All of these transactions enabled landholders to undertake actions or engage in transactions that they could not give the existing allocation of rights and laws concerning equitable and real estates. Column (2) indicates the percentage of estate acts that authorized these transactions. The percentages sum to more than 100 percent because estate acts could authorize more than one type of transaction. The most frequently authorized transaction was the sale of property. Acts authorizing sales usually contained conditions. In many cases, the life tenant

18


had to use a portion of the proceeds to purchase property with a value equivalent to the land just sold and incorporate that property into the estate. This provision protected the income of beneficiaries, such as current dependents and future heirs. In some cases, trustees used the proceeds of the sale to pay debts, often accumulated by the recently deceased life tenant, or to provide pensions, portions, or jointures for the beneficiaries. Lease acts enabled landowners to lease land for terms extending beyond the life of the current landholder and for a fixed number of years, often up to 99. Such long-term leases ensured tenants that they would not have to renegotiate the lease (or be kicked off the property) when an heir inherited the estate.30 Without such assurances, tenants would be reluctant to invest, improve, or maintain property, because the sums that they sunk into the project could be expropriated when their landlord died or when the life tenant and heir apparent renegotiated the settlement and transferred possession of the property to the next generation. This disincentive – known as the hold-up problem – is well documented in transaction cost economics (Williamson, 1985). Acts often authorized a particular type of contract known as a building lease. An act in 1788, for example, enabled Charles Earl Camden to grant ‘‘building leases of the prebendal lands at Kentish Town, in the County of Middlesex.’’31 Building leases authorized the letting of land on long-term contracts, typically 99 years, on which the leaseholders could construct buildings for residential or industrial use and in turn, lease the premises to subtenants who would occupy the buildings. These leases facilitated construction on the edges of rapidly growing towns and cities, where property owners wanted to switch land from rural to urban uses, but lacked contractual forms needed to do so. Table 2 shows that 5.9 percent of all estate acts specifically authorized building leases. A small number of acts authorized leases on land where metal ores and coal lay underneath. In 1736, for example, an act enabled

Table 2.

Types of Leases Authorized by Estate Acts, 1600–1830.

Building lease Housing lease Mining lease Other or no type stated Source: See text.

Number of Acts

Percent of Estate Acts

209 55 33 262

5.9 1.6 0.9 7.4


19

‘‘the Guardians of Anthoney Langley Swymmer, an Infant, to join in making Leases of certain Mines, in the County of Flint, with the other Owners thereof, during the Minority of the said Infant’’32 In 1783, an act empowered ‘‘Nigel Bowyer Gresley y to lease Part of his settled Estates in Staffordshire, pursuant to an Agreement entered into with George Parker, and others, Iron Masters; and also to grant Leases of Lands and Mines within the same Estates.’’33 This lease authorized the opening of a mine in the heart of the midlands metal-mining area during the early stages of the Industrial Revolution. Mining leases typically lasted from 40 to 60 years. This long-time horizon provided lease holders with substantial incentives for making location-specific investments. Acts authorizing the exchange of land had many of the same effects as sale acts with conditions to purchase and settle property of equivalent value. The difference between these acts stemmed from the details of the transaction. Exchanges specified that the land that would be placed in the settlement to replace the land withdrawn from the equitable estate. Many sales indicated that land (or another asset) should replace the property withdrawn from the equitable estate, but specified only the value and use of what the remuneration should be. Acts discharging one property from a settlement and settling another property of equal value were also similar in their effects. If a family wanted to free a property from the confines of a settlement they might want to exchange it with some of their other properties that were not restricted. In this case they were exchanging two or more properties within their own estate rather than exchanging with a different party. Acts authorizing the partitioning of land were particularly useful when shifting agricultural land to urban uses. Residences required less space than farms. Property owners partitioned farms so they could divide the property into multiple building sites. An example from 1809 was an act to ‘‘Partition certain Settled Estates of John Wharton Esquire, situated in the Counties of York, Westmorland, and Durham.’’34 An act from 1827 confirmed a ‘‘Partition made by Mary Bainbrigge Spinster, with the Reverend Richard Fawcett Clerk and Anna Maria his Wife, and others, of an Estate situate in the Township of Headingley-cum-Burley in the Parish of Leeds in the County of York.’’35 The latter act occurred at a time when the rapidly expanding population of Leeds was engulfing neighboring townships like Headingley-cum-Burley. Another transaction facilitated by estate acts was the mortgaging of land. Equitable estates sometimes restricted life tenants from borrowing against their estate. In most cases, standard mortgages could not be taken out

20


because in the event of a default the creditor could not seize settled land. Some settlements authorized borrowing or the ‘‘raising of money,’’ but usually only for specific purposes, such as raising funds for daughters’ dowries or sons’ educations, to refinance existing debts, or to pay for improvements. Landholders might need to borrow for other reasons, and in those instances too, they turned to estate acts. An estate act could authorize a mortgage in several ways. One was to vest authority in trustees who raised the funds. For example, an act in 1741 vested ‘‘Part of the Marriage Portion of Mary, late Wife of John Walcot Esquire, and also part of his settled Estate in Trustees for raising Money to pay Debts.’’36 Another was to give the life tenant authority to borrow. For example, an act in 1726 enabled ‘‘Daniel Dunn Esquire, by Sale or Mortgage of Part of his Estate, to raise Money to pay off and discharge the Portions of his Brothers and Sister.’’37 An act in 1807 enabled ‘‘Rear Admiral Bentinck, Tenant for Life under the Will of his late Father John Albert Bentinck Esquire deceased, to charge his Estates in the County of Norfolk with the Sums y for the embanking, improving, and increasing the same Estates by the Means therein mentioned.’’38 Estate acts authorized an array of additional transactions intended to improve the efficiency of estates. Some specifically authorized the opening of mines and extraction of coal. Many of these authorized the extraction of coal in midlands and northern counties on the eve of the Industrial Revolution. A handful authorized the exploitation of tin deposits in southwestern England. Property owners requested these acts when the rules under which they inherited their estate did not explicitly authorize the mining of minerals. The default rule prohibited mining, which was considered to be a legal act of ‘‘waste,’’ since it consumed a nonrenewable resource, which law and custom preserved for the tenants in tail (i.e., future heirs). Property owners seeking to raise capital to open mines (or to rent the rights to work the mines to third parties) needed acts of Parliament to ensure partners that capital invested in the mine could not be expropriated by individuals with beneficial interests in the estate, and thus, the right to sue in Chancery court and shut down the mine during the (usually very long) period of time while Chancery pondered the case. Over 60 estate acts authorized the harvesting (i.e., cutting and sale) of timber. For example, an act in 1798 called for ‘‘Timber to be cut upon the settled Estates of Le Gendre Pierce Starkie Esquire, and applying the Money to arise there from in the Purchase of other Estates, to be settled to the same Uses.’’39 Property owners requested these acts when the rules under which they inherited their estate did not clearly authorize harvesting trees from

21


particular plots of land. The default rule prohibited the cutting of timber, which was considered to be a legal act of ‘‘waste,’’ since it consumed a nonrenewable resource, which law and custom preserved for the future heirs and other beneficiaries. But restrictions on harvesting timber could harm long-term financial prospects when the price of wood changed or when oldgrowth trees stunted new growth. In such circumstances, property owners could approach Parliament for authority to harvest the timber. Authority often came in the form of an act requiring the life tenant to set aside part of the proceeds of the sale of timber for the dependents of the estate and to make sure that no one lost income from the arrangement. To summarize, estate acts’ primary task was to break the fetters imposed by settlements. The acts authorized a variety of transactions prohibited by strict settlements, including sales, leases, exchanges, partitions, and mortgages on property. Estate acts also authorized the mining of iron, coal, and other ores as well as the cutting and sale of timber which were forbidden in many settlements before the nineteenth century. Of course the fact that estate acts enabled these transactions does not necessarily imply that they would not have been undertaken in the absence of an estate act. Individuals could have conducted transactions on settled lands even though they were open to legal challenges in court.40 A definitive causal assessment of the effects of estate acts on property transactions is still missing in the literature and is an important topic for future research.

6. TRENDS The number and composition of estate acts changed over time. Table 3 lists the number of estate acts and the types of transactions they facilitated in each year from 1600 to 1830. These statistical series are especially useful for doing time-series analysis on the British economy.41 Fig. 3 plots the annual number of sale acts by year from 1600 to 1830. Like all acts of Parliament, sale acts were rare in the early 1600s. No acts appear to have been passed during the Interregnum of the 1640s and 1650s, because after the Restoration the monarchy annulled all acts passed in its absence.42 After the Restoration, the number of sale acts increased for a brief period, but then declined during the political instability of the 1670s and 1680s. During the 1690s, the number of sale acts surged and remained at a high level for several decades. During the eighteenth century, the number of sale acts averaged 11 per year, with much variance around that mean.

22


Number of Estate Acts by Year and Type of Transaction, 1601–1830.

Table 3. Year

Sale

Lease

Exchange

Discharge

1601 1602 1603 1604 1605 1606 1607 1608 1609

1 0 5 0 7 5 0 0 9

0 0 1 0 1 0 0 0 0

0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0

1610 1611 1612 1613 1614 1615 1616 1617 1618 1619

0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0

1620 1621 1622 1623 1624 1625 1626 1627 1628 1629

0 0 0 11 0 0 0 2 0 0

0 0 0 0 0 0 0 0 0 0

1630 1631 1632 1633 1634 1635 1636 1637 1638 1639

0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0

Mortgage

Partition

Timber

0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0

0 0 0 1 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0

23


Table 3. (Continued ) Year

Sale

Lease

Exchange

Discharge

1640 1641 1642 1643 1644 1645 1646 1647 1648 1649

2 0 0 0 0 0 0 0 0 0

3 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0

1650 1651 1652 1653 1654 1655 1656 1657 1658 1659

0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0

1660 1661 1662 1663 1664 1665 1666 1667 1668 1669

8 7 11 4 11 0 5 2 0 0

1 0 0 2 1 0 0 2 0 0

1670 1671 1672 1673 1674 1675 1676 1677 1678 1679

21 0 2 0 0 3 0 18 4 1

2 0 1 0 0 0 0 2 0 1

Mortgage

Partition

Timber

0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0

0 0 2 0 0 0 0 2 0 0

0 1 1 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 1 0 0

0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 1 0 0

0 0 0 0 0 0 0 0 1 0

0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0

24



Sale

Lease

Exchange

Discharge

1680 1681 1682 1683 1684 1685 1686 1687 1688 1689

0 0 0 0 0 1 0 0 7 9

0 0 0 0 0 0 0 0 2 1

0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0

1690 1691 1692 1693 1694 1695 1696 1697 1698 1699

10 19 15 2 19 12 15 21 34 0

3 5 2 0 7 3 1 4 7 0

0 0 2 0 1 2 1 1 0 0

1700 1701 1702 1703 1704 1705 1706 1707 1708 1709

7 1 30 17 31 30 15 7 17 12

0 0 7 5 5 5 3 1 2 1

1710 1711 1712 1713 1714 1715 1716 1717 1718 1719

21 15 9 12 9 0 9 4 7 9

6 3 2 1 2 0 1 1 1 1

Mortgage

Partition

Timber

0 0 0 0 0 0 0 0 2 1

0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0

0 0 0 0 1 1 0 2 1 0

1 4 1 0 5 1 0 1 4 0

0 0 0 0 0 0 0 0 0 0

0 0 1 0 1 0 1 0 1 0

0 0 2 1 0 1 2 0 1 0

0 0 1 1 0 0 3 1 0 0

3 0 3 0 5 1 0 0 2 0

0 0 1 2 0 0 1 0 0 1

0 0 1 0 0 0 0 0 0 0

1 0 1 0 0 0 2 1 0 1

1 1 2 0 0 0 0 2 1 1

0 1 3 2 2 0 0 0 1 1

1 1 0 1 0 0 2 0 0 1

1 2 0 0 0 0 1 0 0 0

25



Sale

Lease

Exchange

Discharge

1720 1721 1722 1723 1724 1725 1726 1727 1728 1729

15 5 8 10 17 13 15 8 9 6

0 1 0 1 2 4 1 4 3 1

1 0 1 0 1 0 2 0 2 0

2 1 0 0 2 2 1 1 2 1

1730 1731 1732 1733 1734 1735 1736 1737 1738 1739

8 13 10 6 5 6 9 4 12 1

2 3 4 1 0 1 4 3 3 1

2 3 1 1 2 1 0 1 2 2

1740 1741 1742 1743 1744 1745 1746 1747 1748 1749

9 11 4 12 11 9 15 11 14 10

2 2 4 0 0 1 4 0 2 4

1750 1751 1752 1753 1754 1755 1756 1757 1758 1759

10 5 0 12 13 11 14 18 7 9

5 4 0 4 3 4 2 5 6 8

Mortgage

Partition

Timber

0 0 1 1 1 1 2 1 0 1

0 0 0 0 0 1 0 1 0 0

0 0 1 0 0 0 0 0 0 1

0 0 1 1 1 1 0 2 5 0

0 3 1 0 0 0 1 0 0 0

0 0 0 0 1 0 2 0 0 0

0 0 0 0 0 0 0 0 0 0

0 1 2 1 1 1 0 1 0 0

3 0 2 0 1 3 1 0 3 0

1 2 0 1 1 1 3 1 1 2

1 1 1 0 0 0 0 1 1 1

0 0 0 0 0 0 0 0 0 1

0 2 0 2 2 5 1 0 2 2

3 3 0 1 0 0 1 2 3 2

1 1 0 0 1 3 7 2 0 0

1 0 0 2 0 2 1 2 1 0

0 0 0 0 1 0 0 1 2 0

26



Sale

Lease

Exchange

Discharge

1760 1761 1762 1763 1764 1765 1766 1767 1768 1769

6 0 13 3 7 18 16 19 10 14

2 0 4 1 6 0 5 7 3 5

2 0 0 0 0 3 0 2 0 2

2 0 4 2 3 1 0 5 1 5

1770 1771 1772 1773 1774 1775 1776 1777 1778 1779

11 14 19 17 16 21 22 14 11 9

4 4 9 4 6 5 3 4 2 3

1 3 3 4 4 11 2 2 3 2

1780 1781 1782 1783 1784 1785 1786 1787 1788 1789

9 4 5 6 7 10 13 3 3 13

0 1 0 5 3 3 5 0 5 2

1790 1791 1792 1793 1794 1795 1796 1797 1798 1799

6 6 10 8 7 10 13 14 11 11

2 6 7 10 7 5 5 4 3 3

Mortgage

Partition

Timber

1 0 1 1 1 1 1 2 0 3

1 0 2 0 2 1 0 2 0 1

0 0 0 0 0 0 0 1 0 1

5 2 4 5 2 2 5 5 1 0

0 2 3 0 1 1 2 0 0 2

2 0 1 2 0 1 0 0 2 1

1 0 3 2 0 0 2 0 2 1

3 1 0 1 4 4 3 1 4 1

0 4 2 1 2 3 0 4 1 0

3 0 2 0 0 1 1 0 0 1

0 0 0 0 0 1 0 0 0 0

0 0 0 0 1 0 1 0 2 2

3 4 2 3 2 5 3 5 2 2

2 1 1 1 1 2 0 3 2 2

0 1 1 0 0 1 2 0 0 2

1 0 1 0 1 3 2 2 0 2

0 1 1 0 1 0 1 1 2 1

27



Sale

1800 1801 1802 1803 1804 1805 1806 1807 1808 1809

17 14 16 15 5 14 16 23 18 20

3 3 7 3 3 7 5 5 7 4

2 1 6 7 1 5 3 2 6 1

0 0 1 4 1 2 0 0 2 1

1810 1811 1812 1813 1814 1815 1816 1817 1818 1819

11 12 14 29 16 10 11 12 7 13

5 4 7 7 3 3 1 4 3 6

5 3 6 4 6 1 6 3 3 2

1820 1821 1822 1823 1824 1825 1826 1827 1828 1829

12 11 14 9 14 15 11 16 16 14

4 5 4 3 8 24 12 13 6 10

1830

10 1,814

Total

Lease

Exchange

Discharge

Mortgage

Partition

Timber

1 3 2 0 1 0 1 1 0 0

2 1 2 1 0 0 0 1 1 2

0 1 2 0 0 0 0 1 0 3

2 2 0 2 1 2 2 2 0 2

0 0 1 1 0 0 0 0 0 0

3 2 0 0 0 2 1 1 2 1

3 1 0 0 0 0 0 0 0 1

4 3 1 7 2 4 2 2 3 2

0 0 0 1 1 1 0 4 1 0

0 0 1 0 0 0 0 1 0 0

0 1 1 0 0 1 0 3 2 0

0 1 0 0 0 0 0 0 1 2

8

3

0

1

1

0

538

273

193

137

93

60

Fig. 4 shows the number of acts authorizing leases in each year from 1600 to 1830. The pattern resembles sale acts. The correlation between these series is 0.61. Like sale acts, lease acts were rare during the early 1600s but increased after 1660. From 1715 to 1790, the number of lease acts averaged 3 per year. After 1790, as the pace of urbanization and industrialization

28

DAN BOGART AND GARY RICHARDSON 35

30

Acts per Year

25

20

15

10

5

0 1600

1650

1700

1750

1800

Fig. 3. Number of Estate Acts Authorizing Property Sales, 1600–1830. Source: See text. Note: The gray dots indicate the number of acts authorizing property sales passed in each year. The black line is an 11-year moving averages of the annual data as defined in the note for Fig. 1.

increased, the number of lease acts rose. For the next four decades, the number of lease acts averaged 6 per year. The peak occurred in 1825, when 25 lease acts were passed. These patterns should interest scholars for several reasons. First, they reflect political events, such as the Restoration of 1661 and the Glorious Revolution of 1689. After the Restoration, the number of estate acts surged. During the political turmoil of the 1670s and 1680s, the number declined. After the Glorious Revolution of 1689, sales surged again. There are several potential explanations as to why the number of estate acts surged following the Glorious Revolution. One class of arguments focuses on the behavior of Parliament. For example, Parliament began meeting regularly and streamlined procedures for reviewing estate bills in the late 1690s and early 1700s, providing a quick and inexpensive forum for modifying rights to equitable estates.43 The shift to regular sessions of Parliament and the procedural improvements adopted after 1689 can be interpreted as a positive supply shock which effectively lowered the cost of getting estate acts.

29

Estate Acts, 1600–1830 25

Acts Per Year

20

15

10

5

0 1600

1650

1700

1750

1800

Fig. 4. Number of Acts Authorizing Property Leases, 1600–1830 Source: See text. Note: The gray dots indicate the number of acts authorizing property leases passed in each year. The black line is an 11-year moving averages of the annual data as defined in the note for Fig. 1.

Another class of arguments focuses on changes in the demand for estate acts. This explanation would suggest that the increased quantity of acts was driven by economic developments like population growth. The composition of estate acts in the period before and after the Glorious Revolution sheds light on the relative importance of supply and demand changes. Table 4 shows that the composition of acts from 1660 to 1688 resembled the composition from 1689 to 1719. A little over 50 percent of all estate acts authorized sales in both periods. A little less than 10 percent authorized leases in both periods. These similar distributions suggest that the Glorious Revolution did not change the types of estate acts that were passed, or the composition of demand for acts, but it did change the volume of estate acts that were passed, probably by relaxing constraints on the supply of acts. Table 4 also shows that the composition of estate acts changed after 1719. After that date, the proportion of acts authorizing leases, exchanges, and discharges increased. The rate of increase accelerated during the Industrial Revolution period spanning from 1760 to 1830. The most rapid increase

30

Table 4.


Percentage of Estate Acts by Period and Type of Transaction.

Transaction

1660–1688

1689–1719

1720–1759

1760–1830

1660–1830

Sale Lease Exchange Discharge Mortgage Partition Harvest timber Mine ore/coal

55.0 7.3 2.1 1.6 1.0 0.0 0.5

52.7 10.1 2.5 2.4 5.0 1.3 1.1

51.5 13.2 6.0 6.8 5.2 2.7 0.9

51.1 20.1 12.0 7.0 3.0 3.7 2.5

51.5 15.3 10.2 5.5 3.8 2.6 1.7

Total number of acts

191

793

748

1,692

Source: See text.

occurred in the number of lease acts. Lease acts frequently facilitated construction in urbanizing and industrializing areas and played an important role in reallocating resources toward infrastructure, industry, and mining. The time trends have implications for a number of debates in social history. Habakkuk (1980, 1994) famously argued that aristocratic families struggled to maintain their social position relative to a rising entrepreneurial and mercantile class. In this struggle, aristocratic families supposedly sold valuable properties to pay for debts incurred maintaining expensive lifestyles. Estate acts served as a mechanism for facilitating these sales. The information in estate acts can address the validity of Habakkuk’s hypothesis. Sale acts often described the conditions that Parliament imposed upon the property sales. One common condition was that the proceeds of the sale be used to pay debts. The phrase ‘‘to pay debts’’ is one indicator of a crisis among the aristocratic families. Social competition with the bourgeoisie forced aristocratic families to sell land in order to finance lavish lifestyles and maintain relative social rankings. As entrepreneurs and merchants accumulated capital and acquired the trappings of privileged lifestyles, competition among classes became increasingly intense, and the aristocracy acquired even larger debts, forcing them to sell even more of their most productive lands. The class dynamic driving this model should leave a clear pattern in the evidence. As the rise of the bourgeoisie forced the aristocracy to acquire larger debts, more estate acts should refer to debts as the reason for selling property. Fig. 5 addresses this dynamic by reporting the percentage of sale acts indicating the proceeds should be used to pay debts. Before 1690 more than

31

Estate Acts, 1600–1830 60

Percentage of Sale Acts

50

40

30

20

10

0 1660

1685

1710

1735

1760

1785

1810

Fig. 5. Percentage of Acts that Authorized Sales of Land and Required the Proceeds to be Dedicated to Either the (a) Payment of Debts or (b) Purchase of Land Settled to the Same Use. Source: See text. Note: The filled circles indicate an 11-year moving average for the percentage of acts authorizing property sales that indicated the proceeds should be used for the payment of debts. The open circles indicate an 11-year moving average for the number of acts authorizing property sales passed in each year that required the proceeds to be used to purchase land of equivalent value settled to the same use.

three-fourths of all acts listed payment of debts, but the proportion fell continuously after the Glorious Revolution. By the time of the Industrial Revolution, less than one act in four listed payment of debts as a rationale for the transaction, and then, often one among many rationales. Most acts mention debts in the context of raising portions and/or jointures (i.e., payments to dependents such as widows and younger siblings, often upon reaching adulthood, to pay for education, to establish them in careers, or to fund dowries). Many estate acts also mention debts in the context of raising funds to invest in improvements.44 Another dynamic is clearly evident in Fig. 5. As time progressed, estate acts increasingly required the first use of the proceeds of land sales to be purchasing equivalent land (or at times other assets) which would provide beneficiaries of the equitable estate with incomes equivalent to what they

32


would have received under the old arrangement. After beneficiaries’ interests had been accommodated, the remainder of the proceeds from the sale of lands could be dedicated to the purchase of additional land or to new uses, such as investment in infrastructure, mercantile ventures, or industrial concerns. The motivation for sale acts thus appears to be quite different in the industrial revolution period. The following section on geographic patterns provides addition evidence that estate acts were interconnected with industrialization.

7. GEOGRAPHIC DISTRIBUTION Estate acts affected particular pieces of property in particular places. Our database indicates the region, county, city, and/or street of affected properties for approximately two-thirds of all estate acts. Table 5 reveals the regional distribution of these acts. Column (3) examines acts for which our database contains geographic information. Of these acts, 85 percent pertained to property in England, approximately 5 percent referred to property in Ireland, 5 percent to property in Scotland, 4 percent to property in Wales, and 1 percent to property in colonies overseas.45 The concentration of acts in England was due, in part, to the concentration of population and land area in that part of the United Kingdom. The concentration may also reflect the difficulty of requesting an estate act if one lived on the periphery of the empire and far from the center of power in London. Table 5.

Regional Distribution of Estate Acts Relative to Population and Land Area. Number of Percent of Percent of Population Percent of Land Area Percent of Estate Acts Estate Located in 1801 Population in Square Land in (1) Acts Acts (000s) (5) Kilometer UK (2) (3) (4) (6) (7)

England Ireland/Isle of Man Scotland Wales Colonies Location unidentified

Source: See text.

2,063 131

58.6 3.7

85.3 5.4

8,308 1,388

70.2 11.7

130,395 13,843

53.6 5.7

117 93 14 1,135

3.3 2.6 0.4 32.2

4.8 3.8 0.6 –

1,550 587

13.1 5.0

78,313 20,754

32.2 8.5

–

–

–

–

33


In addition, Ireland (before 1801) and Scotland (before 1707) had indigenous Parliaments. Acts passed by these local legislatures do not appear in our database, which contains information only from the Parliament at Westminster.46 For estate acts affecting land within England, our database often identifies the county (or counties) in which the property (or properties) was (or were) located. An estate act could pertain to property located in multiple counties. Such an act would not be unusual, because the estates of landed families often spanned several counties. A landed family might possess a house in London, a country estate near their ancestral lands, agricultural manors scattered in several counties, and rights to revenues from fairs, markets, tithes, or townships scattered around the realm. Table 6 indicates the number of counties referred to in each act. The preponderance, nearly 80 percent, affected properties in a single county; roughly 15 percent affected properties in two counties; the rest affected properties in 3 or more counties. These percentages indicate estate acts typically affected land in one locality, although they could be used to reorganize rights to land over broad areas. The proportion of acts affecting properties in more than one county remained stable over time. Table 7 indicates the types of estate acts in each county from 1600 to 1830. The county with the greatest number of acts authorizing sales, leases, and partitions was Middlesex. The obvious reason for this concentration was the rapid expansion of London. The counties with the fewest acts Table 6.

Geographic Breadth of Estate Acts Within England, By Number of Counties and Time Period. Number of Counties Named in Act

Number of acts Percent of acts Period

1600–1688 1689–1719 1720–1759 1760–1830 Source: See text.

1

2

3

4

5þ

1,422 78.4

287 15.8

80 4.4

20 1.1

10 0.6

Percent of acts naming one county

Percent of acts naming more than one county

82.8 78.7 75.0 77.2

17.2 21.3 25.0 22.8

34

Table 7.


Number of Estate Acts Changing Property Rights by County and Type of Transaction, 1600–1830.

County

Sale

Lease

Exchange

Discharge

Mortgage

Partition

Timber

Bedfordshire Berkshire Buckinghamshire Cambridgeshire Cheshire

16 32 37 22 37

0 6 0 1 15

4 4 7 6 6

7 4 5 5 4

1 0 1 4 2

1 0 1 0 2

0 0 0 1 2

Cornwall Cumberland Derby Devon Dorset

12 3 22 53 26

13 1 4 17 6

3 1 9 9 10

3 1 3 7 4

1 0 1 1 1

0 0 1 1 3

0 1 1 0 1

Durham Essex Gloucester Hampshire Hereford

17 54 39 32 9

3 8 10 5 1

2 5 9 7 6

1 6 5 4 5

0 0 2 2 0

3 6 3 2 0

1 3 2 4 0

Hertford Huntingdonshire Kent Lancashire Leicestershire

38 12 90 42 38

4 2 21 42 3

10 4 17 4 6

9 0 13 4 4

2 2 1 3 1

3 2 7 4 4

1 0 2 3 0

52 104 49 26 19

3 93 2 1 4

10 13 16 6 5

9 14 10 6 0

5 5 4 2 2

9 13 1 4 0

1 2 1 3 1

Nottinghamshire Oxfordshire Rutland Shropshire Somerset

23 25 4 22 41

5 5 0 1 11

11 16 1 5 8

5 6 0 1 5

2 2 0 1 0

2 2 2 0 4

2 0 0 0 1

Staffordshire Suffolk Surrey Sussex Warwick

35 43 69 47 28

10 2 47 6 9

9 15 9 16 6

5 4 9 7 6

1 0 3 4 3

1 5 9 4 1

1 2 2 3 0

Westmoreland Wiltshire Worcester York

3 45 20 82

0 10 4 14

2 8 7 24

1 6 5 16

0 2 1 11

1 5 2 10

1 3 1 5

Lincolnshire Middlesex Norfolk Northamptonshire Northumberland

Source: See text.


35

authorizing sales or leases were Westmoreland and Cumberland. These counties were sparsely populated and far from London and other industrializing areas. The county with the highest number of acts authorizing exchanges, discharges, mortgages, and the harvesting of timber was York, which was large in terms of land area, densely populated, and contained expanding industrial cities such as Leeds and Sheffield. To control for differences in the land area of each county, Table 8 ranks the number of acts per square mile in each county. Middlesex ranked highest for every type of transaction. Surrey, which lies just south of Middlesex and just across the Thames from the City of London, ranked second in terms of sale and lease acts per square mile. Hertfordshire, which lies just north of Middlesex, ranked third for acts authorizing sales, exchanges, and discharges per square mile. Lancashire – home to the cotton textile industry and the rapidly expanding industrial centers of Manchester and Liverpool – ranked third in acts authorizing leases per square mile. Its rapid urban growth must have been one reason for the large number of acts authorizing building leases. Over time, the geographic distribution of estate acts changed, as Table 9 shows, by indicating the number of sale and lease acts per square mile in each county during the 160 years from 1600 to 1759 and the 70 years from 1760 to 1830. During the latter period of industrialization acts authorizing sales and leases became more concentrated close to London, in the counties of Middlesex and Surrey, and more concentrated in the industrializing counties to the north and west of London. Lancashire experienced the most rapid rate of growth in acts per square mile, probably, it is worth repeating, due to the expansion of the cotton centers of Manchester and Liverpool. Cheshire, which was adjacent to Lancashire, also experienced rapid growth in lease acts per square mile. Counties lacking industrial centers, such as Hertfordshire and Buckinghamshire, which ranked highly in acts per square mile before 1760, declined in rank.47 These tables demonstrate that estate acts became increasingly concentrated in industrializing and urbanizing regions during the first Industrial Revolution era from 1760 to 1830. This pattern reveals a correlation between changes in property rights and economic development in England. The nature of this link remains a matter of research. The restrictions associated with strict settlements should have become more binding as the pace of urbanization and industrialization increased and as switching land from agricultural to urban or industrial uses became more profitable. Moreover, it would have been difficult, if not impossible, for landowners to reallocate resources in response to urbanization and industrialization

36


Table 8. Rank of Counties According to the Number of Estate Acts Changing Property Rights Per Square Mile, By Transaction Type, 1600–1830. County

Sale Lease Exchange Discharge Mortgage Partition Timber Mining All


8 7 5 22 9

37 8 36 29 4

14 26 13 20 23

2 17 8 11 21

9 31 20 2 14

17 37 26 31 20

31 38 33 23 10


37 39 28 29 23

7 35 23 14 18

37 39 12 32 10

32 36 25 29 18

27 38 25 35 22

33 39 28 38 15

35 25 21 39 15


34 10 15 30 35

26 19 9 24 34

38 33 18 28 21

35 23 20 30 12

37 32 13 26 30

24 10 19 29 36

19 6 9 12 37


3 13 4 26 6

15 20 5 3 21

3 7 6 36 19

3 37 6 31 16

5 3 18 17 21

5 4 8 22 6

4 27 13 11 28


31 1 27 24 36

32 1 31 30 27

30 1 16 25 34

28 1 15 13 39

16 1 12 11 29

13 1 32 7 35

30 1 34 3 24


20 11 21 32 25

17 11 38 33 13

4 2 22 31 27

10 5 38 33 26

10 6 39 28 34

18 11 2 34 23

8 29 26 36 18


17 18 2 14 16

10 28 2 22 6

15 11 5 8 17

19 24 4 14 9

23 33 4 8 7

27 16 3 12 25

20 16 2 7 32


38 12 19 33

39 12 16 25

35 24 9 29

34 22 7 27

36 24 19 15

30 9 14 21

22 5 14 17

Source: See text.

37


Table 9.

Sale and Lease Acts Per 100 Square Miles in Each County, 1600–1759 and 1760–1830.

County

Sale

Lease

Before 1760

After 1760


2.4 2.3 2.7 1.6 1.5

1.1 2.0 2.3 0.9 2.0


0.4 1.1 1.4 1.5

0.4 0.2 1.1 0.6 1.1


0.6 2.0 1.4 1.0 0.5

0.6 1.6 1.7 0.9 0.6

0.1 0.3 0.1

0.2 0.5 0.5 0.2 0.1


3.3 2.4 2.7 0.8 3.3

2.7 0.8 3.1 1.5 1.4

0.3 0.3 0.2 0.2 0.2

0.3 0.3 1.2 2.2 0.1

0.9 13.8 1.1 1.5 0.3

1.1 23.0 1.3 1.1 0.7

0.1 6.4 0.1 0.1 0.1


1.3 1.7 2.7 0.6 1.1

1.4 1.6

0.1 0.1

0.5 0.5

1.0 1.4

0.1 0.5

0 0.2


1.0 1.9 4 1.2 1.4

1.9 0.9 5.1 2 1.7

0.1 0.1 1.1 0 0.2

0.7 0.1 5.1 0.4 0.8


0.1 1.2 1.8 0.5

0.3 2.1 1 0.9

0 0.4 0.3 0.1

0 0.3 0.3 0.2


Source: See text.

Before 1760

After 1760

0.3

0.5

0.1 0.4

0 1.0

0.6

0.4 0.1 0.3 0.2 0.1

0.1 0.4 0.5

26.6

0.2

38


without reorganizing property rights. The Chancery court was not a good substitute for Parliament, especially in the late eighteenth and early nineteenth century (Baker, 1971). Without this parliamentary forum, restrictions on property transactions would have likely remained and many crucial investments would have been forgone to the detriment of both static and dynamic efficiency.

8. RANK, PROFESSION, AND GENDER OF INDIVIDUALS NAMED IN ESTATE ACTS Estate acts reorganized rights held by individuals and families. Who held these rights? Who was able to get acts from Parliament? Did Parliament intercede on the behalf of all property holders, or did Parliament help only particular classes of people, such as aristocrats or the politically powerful? Our database enables us to answer these questions because clerical titles often named the parties involved. Most clerical titles named the life tenant (or the trustee for underage landholders) who possessed the property. Many acts named other parties with beneficial interests in the estate or otherwise involved in the transaction. Some acts named deceased individuals whose settlements were the source of the property in dispute. The social ranks and/or professions of these individuals were often indicated. Ranks indicating membership in the aristocracy included Baron, Count, Countess, Duke, Duchess, Earl, Marquess, Marchioness, Viscount, Lord, and Lady. Ranks indicating membership in the gentry included Baronet, Esquire, Knight, Gentleman, and Dame. Professions included merchants, doctors, and clerks. Members of the clergy were identified as bishops, reverends, and rectors. In some cases, acts named individuals without indicating ranks or professions. These individuals, in all likelihood, did not belong to the nobility, gentry, or clergy, because identifying membership in these orders would have been valuable before the House of Lords, whose members came from these orders and represented their interests. In many cases, the gender of individuals could also be identified. Women belonging to the nobility received feminine titles such as Countess, Duchess, Marchioness, or Lady. Women belonging to the gentry received the title Dame. Women belonging to other orders were often identified by feminine names, such as Mary or Elizabeth, or by labels such as ‘‘wife of ’’ or ‘‘daughter of.’’

39


Table 10.

Social Ranks and Professions of Individuals in Estate Acts.

Rank or Profession Duke Marquess Earl Viscount Baron Lord Countess Duchess Marchioness Lady Total noble rank

Sale Lease Exchange Discharge Mortgage Partition Timber

All

37 19 109 36 7 81 13 3 1 12 306

26 9 45 9 0 43 8 6 3 3 136

19 7 40 2 0 33 4 2 1 2 104

20 1 22 3 0 16 4 3 0 1 62

6 1 14 6 1 2 1 0 1 5 32

2 2 4 2 1 6 2 0 1 0 17

3 0 4 1 0 6 0 0 0 0 14

134 42 295 71 13 220 46 18 9 37 812

Baronet Knight Esquire Gentleman Dame Total gentry rank

7 278 807 107 22 1,198

3 57 179 18 12 257

1 33 135 10 2 178

2 25 86 6 1 118

1 18 75 5 1 99

0 22 48 5 7 76

0 8 34 1 1 44

20 536 1,496 169 67 2,217

Merchant Doctor Clerk Total professions

12 11 29 52

6 8 8 22

0 2 13 15

0 0 7 7

0 1 1 2

0 3 3 6

0 0 1 1

25 30 67 122

Bishop Reverend Rector Total clergy

8 16 4 28

8 6 14 28

7 10 12 29

0 2 3 5

2 0 0 2

0 5 0 5

1 1 0 2

31 41 35 107

321

122

20

17

9

9

3

556

Individuals without profession or rank

Source: See text. Note: The last row indicates individuals who are named in the acts but whose appellations indicate neither social rank nor profession. Given the prevalence of these honorifics and their importance in this class conscious society, we suspect that the ranks and professions of individuals would have been indicated, if they possessed them.

Table 10 reports the number of estate acts and the rank or profession of the individuals involved. Table 11 reports the same information as a percentage of all estate acts. Noble’s names appeared in 812 acts or 23 percent of the total. Within the nobility, Dukes and Earls accounted for 4 percent and 8 percent of the acts, respectively. Nobles obtained acts authorizing the sale of land at a slightly lower rate than the gentry. This may

40

Table 11.


Percentage Distribution of Social Ranks and Professions by Type of Act. Sale Lease Exchange Discharge Mortgage Partition Timber All

Duke Marquess Earl Viscount Baron Lord Countess Duchess Marchioness Lady Total noble rank

2.0 1.0 6.0 2.0 0.4 4.5 0.7 0.2 0.1 0.7 16.9

4.8 1.7 8.4 1.7 0.0 8.0 1.5 1.1 0.6 0.6 25.3

7.0 2.6 14.7 0.7 0.0 12.1 1.5 0.7 0.4 0.7 38.1

10.4 0.5 11.5 1.6 0.0 8.3 2.1 1.6 0.0 0.5 32.3

4.4 0.7 10.2 4.4 0.7 1.5 0.7 0.0 0.7 3.6 23.4

2.2 2.2 4.3 2.2 1.1 6.5 2.2 0.0 1.1 0.0 18.3

5.0 0.0 6.7 1.7 0.0 10.0 0.0 0.0 0.0 0.0 23.3

3.8 1.2 8.4 2.0 0.4 6.2 1.3 0.5 0.3 1.1 23.1


0.4 15.3 44.5 5.9 1.2 66.0

0.6 10.6 33.3 3.3 2.2 47.8

0.4 12.1 49.5 3.7 0.7 65.2

1.0 13.0 44.8 3.1 0.5 61.5

0.7 13.1 54.7 3.6 0.7 72.3

0.0 23.7 51.6 5.4 7.5 81.7

0.0 13.3 56.7 1.7 1.7 73.3

0.6 15.2 42.5 4.8 1.9 62.9


0.7 0.6 1.6 2.9

1.1 1.5 1.5 4.1

0.0 0.7 4.8 5.4

0.0 0.0 3.6 3.6

0.0 0.7 0.7 1.5

0.0 3.2 3.2 6.5

0.0 0.0 1.7 1.7

0.7 0.9 1.9 3.5


0.4 0.9 0.2 1.5

1.5 1.1 2.6 5.2

2.6 3.7 4.4 10.7

0.0 1.0 1.6 2.6

1.5 0.0 0.0 1.5

0.0 5.4 0.0 5.4

1.7 1.7 0.0 3.3

0.9 1.2 1.0 3.1

Individuals without 17.7 profession or rank

22.7

7.3

8.9

6.6

9.7

5.0

15.8

reflect a trend in which noble families were accumulating property over this period, and where the estates of the largest holders were gradually growing (see Beckett, 1984). The share of acts naming nobles can be compared with the percentage of land owned by the ‘‘great landowners,’’ which consisted largely of the nobility. Beckett (1984) reports that great landowners controlled 15–20 percent of the land in 1690 and 20–25 percent of the land in 1790.48 Since 23 percent of the acts involved the property of the nobility, nobles’ access to (or use of ) estate acts seems proportional to the extent of their land


41

ownership. The nobility, in other words, does not appear to be overrepresented in estate acts. This seems surprising. Nobles’ position at the top of the social and political hierarchy gave them great influence in Parliament, particularly in the House of Lords, where most estate acts originated. One might think that they would treat themselves preferentially, use their political power purely for their own immediate benefit, and pass estate acts only for themselves. Of estate acts, 2,217 or 63 percent of the total named members of the gentry, including the ranks of Baronet, Knight, Esquire, Gentleman, and Dame. Within the gentry, esquire was the most common title, appearing in 42 percent of all acts. The title ‘‘esquire’’ included landowners, lawyers, industrialists, and merchants who possessed substantial estates. Beckett reports that the gentry owned 40–50 percent of the land in 1690 and about 50 percent in 1790. ‘‘Small owners’’ held 25–33 percent of all land in 1690 and 15 percent in 1790. Since 63 percent of estate acts involved members of the gentry, the gentry appear to be over represented among those receiving estate acts. Small holders appear to be under represented, even if one assumes that all of the acts failing to indicate the ranks of the participants dealt with the land of small holders belonging neither to the nobility nor gentry. It is interesting to note that the gentry participated disproportionately in acts for partitioning and mortgaging property. These acts allowed them to divide parcels of property, usually when switching agricultural land to urban uses, and to raise money using their land as collateral. The gentry’s emphasis on these endeavors may reflect greater involvement in the expanding capitalist economy. Noble women were named in 110 acts or 3.2 percent of the total. In many cases, Countess, Duchess, Marchioness, Dames, or Ladies were named in acts because they had recently deceased and the act represented an attempt to change the rights associated with their will. For example, an act in 1724 vested ‘‘the Real Estate of Dame Elizabeth Holford Widow, deceased, in the Parish of Saint Olave, Hart Street, London, in Christopher Appleby Gentleman, and his Heirs, for the better enabling him to sell the same, towards the Discharge of the Charitable and other Legacies given by her Will.’’49 In other cases, estate acts gave women expanded powers. For example, an act in 1720 enabled ‘‘the Lady Viscountess Gage and her Trustees, and Thomas Whorwood Esquire to purchase Lands of Inheritance with the Money arising by Sale of their Estate in the County of Bucks.’’50 Overall, however, it was not common for women to be given expanded powers. This could reflect discrimination against women, but it could also

42


be the case that most upper class women held their wealth in personal property, like government securities.51 Therefore, it is not inconceivable that women with the title Countess, Duchess, Marchioness, Dame, or Lady held 5 percent of the land or less. We should make an important qualification to these figures. Our figures only include legislation that fit our definition of estate acts. Our figures exclude any acts dealing with marital arrangements of the aristocracy and gentry which did not alter rights to real property. Among the transactions authorized by estate acts, aristocratic women were most often named in acts partitioning property. These acts divided parcels of land into separate plots. In these cases, the act usually divided a parcel into two pieces. The husband received rights to one of the new plots. The wife received ownership of the other. For example, an act in 1757 was for ‘‘confirming a Partition between William Earl of Dartmouth, and Frances Catherine Countess of Dartmouth, his Wife, and Sir William Maynard Baronet.’’52 Few estate acts named professionals, like doctors, clerks, businessman, and merchants. Their scarcity may be due to their accumulation of personal property, such as cash and luxuries, and urban real estate, which was settled far less often than rural land. Moreover, merchants that accumulated substantial rural estates would often assume the title of esquire or gentleman. The bottom rows of Tables 10 and 11 indicate the number of acts failing to refer to individuals with ranks or professions. These acts refer to individuals that do not identify as members of the aristocracy, gentry, or profession. An act in 1826, for example, enabled ‘‘the Trustees under the Will of Benjamin Griffin, deceased, to grant Building and other Leases of Parts of the Estates thereby devised.’’53 Acts such as this probably referred to individuals ineligible for honors or titles, since such symbols of status received prominent placement in legal documents and had substantial value in class conscious Georgian society. A small number of acts refer to the property of organizations rather than individuals. For example, an act in 1825 was for ‘‘confirming an Exchange made of certain Parts of the Glebe Lands of the Rectory of Stowlangtoft.’’54 In this case, the rectory as an institution received the authority to exchange land, presumably at the behest of the rector or his superiors. The distribution of ranks, professions, and genders remained stable in the long run. Table 12 indicates the distribution of these categories across time periods. Nobles were named in 25 percent of the acts in the Restoration period and 24 percent in the Industrial Revolution period. The gentry were named in 61 percent of the acts in the Restoration period and 61 percent in

43


Table 12.

Percentage Distribution of Social Ranks and Professions by Time Period. 1660–1688

1688–1719

1720–1759

1760–1830

Duke Marquess Earl Viscount Baron Lord Countess Duchess Marchioness Lady Total noble rank

1.1 0.6 10.2 2.8 0.6 9.0 0.6 0.6 0.0 1.7 24.9

2.6 0.6 6.8 1.8 0.5 4.0 0.8 0.4 0.1 1.6 17.7

4.9 1.1 9.0 2.8 0.1 6.3 1.2 0.8 0.3 1.5 25.4

4.2 1.6 8.2 1.7 0.3 6.6 1.5 0.5 0.4 0.5 23.5


0.6 32.2 26.0 2.8 3.4 61.0

1.1 15.5 41.0 9.1 1.9 65.8

0.7 15.8 46.0 4.1 2.1 66.4

0.3 12.2 44.4 3.3 1.4 60.5


0.0 1.1 0.0 1.1

1.4 0.5 1.3 3.2

0.8 0.7 2.4 3.9

0.5 1.1 2.2 3.9


0.6 0.0 0.6 1.2

1.0 0.0 0.5 1.5

0.8 0.0 0.8 1.6

0.8 2.4 1.4 4.6

15.3

17.9

11.1

17.2

Individuals without profession or rank

the Industrial Revolution period. The shares of smaller categories, such as the clergy, remained small in all periods. In the short run, some patterns appear noteworthy. The proportion of acts naming nobles fell after the Glorious Revolution. This is significant because many estate acts from 1689 to 1719 authorized property sales. If the nobility were less likely to be named in these years, then the nobility may have sold less of their property in this period, and it may be one reason that the share of the land owned by the nobility expanded. Another pattern of interest is the decline in the number of acts not naming ranks or professions

44


during the period from 1720 to 1759 followed by an increase in this group during the Industrial Revolution. One possible explanation is that the greater political control by landed interests and the ‘‘Whig Oligarchs’’ limited access by nonelites to some degree in the mid-eighteenth century. Ultimately, estate acts’ distribution across ranks and professions appears to be a reflection of property’s distribution across these groups. Perhaps this should not have been surprising. While the nobility often employed strict settlements to establish long-lasting estates, settlements were also employed by the rural gentry, and at times, were even used on single-family farms (English & Saville, 1983, p. 12). The arrival of unanticipated opportunities, which settlements could not accommodate, and which were often the impetus to request estate acts, may have been relatively random across land holdings. The social distribution of estate acts suggests that Parliament was readily accessible to a broad cross section of society. As expected the aristocracy certainly made use of estate acts, but so did the gentry and even smaller holders. The accessibility of Parliament is perhaps surprising given that aristocrats dominated the House of Lords and the Commons. Aristocrats might have restricted access to members of their class in an effort to foster their political power. The fact that they did not is of major political and economic importance and remains a puzzle for future researchers to solve.

9. CONCLUSION This essay quantifies the legal, economic, geographic, and social characteristics of estate acts. Estate acts reorganized individuals’ and families’ rights to real and equitable estates. Estate acts allowed landholders to take some action or complete an economic transaction that they could not under the prevailing property-rights regime. The majority of estate acts allowed land to be put on the market. These market-oriented acts facilitated transactions such as the sale, long-term leasing, exchange, partition, and mortgaging of property. Estate acts also enabled landholders to harvest timber and mine metal and coal. The number of acts authorizing property sales increased substantially during the late seventeenth and early eighteenth centuries, with a notable rise following changes in the operation of Parliament after the Glorious Revolution. After 1760, when the pace of industrialization and urbanization increased, estate acts were particularly concentrated in counties such as Middlesex and Lancashire, which were the cradle of the Industrial Revolution.


45

The estate-acts data presented in this paper address an array of additional questions of interest to economic, social, and political historians. Consider the three following examples. First, during the seventeenth, eighteenth, and nineteenth centuries, whose interests did Parliament promote? Who had access to Parliamentary legislation? The proportion of estate acts pertaining to the nobility, gentry, and other social groups corresponds with the share of these groups in total landownership. This finding suggests that access to estate acts was open to all landowners, even those in the middling ranks of the social and political hierarchy. The shares were roughly constant over time. The only anomaly comes in the years (1688–1719) following the Glorious Revolution, when the nobility’s share of estate acts fell below the long-run average and the lower ranks’ share of estate acts rose above their long-run average. This anomaly probably reflects political forces at work during the early period of Parliamentary ascendance. Second, did Parliament’s actions match its rhetoric about acting in the public’s interest and to increase the realm’s wealth? Or alternatively, was Parliamentary legislation primarily a tool for redistribution from the socially powerless to the politically powerful? A great deal of evidence indicates that estate acts served constructive purposes. Acts authorizing long-term leases, for example, typically described the projects, such as the opening of mines or construction of residences, that the leases facilitated. Acts authorizing the sale of property (or otherwise releasing property from the strictures of settlement) typically specified that a portion of the proceeds of the sale must be dedicated to purchasing lands and settling them to the old usage (or taking other actions that would ensure all beneficiaries of the estate remained as well off financially as they had been in the past). The texts of the acts, in other words, reveal Parliament’s intentions. Parliament approved reallocating resources to new and more productive uses, as long as the financial interests of beneficiaries to estates were protected. Third, how did political changes influence the adaptation of property rights via estate acts? Did the Glorious Revolution mark a significant change in ‘‘supply’’ of property rights legislation or was it simply correlated with changes in the ‘‘demand’’ for such legislation. Our findings give preliminary evidence suggesting that the Glorious Revolution represented a positive supply shock. The data show that the composition of transactions authorized by estate acts (i.e., sales, leases, mortgages, etc.) changed little in the decades surrounding 1689. If changes in demand were a more significant driver then one would expect a change in the level and the composition of estate acts. In fact only the level increased. In future research, we plan to investigate the link between political changes and adaptable rights more

46


closely by explicitly controlling for economic factors – like interest rates and domestic trade – which influenced the demand for property rights legislation. Finally, we note that our formulation of estate acts resembles the Paretoimproving approach to allocating property rights which Ronald Coase observed in common law courts and Parliamentary decisions during the nineteenth century (Coase, 1960, 1974). According to Coase, attaining economic efficiency in the presence of transaction costs requires the proper definition and allocation of rights. In the nineteenth century, when Britain was the wealthiest nation in the world, British common and statutory law recognized this principle, and Britain’s Parliament and common law courts assigned rights to maximize productivity. The evidence reveals that Parliament first employed this principle en masse in the early eighteenth century, when Parliament began to pass unprecedented volumes of estate acts, and when the British economy entered an era of expansion preceding industrialization.

NOTES 1. Our quantitative compilation of estate acts builds on a large literature which counts and categorizes acts of Parliament. See Tate (1967, 1978), Turner (1980, 1984), Habakkuk (1980), Wordie (1983), Langford (1991), Hoppit (1996, 2003), Hoppit and Innes (1997), and Innes (1998). 2. See, for example, the classic works by North and Weingast (1989), Neal (1990), and Clark (1998a). 3. Academic inspirations for our evidentiary endeavor include Tate’s (1978) Domesday of Enclosures and Greg Clark’s (1998b) compilation of the Charity Commission Records. These sources form the foundation for rigorous analysis of enclosure acts. 4. For more information about strict settlements, see Bogart and Richardson (2009), from which this section was abstracted. 5. Some of the best known works describing the system of strict settlements are Thompson (1963, 1994), Spring (1964, 1983, 1993), Baker (1971), English and Saville (1983), Beckett (1984), Habakkuk (1994), and Cannadine (1994). 6. The fact that until the Conveyancing Act of 1881, solicitors were paid for conveyances by the word (1s for every 72 words in 1862), did not encourage conciseness (English & Saville, 1983, p. 18). 7. http://www.portcullis.parliament.uk. The Parliamentary Archives provided us with the database underlying Portcullis to facilitate our research. The clerical titles contained within Portcullis were first published in nineteenth-century compilations of Parliamentary legislation, like the Statutes of the Realm (Great Britain, 1810). 8. See, for example, Parliamentary archive reference number HL/PO/PB/1/1707/ 6&7An26.


47

9. See, for example, HL/PO/PB/1/1710/9&10An47, HL/PO/PB/1/1809/49G3n399, and HL/PO/PB/1/1809/49G3n208. Note that in acts of all types, minor variations in the use of articles exists. 10. HL/PO/PB/1/1702/13&14W3&1As1n33. 11. HL/PO/PB/1/1759/33G2n55. 12. HL/PO/PB/1/1692/4&5W&Mn43. 13. HL/PO/PB/1/1725/12G1n35. 14. HL/PO/PB/1/1695/7&8W3n50. 15. HL/PO/PB/1/1749/23G2n57. 16. For a sample of legal opinions on trustees’ powers see Great Britain, House of Commons (1829). 17. HL/PO/PB/1/1692/4&5W&Mn41. 18. HL/PO/PB/1/1773/13G3n179. 19. HL/PO/PB/1/1702/13&14W3&1As1n53. 20. HL/PO/PB/1/1739/13G2n44. 21. HL/PO/PB/1/1785/25G3n104. 22. HL/PO/PB/1/1677/29&30C2n12. 23. HL/PO/PB/1/1733/7G2n40. 24. HL/PO/PB/1/1697/9&10W3n53. 25. HL/PO/PB/1/1735/9G2n33. 26. HL/PO/PB/1/1692/4&5W&Mn26. 27. HL/PO/PB/1/1738/12G2n53. 28. HL/PO/PB/1/1790/30G3n142. 29. HL/PO/PB/1/1719/6G1n25. 30. See Bogart and Richardson (2009) for a discussion of the effects of lease acts. 31. HL/PO/PB/1/1788/28G3n132. 32. HL/PO/PB/1/1736/10G2n59. 33. HL/PO/PB/1/1783/23G3n88. 34. HL/PO/PB/1/1809/49G3n363. 35. HL/PO/PB/1/1827/7&8G4n214. 36. HL/PO/PB/1/1741/15G2n55. 37. HL/PO/PB/1/1726/13G1n70. 38. HL/PO/PB/1/1807/47G3s2n199. 39. HL/PO/PB/1/1798/38G3n212. 40. Bogart and Richardson (2009) discuss some cases where buyers suffered great losses when challenged in courts for purchasing settled land without authorization. 41. See Bogart and Richardson (2010) for a time-series analysis of estate, statutory authority, and enclosure acts. 42. The standard reference on the Acts and Ordinances of the Interregnum (Firth & Rait, 1911) does not contain information about private acts. 43. See Langford (1991) and Hoppit (1996). 44. Another explanation for the appearance of the phrase ‘‘to pay debts’’ may be the legal procedure used to break entails, the process of common recovery. This process involved a suit over a fictitious debt which resulted in the conversion of a settled estate into a fee simple holding. The ubiquity of this standardized procedure – which involved suits over fictitious debts – should make scholars wary of using legal and political records as evidence of the debts of landed families.

48


45. We expect that this distribution will hold for all of England, since we believe that the distribution of locations for acts lacking location information in the clerical title is similar to the distribution for acts whose clerical title contains location information. Further research with the manuscripts of the full acts will eventually indicate the location of all property effected by estate acts. 46. See Hoppit (2003). 47. See Bogart and Richardson (2009) for more on the correlation between estate acts and urbanization. 48. Beckett’s figures come from Mingay (1963), Thompson (1963), and Cooper (1967). 49. HL/PO/PB/1/1724/11G1n72. 50. HL/PO/PB/1/1732/6G2n33. 51. See Green and Owens (2003) for a discussion of women’s wealth holdings in the early nineteenth century. 52. HL/PO/PB/1/1757/30G2n123. 53. HL/PO/PB/1/1826/7G4n242. 54. HL/PO/PB/1/1826/7G4n238.

ACKNOWLEDGMENTS We thank Gregory Clark, Jean-Laurent Rosenthal, Claudia Goldin, Gary Libecap, Lee Alston, Stergios Skaperdas, Linda Cohen, John Wallis, Tim Leunig, Richard Sylla, Mauricio Drelichman, and Joel Mokyr for helpful comments on earlier drafts. We also thank seminar participants at the University of British Columbia, New York University and conference participants at meetings of the Economic History Association, Economic History Society, the All-UC Group in Economic History, and the International Economic History Association. We thank Francesca Labordo and Patricia Suzuki for research assistance. We thank the Parliamentary Archives for assistance and advice. We thank the University of California for financial support.

REFERENCES Baker, J. H. (1971). An introduction to English legal history. London: Butterworth. Beckett, J. V. (1984). The pattern of landownership in England and Wales, 1660–1880. Economic History Review, 37, 1–22, New Series. Bogart, D., & Richardson, G. (2009). Making property productive: Reorganizing rights to real and equitable estates in Britain, 1660 to 1830. European Review of Economic History, 13, 3–30.


49

Bogart, D., & Richardson, G. (2010). Property rights and parliament in industrializing Britain. NBER Working Paper No. 15697. Cannadine, D. (1994). Aspects of aristocracy. New Haven, CT: Yale. Clark, G. (1998a). Common sense: Common property rights, efficiency, and institutional change. Journal of Economic History, 58(March), 73–102. Clark, G. (1998b). The charity commissioners as a source in English economic history. Research in Economic History, 18, 1–52. Coase, R. (1960). The problem of social cost. Journal of Law and Economics, 3, 1–44. Coase, R. (1974). The lighthouse in economics. Journal of Law and Economics, 17, 357–376. Cooper, J. P. (1967). The social distribution of land and men in England, 1463–1700. Economic History Review, 20, 419–440. English, B., & Saville, J. (1983). Strict settlement: A guide for historians. Hull: University of Hull Press. Firth, C. H. & Rait, R. S. (1911). Acts and ordinances of the interregnum. London: TannerRichie. Great Britain. (1810). Statutes of the realm, 1101 to 1713. London: G. Eyre and A. Strahan. Great Britain, House of Commons. (1829). The first report of the royal commission on real property, British Parliamentary Papers (Vol. X). Green, D., & Owens, A. (2003). Gentlewomanly capitalism? Spinsters, widows, and wealth holding in England and Wales, c.1800–1860. Economic History Review, 56, 510–536. Habakkuk, J. (1980). The rise and fall of English landed families, 1600–1800. Transactions of the Royal Historical Society, 30, 199–221. Habakkuk, J. (1994). Marriage, debt, and the estates system: English landownership, 1650–1950. Oxford: Clarendon Press. Hoppit, J. (1996). Patterns of parliamentary legislation, 1660–1800. Historical Journal, 39, 109–131. Hoppit, J. (1997). Failed legislation, 1660–1800. London: Hambledon Press. Hoppit, J. (Ed.) (2003). Parliaments, nations and identities in Britain and Ireland, 1660–1850. Manchester: Manchester University Press. Hoppit, J., & Innes, J. (1997). ‘Introduction,’ in Hoppit, Julian. Failed legislation, 1660–1800. London: Hambledon Press. Innes, J. (1998). The local acts of a national parliament: Parliament’s role in sanctioning local action in eighteenth-century Britain. In: D. Dean & C. Jones (Eds), Parliament and locality, 1660–1939 (pp. 23–47). Edinburgh: Edinburgh University Press. Langford, P. (1991). Public life and the propertied Englishman. Oxford: Oxford University Press. Mingay, G. E. (1963). English landed society in the eighteenth century. London: Routledge and Kegan Paul. Neal, L. (1990). The rise of financial capitalism. Cambridge: Cambridge University Press. North, D., & Weingast, B. (1989). Constitutions and commitment: The evolution of institutions governing public choice in seventeenth century England. Journal of Economic History, 49, 803–832. Spring, E. (1964). The settlement of land in nineteenth-century England. American Journal of Legal History, 8, 209–223. Spring, E. (1983). The family, strict settlement, and historians. Canadian Journal of History, 18, 379–398. Spring, E. (1993). Law, land, and family: Aristocratic inheritance in England, 1300–1800. Chapel Hill, NC: University of North Carolina Press.

50


Tate, W. E. (1967). The enclosure movement. New York: Walker. Tate, W. E. (1978). A domesday of English enclosure acts and awards. Reading, MA: University of Reading. Thompson, F. M. L. (1963). English landed society in the nineteenth century. London: Routledge & Kegan Paul. Thompson, F. M. L. (Ed.) (1994). Landowners, capitalists, and entrepreneurs: Essays for Sir John Habakkuk. Oxford: Clarendon Press. Turner, M. E. (1980). English parliamentary enclosure: Its historical geography and economic history. Folkestone: Dawson-Archon Books. Turner, M. E. (1984). Enclosures in Britain: 1750 to 1830. London: Palgrave. Williamson, O. (1985). The economics institutions of capitalism: Firms, markets, and relational contracting. New York: Macmillan Free Press. Wordie, J. R. (1983). The chronology of English enclosure, 1500–1914. The Economic History Review, 36(4), 483–505, New Series.

THE MACROECONOMIC AGGREGATES FOR ENGLAND, 1209–2008 Gregory Clark ABSTRACT Estimates are developed of the major macroeconomic aggregates – wages, land rents, interest rates, prices, factor shares, sectoral shares in output and employment, and real wages – for England by decade between 1209 and 2008. The efficiency of the economy in the years 1209–2008 is also estimated. One finding is that the growth of real wages in the Industrial Revolution era and beyond was faster than the growth of output per person. Indeed until recently the greatest recipient of modern growth in England has been unskilled workers. The data also create a number of puzzles, the principal one being the very high levels of output and efficiency estimated for England in the medieval era. These data are thus inconsistent with the general notion that there was a period of Smithian growth between 1300 and 1800 which preceded the Industrial Revolution, as expressed in such recent works as De Vries (2008).


51

52

GREGORY CLARK

1. ESTIMATING ECONOMIC GROWTH FROM PAYMENTS TO FACTORS English nominal net domestic income, NDI, is estimated as NDI ¼ wages þ farmland and mineral rents þ tithe payments þ net mine; canal; road; rail; and ship rents þ net house rents þ other net capital incomes þ indirect taxes Real NDI is NDI deflated by the average price of domestic expenditures, PDE. Real net domestic output is NDI deflated by the price of net domestic production, PNDP. These two output measures can differ if export and import prices move differently. Dividing by population we get all of these in per capita terms.

2. MEASURING EFFICIENCY, 1209–1869 The basic measure of the efficiency (total factor productivity) of the economy is an index: At ¼

ðrt þ lÞa PaKt wbt sct Pt ð1 tt Þ

(1)

where r is the return on risk-free capital, l a risk premium, PK the index of price of capital, w the index of wages, s the index of farmland rents, P the price index, and t the share of national income collected in indirect taxes. The price indexes for outputs, wages, and rents are all measured as geometric indexes, with weights changing from year to year. a, b, and c are the shares in factor payments of capital, labor, and land, respectively. These shares are changed every 10 years to reflect changes in the earnings of the different factors over time. Thus, though the index has the Cobb–Douglas form, the changing weights imply that there is no underlying assumption of a Cobb–Douglas technology. In fact, the index is agnostic on the form of the production function, except for an assumption that the capital share is unchanging.

Macroeconomic Aggregates for England

53

3. WAGE INCOME AND THE LABOR MARKET Wages are the most important share of GNI throughout the years 1209–1869, and the most important cost in the index given by (1) with a weight of 50–75 percent. To estimate aggregate wages and labor costs before 1870, the approach here is to first estimate a separate national index of farm day wages, and of non-farm day wages. Then these are aggregated into a national wage using estimates of the structure of occupations, and the relative average wage in the primary sector compared to the rest of the economy. It is shown that for the years 1820–1869 the average national wage estimated on this basis correlates well with a more detailed wage index constructed by Feinstein (1998a, 1998b).

3.1. Farm Wage Index The details of the construction of the farm day wage index are given in Clark (2007a). The wage estimated is the average day wage of farm workers outside harvest. Farm workers typically earned extra income at hay time and the grain harvest. The average premium at harvest (for 6 weeks) was 61 percent, and at hay (for 2 weeks) was 32 percent. Assuming a 300-day (50-week) year, this implies that the average day wage was 8.6 percent greater than the level reported in Table 1. The reported average male farm day wage reflects this adjustment. The prices and wages reported for the earlier years are frequently dated only by an account year which differs from a calendar year. Thus, the most common account year in the medieval period ran from Michaelmas (29 September) to Michaelmas. This was because the harvest was complete only shortly before this quarter feast, and was the natural time for an account to be drawn of the success of the previous harvest season. Later parish accounts often ran from Lady Day (25 March) to Lady Day, or from Easter to Easter, where Easter had no fixed date. In all cases where the exact date of a recorded wage or price is unknown, it is attributed to the calendar year in which the majority of the account year falls.

3.2. Non-farm Wage Index These are estimated as an average of the day wages of skilled and unskilled building workers, as reported in Clark (2005). Again Table 1 shows the data

54

Decade

1200 1210 1220 1230 1240 1250 1260 1270 1280 1290 1300 1310 1320 1330 1340 1350 1360 1370 1380 1390 1400 1410 1420 1430 1440 1450 1460 1470 1480 1490 1500 1510 1520 1530 1540 1550 1560 1570 1580 1590 1600 1610

GREGORY CLARK

Table 1.

Wages, 1200s–1860s.

Farm Wages (d./day)

Building Laborers (d./day)

Building Craftsmen (d./day)

Share in Primary Production

Average Wage (d./day)

Skill Premium (Craft/Labor)

1.37 1.26 1.25 1.18 1.25 1.31 1.33 1.28 1.35 1.32 1.34 1.44 1.54 1.52 1.49 2.80 2.88 3.16 3.17 3.05 3.45 3.47 3.50 3.68 3.66 3.79 3.54 3.56 3.53 3.52 3.38 3.37 3.42 3.32 3.94 5.06 6.15 6.64 6.67 7.13 7.52 7.97

– – 1.63 – 1.88 1.71 1.77 1.45 1.43 1.42 1.57 1.73 1.67 1.70 1.61 2.28 2.75 2.94 2.95 2.88 3.15 3.17 3.31 3.43 3.54 3.62 3.59 3.45 3.45 3.43 3.36 3.47 3.48 3.64 4.04 5.23 6.10 6.42 6.67 6.76 7.61 8.03

2.78 2.08 2.60 – 2.89 3.17 3.10 2.70 2.84 2.83 3.01 3.27 3.23 3.26 2.89 4.06 4.45 4.72 4.62 4.56 4.72 4.89 4.96 5.06 5.29 5.19 5.03 5.13 4.99 5.09 4.93 5.12 5.29 5.45 5.78 7.62 8.63 9.06 9.76 10.0 10.9 12.0

0.60 0.60 0.60 0.60 0.60 0.60 0.60 0.60 0.60 0.60 0.60 0.60 0.60 0.60 0.60 0.60 0.60 0.60 0.60 0.60 0.60 0.60 0.60 0.60 0.60 0.60 0.60 0.60 0.60 0.60 0.60 0.60 0.60 0.60 0.60 0.60 0.60 0.60 0.60 0.60 0.60 0.60

2.02 1.97 1.96 1.88 2.03 2.12 2.15 1.94 2.02 2.00 2.09 2.26 2.31 2.31 2.18 3.61 3.87 4.19 4.17 4.05 4.46 4.52 4.59 4.77 4.84 4.94 4.70 4.72 4.65 4.70 4.50 4.57 4.64 4.65 5.29 6.84 8.12 8.66 8.90 9.33 10.0 10.7

– – 1.56 – 1.92 1.93 1.86 1.89 2.00 2.01 1.93 1.90 1.95 1.92 1.80 1.79 1.63 1.61 1.57 1.59 1.50 1.55 1.50 1.48 1.50 1.44 1.40 1.49 1.45 1.49 1.47 1.48 1.52 1.50 1.43 1.46 1.42 1.42 1.47 1.48 1.44 1.49

55


Table 1. (Continued ) Decade

1620 1630 1640 1650 1660 1670 1680 1690 1700 1710 1720 1730 1740 1750 1760 1770 1780 1790 1800 1810 1820 1830 1840 1850 1860

Farm Wages (d./day)

Building Laborers (d./day)

Building Craftsmen (d./day)

Share in Primary Production

Average Wage (d./day)

Skill Premium (Craft/Labor)

8.31 8.93 9.36 10.1 10.6 9.84 9.89 9.62 9.81 10.0 9.84 10.8 10.6 10.9 11.5 12.3 13.1 15.3 19.4 23.1 20.3 20.0 21.1 22.1 23.6

8.24 9.07 10.0 11.1 11.5 12.0 12.2 12.3 11.9 12.1 12.4 12.6 12.6 13.1 13.9 15.1 15.3 17.9 23.9 29.8 27.0 28.0 29.0 30.1 34.5

12.5 13.3 15.0 16.6 17.6 17.7 17.8 18.5 19.0 19.7 20.0 20.3 20.6 20.5 21.3 22.3 23.4 26.8 35.9 43.8 42.1 42.7 43.3 45.6 52.7

0.60 0.60 0.60 0.60 0.60 0.60 0.59 0.57 0.56 0.54 0.52 0.51 0.49 0.47 0.46 0.45 0.45 0.44 0.43 0.42 0.40 0.38 0.36 0.32 0.28

11.2 12.0 12.9 14.1 14.8 14.4 14.6 14.6 14.9 15.4 15.5 16.4 16.4 16.8 17.8 19.0 19.9 23.1 30.4 37.1 34.4 35.2 36.7 38.9 45.0

1.51 1.47 1.50 1.50 1.53 1.48 1.46 1.50 1.60 1.63 1.62 1.62 1.63 1.57 1.53 1.48 1.53 1.50 1.51 1.47 1.56 1.53 1.50 1.52 1.53

by 10-year averages. It is assumed that the ratio of numbers of skilled to unskilled stays the same throughout the years 1209–1869.

3.3. Share of Labor Force in the Primary Sector For the years 1750–1869 the numbers employed here for employment in the primary sector are those of Shaw-Taylor and Wrigley (2008). Table 2 shows their benchmark estimates for the years 1755, 1817, 1851, and 1871. I interpolated between these benchmarks by assuming the same change in employment share in each year between the benchmarks. The assumed share in primary production in the years before 1680 of 0.60 is much less than is

56

GREGORY CLARK

Table 2. Assumed Sectoral Employment Distribution. Year

Share Primary Sector Assumed Here

Broadberry et al. (Agricultural Population)

Share in Wills (Primary Sector)

0.34 0.42 0.47 0.48 0.50 0.60 0.60 0.60 0.60 0.60

– – – 0.54 – 0.70 – 0.75 0.76 0.79

0.36 0.44 0.45 0.48 0.50 0.60 0.60 – – –

1851 1817 1755 1700 1680 1600 1560/1570 1520/1530 1380 1300

Note: Broadberry et al. (2009, Table 18) and Shaw-Taylor and Wrigley (2008).

Share in Primary Sector

1.0 0.8 0.6 0.4 0.2 0.0 1530

1570

1610

1650

1690

1730

1770

1810

1850

Fig. 1. Share of Will Makers with Primary Sector Occupations, 1580s–1860s. Note: The variation of the measure in the decades 1640–1729 is due to the small numbers of observations in these years.

assumed in a recent paper by Broadberry, Campbell, Klein, Overton, and van Leeuwen (2009), whose assumed shares are shown also in Table 2. The reasons for assuming this smaller primary share are twofold. First, for the years 1510–1800 we can get some ancillary information on occupational structure from the stated occupations of a large number of testators. Fig. 1 shows the share of these testators who reported primary sector occupations by decade from the 1520s to the 1860s, calculated as: s ¼ ð1 jÞsnon-London þ j


57

where snon-London is the share of testators outside London who list primary sector occupations and j the share of the population in London. The share for 1860–1869 is 0.21, close to the numbers reported by Shaw-Taylor and Wrigley. The share for the 1810s is 0.44, compared to 0.42 for Shaw-Taylor and Wrigley. The earlier share of primary sector occupations is never as high as the 70 percent Broadberry et al. assume for 1600 and earlier. Second, the structure of the economy will be closely connected to real incomes (see Clark, Huberman, & Lindert, 1995). We will see below that estimated real incomes in the years 1400–1550 exceed those of 1700–1750, so a much largely primary share for occupations seems unlikely.

3.4. Average Implied Day Wage To estimate the average day wage in the economy as a whole, we need to know what average earnings in the primary sector were compared to the economy as a whole. Leone Levi estimated the average adult male wage in England in 1866 as 45 d. per day (1867, p. 9). The average wage for male farm laborers was, however, only 25.7 d. per day (including an allowance for the harvest premium). However, for each six farm workers there was a bailiff or farmer, many of whom worked on their own accounts, and others who supervised hired labor, who would have higher implied labor incomes. For want of better information, let us assume that premium was 100 percent. This implies an average effective implied male wage in the primary sector of 30.0 d.1 The share of employment in the primary sector, mainly agriculture, in the 1860s was 0.279. This in turn implies an average male day wage in the rest of the economy of 50.8 d. in 1866, a 69 percent premium on average labor income in the primary sector. The average wage in the economy is thus estimated as: W ¼ bosW a þ ð1 bÞsðW c þ W l Þ where b is the share of labor employed in primary occupations, Wa the farm laborer wage, Wc the wage of building craftsmen, Wl the wage of building laborers, and adjustment factors of o ¼ 1.169 and s ¼ 0.582, to set the correct average wage levels in each sector. This wage is shown by decade in Table 1. How reasonable is this approximation of (implied) day wage in the economy in years where we have alternative measures? Charles Feinstein constructed such a series for Britain (England, Scotland, and Wales) for 1770 onwards, using wages from a variety of sectors (1998a, 1998b). Fig. 2

58

GREGORY CLARK 60

d. per day

50 40

Feinstein

30 Clark

20 10 0 1770

1780

1790

Fig. 2.

1800

1810

1820

1830

1840

1850

1860

1870

Estimated Average Day Wages, 1770–1869.

30

d. per day

25 20

Bowley

15 Clark

10 5 0 1770

Fig. 3.

1790

1810

1830

1850

Wages in English Agriculture, 1770–1850. Source: Bowley (1898) and Clark (2001a, 2001b).

shows the respective estimates of the average wage for the years 1770–1869, where the two estimates are set to equality in 1860–1869. The present average wage income series rises 9 percent more between 1770–1779 and 1860–1869 than the Feinstein series. But after 1820 the movements of the two series are very similar. Indeed for those years the R2, when we predict one series from the other, is 0.92. The major differences between the two series arise in the years 1795–1815. This arises largely from the very different wage series for agriculture that is derived in Clark (2001a, 2001b), as compared with the older Bowley series used by Feinstein. As Fig. 3 shows, these two series diverge wildly in these years. Clark (2001a, 2001b) explains

59


Table 3. Category

Men, 20–64 Men, o20 Women, 20–64 Women, o20 All

Labor Income 1866 from Levi.

Number 1860–1869 (Million) 5.37 1.36 2.45 0.90 10.08

Wage per Day, 1866 (d.)

Male-equivalents

Total Earnings, 1866 (Million d)

45 13 25 17

1.00 0.29 0.56 0.38

302 22 77 19

32.1

0.71

420

why this new series is to be preferred. Feinstein himself noted, of the Bowley series on which he relies in these years, that: The most worrying feature of this series is the absence of a reliable benchmark between 1795 and 1824. During these years the index first climbs by some 56 percent y then falls sharply y These large movements are entirely dependent on Bowley’s interpolation on the basis of very limited information. (1998b, p. 187)

Having derived this wage index, total implied labor income in England in 1866 is estimated from Levi (1867) as in Table 3, at d420 million. This is substantially larger than the implied English wages total for England from Feinstein (1972, Table 1), of d298 million. However, this amount includes an allowance for the labor income of the self-employed which is included by Feinstein along with profits up until 1889. When Feinstein first separates these figures in 1889, total labor income is 44 percent greater than income from employment. Applying this adjustment to the 1860s, Feinstein’s d298 million of employment income would translate into a total of d428 million of all labor income, which is very similar to the number calculated here. 3.5. Skill Premium The last column of Table 1 shows the skill premium, which is measured here by the relative wage of skilled building workers compared to building laborers. 3.6. Population and Labor Supply The population before 1540 is estimated as in Clark (2007a). Since this gives decadal estimates, the other years are interpolated geometrically, except in the periods 1310–1319 and 1340–1349 where the timing of the shocks to

60

GREGORY CLARK

Table 4.

Alternative Population Estimates, England, 1300–1600.

Year

Broadberry et al. Population (Million)

Clark (2007a) Population (Million)

Ratio

1300 1348 1351 1380 1520 1600

4.25 3.83 2.56 2.37 2.20 4.12

5.32 4.48 3.54 2.98 2.87 4.28

1.25 1.17 1.38 1.26 1.30 1.04

Source: Broadberry et al. (2009, Table 18).

population in those decades is known. Thus, population is assumed to have fallen to the 1320s level by 1318, and to the 1350s level by 1349. The population after 1540 is estimated from Wrigley, Davies, Oeppen, and Schofield (1997, p. 614) to 1805. Thereafter, the census totals for England including Monmouth are used, interpolating between the census dates. These estimates by decade are shown as the last column in Table 1. These population numbers for the years before 1500 are controversial. Whereas these estimates imply a population for England ca. 1300 of 5.3 million, Bruce Campbell has argued for a much lower population of only 4.25 million. Table 4 shows the population assumed here versus that of Broadberry et al. As can be seen for the years 1520 and earlier, the estimates here are of a consistently higher population by a margin of 17–30 percent. Clark (2007b) explains and defends these population estimates.

3.7. Days Worked per Year This is a very difficult issue. There is widespread belief that the numbers of days per year worked by the population increased greatly between 1200 and 1800, but surprisingly there is little direct evidence of any substantial increase in work days over this interval (Clark & Van der Werf, 1998; de Vries, 1994, 2008; Voth, 2001a, 2001b). Despite the lack of direct evidence of much change in days worked per year, Broadberry et al. assume widely varying days worked per farm family over the years 1250–1850, as is shown in Table 5. They assume an ‘‘industrious revolution’’ in the years 1700–1850, with a one-third increase in days worked per farm family. But they also assume a ‘‘de-industrious revolution’’ in the years 1300–1450, when work days are assumed to decline by nearly 30 percent. Thus, assumed

61


Table 5. Broadberry et al.’s Assumptions about Farm Labor Inputs. Date

Assumed Work Days per Farm Family

Farm Families (Million)

Farm Output (Index)

Output per Work Day (Index)

1250 1300 1380 1450 1600 1700 1800 1850

315 381 331 266 404 405 473 539

0.68 0.74 0.40 0.38 0.64 0.62 0.69 0.73

1.00 1.17 1.07 0.94 1.51 2.18 2.52 4.66

1.00 0.89 1.72 1.97 1.25 1.86 1.64 2.51

Source: Broadberry et al. (2009, Table 24).

work days per year per farm family are double in 1850 what they are in 1450. The reason they make the assumptions about work days per year is in order to reconcile their estimates of farm outputs directly with estimates of farm output from factor payments (wages, land rents, etc.). Farm wages are so high in 1450, for example, that the total farm output implied if all workers were fully employed would greatly exceed the directly estimated output. This mismatch, however, in part stems from the very high assumed farm share of employment that the authors adopt for the years before 1700. If there was a rise in work hours per person in England in the years 1650–1800, we would think that it would be possible to demonstrate it in data from the labor market. However, the evidence here for England, even for male workers, the easiest to observe, is at best ambiguous. At worst it suggests no significant increase in work hours for adult males between even 1250 and 1800. Clark and van der Werf (1998), for example, find evidence for only a very modest rise in days worked per year by men in the years 1560–1860. If workers were employed by the year and by the day, the days per year of the annual workers should be: Days per year ¼

annual wage day wage

Complicating factors, such as that yearly workers have more security and might thus accept a lower daily wage, will affect the exact ratio here. Or again annual workers may be better workers and so get a higher daily wage.2 But as long as the selection process is the same over time, we can use these payment ratios to look at relative days worked per year over time. Table 6 shows this calculation of the typical number of days in the work year as the

62

GREGORY CLARK

Table 6. Period 1867–1869, England 1867–1869, Wales 1870, Scotlanda 1771, England 1733–1736, Norfolk 1700–1732, England 1650–1699, England 1600–1649, England 1560–1599, England

Implied Days Worked per Year.

Observations

Implied Days per Year

Standard Error of Estimate

7 5 27 10 24 3 16 12 17

293 311 280 280 295 286 276 266 257

13.4 5.9 12.9 12.9 4.6 13.7 6.5 5.9 4.8

Source: Clark and van der Werf (1998, Table 1). Note: The figures in italics are calculated from the wage assessments of local magistrates. a Annual wages in these cases are for plowmen, and day wages for ordinary workers. Ploughmen seem to have been regarded as slightly more skilled, which will bias upwards the estimated days.

ratio between the annual payment of workers and the day wage of similar workers. Since the table is based on small samples of workers paid in both ways, the standard error is also estimated. The true number of work days will lie within two times the standard error of the estimate given 95 percent of the time. The implied work year for farm workers in the 1870s was 280–311. Back in 1560–1599 it was only 257. So the best estimate is of a 10–15 percent increase in work days over this interval. This exercise suggests at best modest increases in days per year between 1560 and 1800. Other measures of likely days per year suggest there may have been no increase. We can, for example, calculate at least the common language interpretation of the number of work days in a week in a similar way to days per year. This is by looking at the ratio of the weekly wages quoted for building workers to their daily wage. When a worker was hired for a ‘‘week,’’ what did that mean? Fig. 4 shows the results of this calculation for English 290 wage quotes for building workers in the decades 1250–1620, prior to de Vries’s industrious revolution. There is no trend. And the notional work week is just as long in 1250 as in 1850: it averages 5.97 days. This does not preclude workers more often working broken or shortened weeks in the earlier years – the records of the construction of Exeter Cathedral in the middle ages show many such interruptions.3 But it does show there was no standard pattern of shorter work weeks in the earlier society. For these reasons the estimates here are constructed using an assumption of a constant 300-day work year for men between 1209 and 1869. One way

63

Macroeconomic Aggregates for England 8

Days per week

6

4

2

0 1250

Fig. 4.

1350

1450

1550

Implied Length of the ‘‘Week’’ for Building Workers in England, 1250–1629.

to interpret this is to say that it is giving a measure of the potential NDI of the English economy over these years. If leisure has a value, and that value is close to the daily wage, then even if workers voluntarily took more holidays earlier, this measure shows what the relative living standards were across time. This assumption about working days per year will not affect the estimates of income from land or house ownership. It will, however, affect the estimates of income from working capital in earlier years because of the ways those are calculated.

3.8. Aggregate (Potential) Labor Income Aggregate (potential) labor income is calculated as: W AG ¼ W 300 ðnNÞ where W is the average male day wage, N the population, and n the fraction of the population economically active measured in male-equivalents (assumed to be 0.340). These numbers are shown in Table 7 (by decade). The numbers are set to match the implied total of Leone Levi for the 1860s, with a labor income for this decade of d420.8 million. This implies a total labor income significantly greater than that estimated by Deane and Cole (1967, p. 152), who give a labor income for Britain in 1861 of only d315.4

64

Table 7. Decade

1200 1210 1220 1230 1240 1250 1260 1270 1280 1290 1300 1310 1320 1330 1340 1350 1360 1370 1380 1390 1400 1410 1420 1430 1440 1450 1460 1470 1480 1490 1500 1510 1520 1530 1540 1550 1560 1570 1580 1590 1600 1610

GREGORY CLARK

Population, Labor Income, Taxes, Rents, and Capital Returns. Population (Million)

Labor Income (Million d)

3.16 3.40 3.74 3.90 3.87 3.84 4.31 4.87 4.88 5.32 5.32 5.61 4.97 4.68 4.39 3.54 3.17 3.16 2.81 2.82 2.64 2.54 2.47 2.51 2.27 2.28 2.32 2.38 2.40 2.31 2.56 2.81 2.94 3.02 2.99 3.24 3.21 3.50 3.55 4.16 4.40 4.73

2.84 2.98 3.26 3.25 3.48 3.62 4.12 4.19 4.38 4.72 4.93 5.61 5.10 4.80 4.25 5.68 5.45 5.89 5.21 5.07 5.23 5.10 5.03 5.32 4.89 5.00 4.85 4.99 4.96 4.84 5.12 5.69 6.07 6.23 7.03 9.85 11.57 13.46 14.05 17.25 19.62 22.54

Local Rates Commodity (Million d) Taxes (Million d) 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.00 0.00 0.03 0.03 0.03 0.08 0.06

0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.04 0.07 0.06 0.08 0.09 0.16 0.27

Farmland Rents (Million d)

Return on Capital (Percent)

1.61 1.61 1.63 1.33 1.89 1.42 1.41 1.23 1.28 1.40 1.45 1.62 1.81 1.86 1.75 1.55 1.80 1.63 1.72 1.53 1.71 1.59 1.47 1.59 1.50 1.51 1.53 1.18 1.35 1.21 1.32 1.27 1.60 2.08 2.05 1.98 2.26 3.11 4.86 4.37 9.05 10.22

9.8 10.7 10.1 11.1 9.7 11.3 11.2 11.0 10.2 10.3 8.5 8.0 12.3 9.7 7.1 7.6 8.1 5.0 5.0 5.0 5.0 5.0 5.0 5.0 5.0 5.0 5.0 5.0 5.0 5.0 5.0 5.0 5.0 5.0 5.0 5.13 5.24 4.84 6.55 6.06 6.52 5.99

65



1620 1630 1640 1650 1660 1670 1680 1690 1700 1710 1720 1730 1740 1750 1760 1770 1780 1790 1800 1810 1820 1830 1840 1850 1860

Population (Million)

Labor Income (Million d)

5.02 5.21 5.42 5.61 5.58 5.46 5.40 5.39 5.51 5.69 5.82 5.73 6.05 6.26 6.66 7.01 7.59 8.28 9.09 10.31 11.98 13.77 15.64 17.59 19.72

24.86 27.77 31.16 35.25 36.79 34.86 34.97 35.05 36.47 38.85 40.10 41.64 44.19 46.86 52.58 59.05 67.16 85.18 123.06 169.96 183.16 215.53 254.62 303.99 394.58

Local Rates Commodity (Million d) Taxes (Million d) 0.11 0.12 0.17 0.22 0.21 0.25 0.40 0.55 0.57 0.79 0.90 0.78 1.01 1.09 1.39 1.74 2.31 4.03 6.29 7.52 7.06 6.55 6.59 7.05 8.90

0.30 0.38 0.38 0.47 0.49 0.90 0.90 1.53 2.32 2.88 3.39 3.51 3.42 4.16 5.60 6.16 7.85 10.84 21.84 28.75 29.16 25.88 26.18 28.39 30.28

Farmland Rents (Million d)

Return on Capital (Percent)

10.35 11.37 12.22 12.49 13.29 11.80 12.41 12.02 11.98 13.18 14.13 13.62 12.83 16.10 16.68 19.32 19.33 24.16 33.85 43.32 38.19 36.56 39.17 39.47 43.18

6.34 5.90 5.68 5.63 5.40 5.50 5.28 4.90 4.67 4.96 4.38 4.14 4.24 4.26 4.04 4.15 3.95 4.10 4.38 4.63 4.48 4.85 4.28 4.10 4.27

million, and for 1871 of d408.4 million (implying for the 1860s an English total labor income of d265–343 million).4 But Deane and Cole have another category of income, ‘‘profits, interest and mixed incomes,’’ which includes income from self-employment which is functionally the same as wages. This category is 39 percent of all income in the 1860s (p. 247). Here I have attributed an estimated wage income to all the occupied population.

4. INDIRECT TAXES Before I calculate the income from working capital and entrepreneurial returns, I need to calculate the income derived by government from indirect taxes.

66

GREGORY CLARK

4.1. Indirect Taxes on Property Occupiers One form of taxation in England was that on the occupiers, as opposed to the owners, of property. These were the various local rates – poor rates, county rates, road rates, church rates, and constable rates – which because of the dominance of poor rates were often referred to just as the ‘‘poor rate.’’ Because these taxes were paid by occupiers as opposed to owners, they do not appear above under land rent and tithes, though their incidence probably lay mainly on the rental value of land and houses. There are totals of such rates for England and Wales in the years 1747–1749, 1775, 1782–1784, 1802, and 1812–1869 (Mitchell & Deane, 1962, pp. 410–411). These are converted to an English basis by multiplying by the share of the population English in 1801 (0.94). To estimate poor rate payments in other years, data were collected from the parish accounts of 33 parishes in Bedford, Dorset, Essex, and Warwick, over the years 1577–1869. Payments in each parish relative to the years 1824–1833 were calculated for each year with data. An average (weighted by the size of payments in the years 1824–1833) was then calculated for each year. Payments on this index are shown by 10-year averages in Table 7. Before 1600 the amounts of these taxes were modest, and they are assumed 0 for the years before 1570 when there are no records of their size. Later I also need to calculate what share of these taxes was paid from farmland. In 1832 poor rates per head were about double in parishes with all the employment in agriculture than they were in parishes where none of the employment was agricultural. I thus assume throughout all these years that this differential was the same. Then I calculate the share of poor rates paid by the farming sector as: Share of poor rates from farmland ¼

2y 1þy

where y is the share of the population employed in farming. This would imply that if y ¼ 0.5, the share paid by the farming sector would be 0.67.

4.2. Commodity Taxes In the eighteenth century indirect taxes on commodities became an important source of government income in England. Under the pressures of war finance demands, the government introduced significant taxation of many commodities – beer, wine, candles, bricks, paper, etc. The revenue from these indirect taxes – customs dues and excise taxes – is reported for

67


the United Kingdom (1801–1869), for Great Britain (1689–1800 and 1807–1816), and for England and Wales (1660–1688) (Mitchell & Deane, 1962, pp. 386–388, 392–393). UK figures are reduced to those for Great Britain by multiplying by 0.92, their relative share in 1807–1816. Totals for Great Britain are reduced to those for England alone by multiplying them by 0.84, based on the population of England relative to Great Britain in 1801. Before 1551 indirect taxes are taken as 0 percent of national income, since in the years 1551–1557 they averaged only 0.2 percent of national income. Table 7 shows decadal totals for indirect taxes.

5. PROPERTY INCOME To get the total gross value of income in the economy, we need to add to wage income the returns from ownership of property: land, houses, shops, industrial buildings, roads, canals, waterways, mines, machines, and working capital such as farm animals and horses. After 1842 we have information on such returns from the property and income tax returns. These returns distinguish income from property of the following types: lands, houses, tithes, manors, fines, quarries, mines, iron works, fisheries, canals, and railways. For the 1860s the average of these reported incomes, reduced to the basis of England, was as follows: Farmland and farm buildings (including tithe) Other houses and buildings Profits from land occupation Profits of mines, canals, railways, etc. Other business and professional incomes

d46.0 d60.3 d23.0 d21.0 d84.4

million million million5 million million6

The tax returns thus do not give any real estimate of property income in agriculture. Also business and professional incomes exclude some incomes under d100 which were earlier exempt from tax, and this exemption limit was lifted to d150 in 1853. For those with incomes of d100 or less in business or a profession, the majority would likely actually be wage income (in the 1860s the annual earnings of a building craftsman would be d66). Thus, the tax reports of business and professional incomes include wage income for small proprietors and professionals. Here I take as property income all business and professional income of d150 or greater in 1842–1869. It is assumed that the exclusions of the property incomes with those with gross

68

GREGORY CLARK

incomes under d150 cancel out the inclusion of wage income for those with income of d150 or more. The proportions of each class in 1855–1856 in England among reported incomes were:7 Less than d100 d100–150 More than d150

d5.9 million d10.6 million d49.4 million7

But there is reason to believe that there may be underreporting of the ‘‘less than d100’’ income group who were not liable for the income tax. The tax returns report the gross income from farmland and housing and other buildings. For business incomes the 1842 Tax Acts allowed deductions for sums expended ‘‘for the repairs of premises and the supply or repair of alterations of any implements’’ (Stamp, 1922, p. 178). Thus, business and professional income was effectively assessed as net income.

5.1. Farmland Rental Income Land rents are estimated from the market rental values, including tithes and land taxes that fell on occupiers, of plots of unchanging area over the years 1209–1869. The rent paid to the owner of land was only one claim on the site value of the land. In addition, there was the tithe due originally to the church, but later to private owners of tithe rights. This was nominally 10 percent of the gross output of the land but was later collected at typically much lower rates. Also increasingly from 1600 onwards there were local parish levies to support the poor, and pay for the roads and other services. By the nineteenth century these were 6–10 percent of the rents paid by occupiers. The rent series used here gives the rental value of farmland inclusive of the tithe, but not including taxes paid by occupiers which are enumerated separately. The details of how this was constructed for the later years are discussed in Clark (2002a). This series thus includes imputed rents in the case where owners were also occupiers. Much land was bundled with dwellings, and the land rents measured here thus include payments for farmhouses and farm buildings. To avoid problems of land quality and varying land measures, the series is constructed by looking at what happens to the same plot over time, except in the medieval period where the less rigorous measure of the same type of land in the same village is used. The rent series thus incorporates and values in earlier years’ communal ‘‘waste’’ land only later brought into private


69

cultivation. It is assumed throughout that there were 28.24 million acres of farmland in England, though in earlier years some of this would be uncultivated waste. From 1842 onwards these rental values are estimated from income and property tax returns (Stamp, 1922, p. 49). Tables 7 and A1 show the value for England of these imputed land rents. It is also assumed that land rents represented rents net of repairs to fences, ditches, and buildings, which are assumed to be made by the tenants. 5.2. Returns to Capital For England evidence on interest rates goes back to about 1170 (Clark, 1988, 1996, 1998a, 1998b). Fig. 5 shows the rate of return on two very lowrisk investments in England from 1170 to 1900. The first is the gross return on investments in agricultural land, R/P, where R is the rental and P the price of land. This can differ from the real return on land: r¼

R þ ðpL pÞ P

Fig. 5. The Return on Land and on Rent Charges, 1170–2003 (by Decade). Note: For the years before 1350 the land returns are the moving average of 3 decades because in these early years this measure is very noisy.

70

GREGORY CLARK

where pL is the rate of increase of land prices and p the general rate of inflation. (pLp) is the rate of increase of real land values. But the rate of increase in real land values in the long run has to be low in all societies, and certainly was low in pre-industrial England. If the rate of increase of real land prices was as high as 1 percent per year from 1300 to 1800, for example, it would increase the real value of land by 144 times over this period. Thus, the rent/price ratio of land will generally give a good approximation to the real interest rate in the long run. The second rate of return is that for ‘‘rent charges.’’ Rent charges were perpetual fixed nominal obligations secured by land or houses. The ratio of the sum paid per year to the price of such a rent charge gives the interest rate for another very low-risk asset, since the charge was typically much less than the rental value of the land or house. The major risk in buying a rent charge would be that since it is an obligation fixed in nominal terms, if there is inflation the buyer gets a lower real rate of return. Again the gross rate of return shown is R/P, where R is the annual payment and P the price of rent charge. The real rate of return, r, in this case is: r¼

R p P

Table 8 shows the assumed risk-free return on capital by decade 1200–1869, taken as the average of these two rates.

5.3. Farm Working Capital Income There are various estimates of the value of the capital supplied by the tenant per acre of land in England in the nineteenth century, with general agreement on the rough magnitudes involved. The most detailed, by Charles Wratislaw in 1861, and the one I use as a benchmark, suggests that the tenant needed to supply on average d8.68 per acre. Other estimates from 1838 and 1878 suggest, respectively, d10 and 12 per acre.8 Wratislaw omits any allowance for the cost of the maintenance of the farmer over the course of the year. Assuming the farmer expends d100 on himself, Wratislaw’s capital per acre would be d9.2. This would be composed as follows: Live stock Implements Seed, labor, horse, and cattle food Rent, tithe, and taxes in advance Maintenance of farmer

60 percent 11 percent 21 percent 3 percent 5 percent

71


Table 8. Decade

1200 1210 1220 1230 1240 1250 1260 1270 1280 1290 1300 1310 1320 1330 1340 1350 1360 1370 1380 1390 1400 1410 1420 1430 1440 1450 1460 1470 1480 1490 1500 1510 1520 1530 1540 1550 1560 1570 1580 1590 1600 1610

Farm and other Property Incomes.

Farm Capital All Farm Income Income (Million d) (Million d)

0.89 0.96 0.95 0.93 1.01 1.02 1.09 1.06 1.05 1.14 0.99 1.07 1.56 1.22 0.88 1.16 1.21 0.89 0.84 0.79 0.85 0.82 0.78 0.84 0.77 0.79 0.77 0.73 0.75 0.72 0.76 0.80 0.89 0.98 1.08 1.35 1.65 2.11 2.55 2.85 4.08 4.56

3.98 4.04 4.19 3.84 4.55 4.19 4.48 4.44 4.58 4.96 4.89 5.45 6.00 5.52 4.88 6.12 6.15 5.96 5.62 5.27 5.69 5.44 5.23 5.60 5.13 5.27 5.12 4.83 5.02 4.77 5.05 5.33 5.95 6.51 7.19 8.96 10.71 13.24 15.58 17.45 24.57 27.79

House Mines, etc., Rents Estimated (Million d) Capital Income 0.09 0.10 0.13 0.14 0.15 0.15 0.44 0.48 0.35 0.32 0.28 0.35 0.30 0.25 0.23 0.09 0.08 0.07 0.06 0.06 0.08 0.08 0.08 0.07 0.04 0.04 0.05 0.06 0.06 0.06 0.07 0.09 0.10 0.11 0.12 0.15 0.17 0.29 0.41 0.47 0.87 1.36

– – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – –

‘‘Other’’ Net Property Capital Incomes Income (Million d) (Million d) 0.52 0.62 0.65 0.72 0.70 0.83 1.05 1.00 0.92 0.98 0.88 0.98 1.23 0.93 0.62 0.69 0.74 0.55 0.49 0.48 0.48 0.47 0.47 0.49 0.46 0.46 0.46 0.47 0.47 0.45 0.49 0.56 0.60 0.64 0.68 0.97 1.15 1.38 1.56 1.86 2.32 2.75

3.10 3.27 3.38 3.13 3.75 3.43 4.00 3.78 3.59 3.85 3.60 4.01 4.90 4.27 3.49 3.48 3.83 3.14 3.11 2.86 3.12 2.96 2.80 2.98 2.77 2.80 2.80 2.43 2.63 2.43 2.64 2.72 3.19 3.81 3.93 4.45 5.24 6.93 9.41 9.59 16.40 18.94

72

GREGORY CLARK


1620 1630 1640 1650 1660 1670 1680 1690 1700 1710 1720 1730 1740 1750 1760 1770 1780 1790 1800 1810 1820 1830 1840 1850 1860

Farm Capital All Farm Income Income (Million d) (Million d)

5.05 5.34 5.71 6.09 6.25 5.69 5.64 5.19 5.02 5.49 5.21 5.14 5.16 5.66 6.12 6.66 7.17 9.11 13.47 17.34 15.63 16.71 16.68 16.94 17.79

29.82 32.80 35.51 38.25 39.98 36.13 36.43 34.63 34.63 36.76 37.00 36.96 36.28 40.54 43.53 48.85 52.53 66.41 92.83 119.41 110.69 113.01 121.02 124.03 127.87

House Mines, etc., Rents Estimated (Million d) Capital Income 1.45 1.61 1.50 1.63 1.99 1.94 2.21 1.72 2.37 2.21 2.85 2.75 2.73 3.78 4.17 4.41 4.05 6.80 10.28 15.27 18.53 23.03 23.60 28.95 40.17

– – – – – – – – – – – 0.86 0.83 0.98 1.27 1.44 1.70 2.16 4.07 5.75 5.28 6.09 9.99 16.39 26.21

‘‘Other’’ Net Property Capital Incomes Income (Million d) (Million d) 3.14 3.36 3.74 4.24 4.41 4.51 4.60 4.73 5.10 5.84 6.05 5.47 6.06 6.67 7.80 8.51 9.72 13.01 21.99 30.93 33.05 40.69 44.10 54.11 75.50

20.10 21.81 23.33 24.65 26.15 24.19 25.27 24.21 25.04 27.50 29.14 28.63 28.63 34.28 37.43 42.08 44.28 59.27 89.95 120.13 117.73 129.63 140.13 162.92 211.75

Allowing the farmer just the return on capital from bonds or mortgages, this would imply a capital cost in the 1860s of d0.39 per acre. However, as with all business enterprises, there has to be an additional return based on the risk of the enterprise. Farming was not a high-return activity, so I set this additional return at 3 percent. This raises the working capital return per acre to d0.69. The land rent actually includes a substantial amount that is a return to capital in the form of buildings and land improvement. Conventionally the farmer’s profit was expected to be half the rental of the land before 1896 (Stamp, 1922, p. 82), though this return included the farmer’s wage which I have included elsewhere. With a land rental in the 1860s of d46 million, that


73

would imply a profit income of d23 million. The net profit income on working capital calculated here for the 1860s is d19 million. To estimate the equivalent capital costs for the other decades, I make the following assumptions. First, the interest cost of the capital employed by farmers was the average of the return on rent charges and land, plus 3 percent for risk. Second, the price of capital goods was the same as the price of farm output. Since live stock, seeds, and animal food were the majority of the capital stock, and implements were a small share, this assumption seems reasonably innocuous. Lastly I assume that the capital–output ratio for the farmer’s capital did not change over time. This last assumption is the most contentious. But again when we consider the importance of animals, fodder, and seeds in farmer’s capital, it does not seem that there was any reason to expect any change in the capital–output ratio over time. With these assumptions I get the implied payments for working capital in agriculture shown in Table 8. In Table 8 the payments to capital in year t are calculated, using these assumptions, as: Capital rentalt ¼

wagest þ land rentst þ taxest ðr0 =rt Þ½ðoutput value0 =capital rental0 Þ 1

where r0 is the return on capital in the 1860s and rt the return on capital in year t.9 Now that we have estimates of farm wages, land rents, poor rates paid by land occupiers, and farm working capital returns, I can also estimate the total output of the agricultural sector. This is shown in Table 8.

5.4. Rental Value of Housing and other Structures To estimate this I start with a measure of the average rental value of a ‘‘dwelling’’ in England. This measure is calculated separately for London and the rest of England because of the much greater value of property in London throughout these years. This requires estimating the share of the population in London and the rest of England throughout these years. ‘‘Dwelling’’ is set in quotes because dwellings were intermixed with shops, pubs, malthouses, barns, etc., in all these years. To get total rentals of all non-farm structures in the economy, I calculate from the property tax returns the total value of these rentals over the years 1842–1869 (Stamp, 1922, p. 50). I project this back before 1842 by estimating the numbers of non-farm dwellings in each decade, and the

74

GREGORY CLARK

average rental value of dwellings over time from the 1260s to the 1860s. The method used here is described in Clark (2002b). There are 14,261 observations on the prices or rent of dwellings in London and elsewhere for the years 1265–1869, and 4,272 from before the year 1800. But there are relatively few observations for the years before 1500, only 659, so that the index is noisy in the earliest years. To get the total implied rental value of dwellings, I need an estimate of the number of non-farm dwellings. In 1801 and later the censuses give the average number of people per dwelling. For earlier years I assume the average number of people per dwelling is the same as in 1801, 5.44. As noted when discussing land rents, farmhouses and farm buildings are already accounted for under the rental value of farmland. I assume throughout all these years that there were 250,000 farmhouses whose rent was already accounted for under land rents throughout the years 1209–1869, based on the number of farmers reported for England in the 1860s. This implies that in the decade where England’s population was estimated to be at its lowest, the 1440s with only 2.27 million people, aside from an assumed 250,000 farmhouses there were only 168,000 other dwellings. The implied rental income reported here is a gross rental. Thus, to get the net rental income, we need to deduct repairs. Clark (1998a) calculates the return on land and housing in England by quinquennia for the eighteenth and nineteenth centuries. The returns on housing average about 2 percent more than those on land, suggesting that the depreciation rate for housing is about 2 percent. To get the net rental of housing, I deduct for each period a fraction of the estimated rental which is: 2 rþ2 where r is the return on land (in percent). The total net rental is also shown in Table 8. 5.5. Other Property Incomes The property tax returns for the years 1842 and later distinguish property incomes from quarries, coal mines, canals, railways, and iron works (Stamp, 1922, p. 220). These reported sums are for Great Britain for the years 1842–1853, and the United Kingdom thereafter. To convert them to an English basis, I assume, unless otherwise indicated, that England was 84 percent of British economic activity and that after 1853 there was no income

75


in Ireland from quarries, coal mines, iron works, canals, railways, and gasworks (there would be some income but it is assumed to be negligibly small). These returns do not include, however, the imputed income from capital invested in roadways. These property incomes are in each case carried back to 1730. In addition, I estimate the property income from ship ownership from 1730 to 1869. 5.5.1. Coal Mining and Quarrying For 1842–1869 I calculate the share of coal production from England from the share in the years 1854–1869 given in Mitchell and Deane (1962, p. 115). For 1842–1853 that is taken as 75 percent, the same share as 1854–1855. I assume quarries had the same output distribution as coal mines. To estimate returns from coal mines earlier than 1842, I use the data from Clark and Jacks (2007), which estimate by decade from the 1730s coal output, pithead prices, coal ground rents, and the share of capital returns in output prices, which through the 1730s–1860s averaged 20 percent. Table 9 shows these estimates. Coal mineral rents are calculated directly by decade. Capital returns are calculated as 20 percent of the total production cost throughout, based on estimates from colliery accounts in the years 1720–1860. For all property income, the last column includes a 12 percent Table 9. Decade

1730 1740 1750 1760 1770 1780 1790 1800 1810 1820 1830 1840 1850 1860 a

Property Income from Mining and Quarrying, 1730s–1860s.

Coal Output (Million Tons)

Coal Output Value (Million d)

Coal Rents (Million d)

Capital Returns (Million d)

All Property Income (Million d)a

3.42 3.83 4.30 5.50 7.05 8.20 9.54 11.10 16.67 19.50 22.82 32.58 46.62 62.82

0.42 0.45 0.56 0.85 1.11 1.26 1.60 2.70 4.51 5.17 5.44 6.98 12.33 17.59

0.039 0.062 0.071 0.122 0.124 0.136 0.170 0.267 0.355 0.451 0.503 0.711 0.985 1.299

0.084 0.090 0.112 0.169 0.222 0.253 0.319 0.540 0.901 1.035 1.088 1.395 2.466 3.518

0.14 0.17 0.21 0.33 0.39 0.44 0.55 0.90 1.41 1.66 1.78 2.36 3.87 5.39

Including an addition of 12 percent for quarries.

76

GREGORY CLARK

addition to account for quarries as well as coal mines, based on the post1842 tax returns (Stamp, 1922, p. 220). 5.5.2. Railways Table 10 shows the estimated property income from railways in the 1840s–1860s from the property tax returns. This income before 1853 is given for Britain, and is reduced to an English basis by multiplying by 0.84 (Stamp, 1922, p. 220). After 1853 the report is for the entire United Kingdom. The share attributed to Britain is calculated using the relative paid-up capital of British and Irish railways (Mitchell & Deane, 1962, pp. 225–228), and is then reduced to an English basis by being multiplied by 0.84. Before 1842 property income from railways is estimated on the basis of the miles of line completed in the United Kingdom relative to 1842 (Mitchell & Deane, 1962, p. 225), multiplied by the same property earnings as in 1842. 5.5.3. Canals Canal and improved river mileage in England from 1750 to 1850 is from Ginarlis and Pollard (1988, Table 8.7, pp. 213–215). For 1730–1749 I assume the same mileage as in 1750. Property income per canal or Table 10. Decade

1730 1740 1750 1760 1770 1780 1790 1800 1810 1820 1830 1840 1850 1860

Property Income from Railways, Canals, and Iron Works, 1750s–1860s.

Railway Property Income Canal Miles Property Iron Output Miles from Railways Income from (Million (Million d) Canals Tons) (Million d) 0 0 0 0 0 0 0 0 0 50 338 2,530 – –

0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.06 0.42 3.33 8.27 14.66

Note: ( ) indicates tentative estimate.

(947) (947) 947 1,028 1,486 1,788 2,149 2,849 3,091 3,258 3,412 3,495 – –

0.15 0.14 0.15 0.17 0.26 0.34 0.47 0.79 1.03 0.95 0.98 1.06 0.80 0.74

20 20 20 20 20 49 75 150 200 249 480 641 – –

Property Income from Iron (Million d) 0.018 0.018 0.018 0.018 0.018 0.023 0.061 0.135 0.208 0.180 0.270 0.342 0.621 0.961


77

waterway mile is assumed to be equal to that of the 1840s, but converted into current terms by the cost of a day of farm labor. These returns are shown also in Table 10. 5.5.4. Iron Works To estimate the profit earnings from iron works over the years 1842–1869, I assume no iron was produced in Ireland, and the earnings in England were the total British earnings multiplied by the share of pig iron produced in England in 1842–1869 (Mitchell & Deane, 1962, p. 131). To project this back as far as the 1720s, I use estimates of the tons of English pig iron output per year at benchmark dates (Mitchell & Deane, 1962, p. 131), multiplied by the price of manufactured iron to get an estimate of the value of output earlier. It is assumed capital returns are the same fraction of the value of output in the 1842 as in each earlier decade. These returns are also shown by decade in Table 10. 5.5.5. Roads Another form of capital was the road system. This generally did not produce an explicit rental income, except where roads were turnpiked and maintained from toll revenue, mainly in the period 1750–1840. In other periods roads were paid for by levies on parish occupants, or by local and county rates on property. The rate payments are calculated below. In earlier years when there were direct labor levies for work on the parish roads, these will be counted in the calculated labor income in the economy. But if we want to count all sources of income, then we need to include the implicit turnpike property income, from the capital invested in turnpike roads and paid for through toll collections, for the years 1696–1869. To do this I first calculate (roughly) the average miles of turnpike road in England in each decade (Pawson, 1977, pp. 155–156; Bogart, 2005, p. 440). I calculate the average capital invested in a turnpike road from Bogart (2005, p. 454): this shows road expenditure per mile in 1819 prices for the 10 years after establishing a new turnpike. The investment in the first year is d260, the second d170, the third d100, fourth d90, and fifth d80. But thereafter there is a steady-state expenditure of d75 per year. I presume the new investment is all of the first year investment, plus all sums thereafter above the presumed maintenance cost of d75 per mile. This gives d400 per mile. I convert this cost into the prices of each decade using the level of farm wages (since labor was the major component of this investment). I assume that the return on this capital throughout these years was 5 percent. These estimates are shown in Table 11.

78

GREGORY CLARK

Table 11. Implied Property Income from Roads, 1720s–1860s. Decade

1700 1710 1720 1730 1740 1750 1760 1770 1780 1790 1800 1810 1820 1830 1840 1850 1860

Turnpike Miles

Total Construction Costs (Million d)

Implied Turnpike Property Income (Million d)

0 495 1,215 2,119 2,844 6,281 11,260 12,216 13,172 13,795 16,020 16,910 17,355 17,835 17,835 17,835 17,835

0.00 0.09 0.21 0.39 0.52 1.18 2.25 2.60 2.99 3.66 5.38 6.76 6.11 6.19 6.51 6.82 7.29

0.00 0.00 0.01 0.02 0.03 0.06 0.11 0.13 0.15 0.18 0.27 0.34 0.31 0.31 0.33 0.34 0.36

5.5.6. Ships The volume of shipping services used by the economy expanded dramatically in the Industrial Revolution era, as England became dependent on imported food and raw materials, and as the cities relied on coal from the northern coal fields as their primary energy source. There are statistics on the net tonnage of ships registered in the United Kingdom, 1788–1869 (Mitchell & Deane, 1962, pp. 217–218). Most of this is accounted for by sailing ships even up to 1869 (when 83 percent of tonnage was still in sailpowered vessels). To get an estimate of what fraction of these ships were operating from English ports, I rely on the data in Davis (1956) on the numbers of sailors paying the ‘‘sailor’s sixpence’’ tax in England and the United Kingdom between 1707 and 1828. This allows me to divide up the tonnage from 1788 and later between England and the rest of the United Kingdom (the English share is typically 85–90 percent of the UK share). For the years 1707–1787, I estimate the English tonnage from the number of sailors alone assuming it had the same ratio to sailors as in 1788–1897. This gives the data reported in column 2 of Table 12 on the total tonnage of English shipping. Estimating the value of that tonnage is difficult. There are various piecemeal estimates of the cost of a ship, fully rigged and outfitted, for the years 1670–1858.10 These give a rough estimate of the cost of a new

79


Table 12. Decade

1700 1710 1720 1730 1740 1750 1760 1770 1780 1790 1800 1810 1820 1830 1840 1850 1860

Property Income from Shipping.

Ship Stock Assumed Cost (Million Tons) per Ton (d)

0.572 0.747 0.784 0.897 0.785 0.920 1.111 1.134 1.279 1.260 1.775 2.070 1.908 1.930 2.593 3.466 4.483

11.0 11.0 11.0 10.1 10.0 10.0 9.7 10.0 10.4 12.6 18.2 21.8 18.3 17.8 17.6 16.5 17.6

Ship Stock (Million d)

Feinstein (1988) Implied Ship Stock (Million d)

Ship Rental Income (Million d)

4.19 5.48 5.75 6.04 5.23 6.14 7.18 7.56 8.82 10.62 21.65 30.10 23.28 22.84 30.39 37.97 52.80

– – – – – – 2.5 2.9 3.8 6.7 13.9 15.1 11.3 14.7 17.6 27.7 38.0

0.61 0.82 0.83 0.86 0.76 0.88 1.03 1.07 1.24 1.50 3.18 4.41 3.35 3.39 4.31 5.35 7.53

ship for various benchmark dates, which I interpolate using a very rough cost index (with a 0.67 weight for wages, and 0.33 for timber). The overall cost index moves in line with the input price index. I assume in calculating the value of the shipping stock that the average working ship had a value two-thirds that of a new ship (there were substantial losses of ships each year from accidents and loss in war). Feinstein (1988, p. 439) gives decadal estimates of the net stock of ships in Great Britain from 1760 to 1850, and from 1851 to 1869 annual estimates of the UK ship stock, though he does not indicate the source for the pre-1851 data. The implied value of the English shipping stock based on his measures is also shown in Table 12. His numbers are much smaller, one reason being that he assumes the value of the net stock is about half that of the gross stock. Davis (1957) and others have investigated the return on ship ownership. Davis’s conclusion is that ships earned a net return substantially in excess of the risk-free return on capital, because of considerable uncertainties on the profitability of voyages due to the hazards of captains, trade, war, and the weather. Losses of shipping could be insured against, but not losses of income from the failures of ventures. I thus assume that the profit rate on this capital was 5 percent beyond that of the return on land or rent changes.

80

GREGORY CLARK

This is in line with the estimate of Barney (1999, p. 137) that the King’s Lynn firm of W. & T. Bragge earned an average net profit of 9 percent in the 30 years, 1766–1795. Column 6 of Table 12 shows the implied rental on ships from the 1700s to the 1860s. The final column of Table 12 shows the sum of all the nonstructure capital incomes – coal mines, quarries, railways, canals, roads, iron works, gasworks, and ships. 5.5.7. Other Capital Income There are other sources of capital income for which it is harder to derive direct evidence. These include income from the machinery and working capital in manufacturing and trade, and capital in road and river transport – horses, wagons, carriage, and harbors. These were incomes captured in the year 1842 by Schedule D of the Property and Income Tax, aside from the property incomes I have already accounted for. For the 1860s, the total income attributable under Schedule D to England and Wales averaged d105.4 million, of which d26.2 million has been accounted for by the sectors discussed above (Stamp, 1922, pp. 218, 504, 509). The remainder of capital income in the non-agricultural economy, d74.1 in the 1860s for England alone, is projected back to 1209 using the formula as in agriculture: Net capital rentalt ¼

wagest þ landrentst þ taxest ðr0 =rt ÞðPK0 =PKt Þðoutput value0 =net capital rental0 Þ 1

Except that now with the term (PK0/PKt) I allow in principle for a variation of the price of capital goods relative to the price of output. For this sector of manufacturing and trade, I assume that the interest cost is the risk-free interest rate plus a risk premium of 5 percent (see, for example, Harley, 2001, pp. 21–28). This produces the numbers in Table 8.

6. NET NOMINAL NATIONAL INCOME I can now calculate the sum of all implied property incomes in the years 1209–1869, and, adding this to wages, the net nominal national income. These estimates by decade are shown in Table 13. Nominal national income for England in the 1860s averages on this calculation d635 million. Feinstein (1972) estimates domestic factor incomes plus indirect taxes in the UK economy as averaging d840 million a year in the 1860s, gross of depreciation. Deducting depreciation, and calculating England’s share of

81


Table 13. Decade Net National Income (Million d)

1200 1210 1220 1230 1240 1250 1260 1270 1280 1290 1300 1310 1320 1330 1340 1350 1360 1370 1380 1390 1400 1410 1420 1430 1440 1450 1460 1470 1480 1490 1500 1510 1520 1530 1540 1550 1560 1570 1580 1590 1600

5.94 5.52 6.65 5.32 7.23 7.05 8.12 7.97 7.97 8.57 8.53 9.63 10.00 9.07 7.73 9.16 9.28 9.03 8.32 7.93 8.35 8.06 7.84 8.30 7.66 7.80 7.64 7.42 7.60 7.27 7.76 8.41 9.26 10.04 10.95 14.33 16.87 20.45 23.54 26.93 36.18

National Income and Factor Shares.

Wage Share in National Income

Land Share in National Income

Capital Share in National Income

Farm share in Income

Output per worker (Agricultural vs. Non-agricultural)

0.478 0.472 0.492 0.509 0.481 0.513 0.506 0.526 0.549 0.551 0.578 0.583 0.510 0.529 0.549 0.620 0.587 0.652 0.626 0.639 0.626 0.632 0.642 0.641 0.638 0.640 0.634 0.672 0.653 0.664 0.660 0.677 0.655 0.619 0.641 0.689 0.688 0.660 0.599 0.642 0.545

0.270 0.260 0.245 0.208 0.262 0.203 0.175 0.155 0.160 0.164 0.170 0.168 0.181 0.206 0.227 0.169 0.194 0.180 0.206 0.193 0.204 0.197 0.188 0.191 0.196 0.194 0.200 0.159 0.179 0.168 0.170 0.151 0.173 0.209 0.188 0.139 0.135 0.154 0.208 0.164 0.253

0.252 0.268 0.263 0.282 0.257 0.284 0.319 0.319 0.290 0.285 0.252 0.249 0.309 0.265 0.224 0.211 0.219 0.168 0.167 0.168 0.169 0.170 0.170 0.168 0.166 0.166 0.167 0.168 0.168 0.169 0.170 0.172 0.172 0.172 0.171 0.172 0.177 0.186 0.193 0.194 0.202

0.65 0.62 0.60 0.57 0.60 0.56 0.54 0.53 0.53 0.53 0.53 0.53 0.58 0.59 0.61 0.66 0.66 0.66 0.67 0.66 0.68 0.67 0.67 0.67 0.67 0.68 0.67 0.65 0.66 0.65 0.65 0.63 0.64 0.65 0.65 0.65 0.62 0.60 0.57 0.60 0.56

1.26 1.11 1.01 0.88 1.01 0.86 0.79 0.75 0.75 0.76 0.75 0.76 0.94 0.97 1.05 1.30 1.27 1.27 1.38 1.29 1.41 1.37 1.33 1.38 1.36 1.40 1.36 1.24 1.31 1.28 1.25 1.14 1.19 1.23 1.27 1.26 1.11 1.01 0.88 1.01 0.86

82

GREGORY CLARK

Table 13. (Continued ) Decade Net National Income (Million d)

1610 1620 1630 1640 1650 1660 1670 1680 1690 1700 1710 1720 1730 1740 1750 1760 1770 1780 1790 1800 1810 1820 1830 1840 1850 1860

41.75 45.26 49.96 54.87 60.38 63.44 59.96 61.15 60.79 63.83 69.23 72.63 73.78 76.25 85.30 95.62 107.29 119.29 155.29 234.85 318.83 330.05 371.03 420.94 495.30 636.61

Wage Share in National Income

Land Share in National Income

Capital Share in National Income

Farm share in Income

Output per worker (Agricultural vs. Non-agricultural)

0.543 0.553 0.560 0.572 0.588 0.584 0.590 0.580 0.591 0.593 0.585 0.579 0.593 0.607 0.578 0.584 0.584 0.603 0.590 0.577 0.586 0.609 0.624 0.645 0.651 0.651

0.248 0.232 0.231 0.227 0.211 0.214 0.203 0.211 0.210 0.201 0.207 0.213 0.201 0.185 0.207 0.195 0.202 0.186 0.184 0.177 0.165 0.140 0.116 0.108 0.092 0.077

0.209 0.215 0.209 0.202 0.201 0.202 0.207 0.209 0.199 0.206 0.208 0.208 0.206 0.208 0.216 0.221 0.215 0.211 0.226 0.246 0.250 0.251 0.260 0.247 0.257 0.272

0.54 0.53 0.53 0.53 0.53 0.53 0.58 0.59 0.61 0.66 0.66 0.66 0.67 0.66 0.68 0.67 0.67 0.67 0.67 0.68 0.67 0.65 0.66 0.65 0.65 0.63

0.79 0.75 0.75 0.76 0.75 0.76 0.94 0.97 1.05 1.30 1.27 1.27 1.38 1.29 1.41 1.37 1.33 1.38 1.36 1.40 1.36 1.24 1.31 1.28 1.25 1.14

UK income by assuming that incomes per person in Wales and Scotland were the same as in England, and incomes in Ireland were half those of England, implies that English net national income in the 1860s was d581 million (Feinstein, 1972, Tables T1 and T120). So the estimate here for the 1860s exceeds Feinstein’s by 9 percent. The difference stems mainly from the larger estimated wage income above. This estimate is also close to the estimate in Deane and Cole (1967, p. 166), that British domestic income (gross of some capital depreciation) was d648 million in 1861 and d877 million in 1871. This implies an English domestic income in the 1860s of d599 million.


83

Table 13 also shows the share of wages, land rents, and capital in national income, where that share of wages is calculated as: Wage income NNI indirect taxes The above calculation suggests the share of wages varied between 48 and 76 percent of national income, reaching the highest share, 76 percent, in the 1860s. The share of land rents is: Land rents NNI indirect taxes The above share before 1800 varied between 15 and 30 percent of national income, but by the 1860s had fallen to 9 percent. The share of capital is calculated as the residual. This method assumes that the burden of indirect taxes was born equally by labor, land, and capital owners. Table 13 also shows the share of income that comes from the agricultural sector over time. Since we assumed a certain allocation of labor between farming and the rest of the economy, I can also estimate the implied relative value of output per worker in agriculture compared to the rest of the economy. That ratio is shown as the last column of Table 13.

6.1. Prices P is an index of the price of output. Here a crucial decision must be made. There are two potential output price indices: PNDP: price of net domestic output; PDE: price of domestic expenditures (including net investment). The two are related in general through the following formula: lnðPNDP Þ ¼

1þy fð1 þ yÞ lnðPDE Þ þ lnðPX Þ y lnðPM Þ 1þf 1þf

(2)

where PM is the price of imports (wholesale prices), PX the price of exports (free on board), y the ratio of the value of imports to NDI, and f the ratio of the value of imports to domestic expenditure. For simplicity I assume throughout that y ¼ f, or that commodity trade is balanced.11 In this case

84

GREGORY CLARK

the above formula becomes much simpler, given as follows: PNDP ¼ PDE

PX PM

y (3)

Should P be the price of the output produced with domestic factors of production, PNDP? Or should P be the price of the goods purchased by English consumers and producers, PDE? Before 1700 this makes little difference, since y is small and the ratio PX/PM deviates little on average from 1. But after 1760, y quickly became much larger. By the 1860s for England it would be 0.25 or more. And the ratio PX/PM declined substantially in the years 1790–1860. As Fig. 6 shows, in the Industrial Revolution era export prices, principally those of textiles and iron and steel, declined relative to import prices. Thus, the productivity of the economy, measured in terms of NDP, was increasing more than the productivity, measured as consumption per unit of input, in decades from the 1790s to the 1860s. The argument for using the NDP measure of prices to measure both output and the efficiency of the economy is that such a measure indexes the output and efficiency of production in England, and this is the core concept that we seek. The argument against this, as we shall see, is that if there is differential productivity advance across industries, then quite accidental factors concerning the location of industrial activity can result in dramatic differences in the measured output and productivity advance across regions,

Export Prices/Import Prices

2.5

2.0

1.5

1.0

0.5

0.0 1690

1730

1770

1810

1850

Fig. 6. Export Relative to Import Prices, England, 1690–1869.

85


differences that in no way measure differences in the inventiveness or dynamism of these economies. The alternative measure of efficiency, using the expenditure price index, indicates better the success of the economy at improving human welfare. But it suffers from the deficiency that events elsewhere in the world, completely unrelated to this economy, will influence this measure. In the case of England in the Industrial Revolution, for example, improvements in the cultivation of cotton in the US South resulted in lower input prices for cotton goods consumed in England, and hence a higher measured efficiency. The price indices are calculated as geometric indices, that is: Y Pai i PM ¼ i

where ai is the shares in import costs of each good, and so sum to 1. Also: Y b Pi i PDE ¼ i

where bi is the shares in domestic purchases of each good, and again sum to 1. With this specification the GDP price index will be of the following form: Y c Pi i PNDP ¼ i

P

where i ci ¼ 1, but the individual weights can be positive or negative. Negative weights will correspond to imported commodities. The domestic expenditure price index is formed from 11 principal component indexes, whose weights for each period are shown in Table 14. But each of these component indices in turn is composed of a weighted average of the price of various commodities. The individual price series were derived as the estimated parameters on year indicators of regressions of the following form: X X bk DTYPEk þ ft Dt þ ikt lnðPit Þ ¼ k

t

where DTYPE is a set of indicator variables for each type of a product, where a type was defined by location, purchaser, characteristics, and measuring unit. In this I try and control for variations in the size of units across sources, and in the quality of the product. This is important because both the quality of the product and the size of the measures varied across sources, even for very homogenous commodities in the same place at the same time. In London in 1827, for example, the Clothworkers Guild paid

86

GREGORY CLARK

Table 14.

The Weights in the Expenditure Price Index.

Commodities 1820–1869 1741–1819 1691–1740 1375–1690 1280–1375 1247–1279 1209–1246 Food Fuel Lodging Light Soap Clothing Services Tobacco Books Manufactures Investment

0.450 0.050 0.100 0.040 0.005 0.110 0.095 0.010 0.020 0.040 0.080

0.540 0.050 0.050 0.040 0.005 0.120 0.080 0.010 0.015 0.030 0.060

0.600 0.050 0.050 0.040 0.005 0.120 0.070 – 0.010 0.020 0.035

0.650 0.050 0.050 0.045 0.005 0.120 0.060 – – 0.020 –

0.635 0.050 0.080 0.040 0.005 0.120 0.050 – – 0.020 –

0.680 0.050 – 0.050 – 0.150 0.050 – – 0.020 –

0.750 – – – – 0.200 0.050 – – – –

20 d. per gallon for milk, Bethlehem insane asylum 13 d., and the King’s Household 24 d., a range in price for a seemingly standard product of nearly 2:1. In earlier years where observations are missing for some years, they were interpolated as an 11-year-centered moving average of the years with prices, where this was possible.

7. EXPENDITURE PRICES The weights of the sub-categories in this price index change over time to reflect two things: first, the changing structure of expenditure as the economy became richer in the years 1820–1869 following the Industrial Revolution; second, the decline in the number of available price series as we move to the earliest years of the thirteenth century. 7.1. Food Index This is the most important of the sub-indices, with a weight of at least 0.45 in the overall expenditure index. This index consists of the weighted average of a number of sub-indices: starches, meat, dairy, fish, fats, drink, sugars, salt, and spices. The weights were as in Table 15. 7.1.1. Starches The component series are wheat bread, barley, oats/oatmeal, peas, potatoes, and rice. The relative weights of each in the starch index are shown in

87


The Weights in the National Food Price Index.

Table 15. Commodities

1820–1869

1650–1819

1275–1649

1209–1274

0.45 0.14 0.14 0.03 0.03 0.12 0.07 0.01 0.01

0.52 0.12 0.12 0.03 0.03 0.12 0.04 0.01 0.01

0.55 0.12 0.12 0.03 0.03 0.12 0.01 0.01 0.01

0.55 0.13 0.13 0.03 0.03 0.12 – 0.01 –

Starches Meat Dairy Fats Fish Drinks Sugars Salt Spices

Table 16.

The Weights within the Starches Index.

Commodities

1820–1869

1790–1819

1760–1789

1720–1759

1660–1719

1209–1659

Wheat bread Barley Oats/oatmeal Peas Potatoes Rice

0.667 0.020 0.030 0.030 0.200 0.050

0.750 0.020 0.050 0.050 0.100 0.030

0.800 0.030 0.050 0.050 0.050 0.020

0.840 0.030 0.050 0.050 0.020 0.010

0.850 0.040 0.050 0.050 – 0.010

0.850 0.050 0.050 0.050 – –

Table 16. Up until 1869 wheat bread was the single most important item of consumption in the economy, getting a weight of at least 9 percent in the domestic expenditures index, and at least 13 percent in the workers’ cost of living (COL) index. However, rather than use bread prices directly, I approximate them based on the prices of wheat, labor (skilled craftsmen), salt, wood fuel, and candles. This is done because there is evidence that government regulation of the bread market before 1815 created changes in the quality of bread sold over time. Bread prices are thus estimated (assuming fixed coefficients) as the weighted average of wheat prices, craft labor, firewood prices, salt prices, and candle prices. The available bread prices before 1816 are mainly those for London, but these were regulated by statute before 1815. The statute stipulated how much flour was to be produced per bushel of wheat, and how many loaves produced from this flour (Webb & Webb, 1904). It also set the ‘‘allowance’’ the baker received to turn the flour into bread.

88

GREGORY CLARK

If bread prices measured bread of constant quality over time, then they should have a very close relationship to the price of wheat. This is because wheat was the overwhelming cost in making bread. A breakdown of the costs of bread baked for the Navy in 1767 suggests that the price of bread should be nearly proportional to that of wheat, since wheat constituted 92 percent of the costs of making bread (Beveridge, 1939, p. 542). Robert Allen objects that this cost share for wheat is too high, leaving out the required managerial and capital returns of the baker (Allen, 2008, p. 966). But if we calculate the share of wheat costs in bread from the details of the London assize in the years 1797–1813, then we still find wheat costs were a full 81 percent of the price of bread (Parliamentary Papers, 1803–1804, pp. 11–12, 1812–1813, pp. 3, 12). The other elements of costs in 1797–1804 were labor, 4.7 percent; fuel, 1.8 percent; yeast, 1.6 percent; salt, 1.5 percent; candles, 0.4 percent; and profits, 10 percent. If bread was of constant quality, then bread prices in other years should be predictable from these costs. Fig. 7 shows the price of bread in London relative to its cost over time: where the cost elements that I can observe are wheat, labor, fuel, salt, and candles. Yeast is assumed to have a cost proportionate to wheat. And profits are assumed always as 10 percent of total costs. The figure shows that the quality of bread cannot be constant over time. After the lifting of the bread assize in 1815, the price of bread quickly rose nearly 10 percent relative to the cost. Around 1790 bread sold for about 8 percent less than its cost of production – so then either bakers were making negative economic profits or the quantity of wheat in the standard loaf had been, in effect, reduced.12 Earlier there are other periods where prices are substantially above costs.13 In this situation it seems much safer to work up the implied bread price from its component costs than to assume that there were vast swings in the compensation of bakers, with those of the late eighteenth century subsidizing their bread sales, and those of other periods garnering substantial profits. Oatmeal prices were used in years where they were available. In other years oatmeal prices were interpolated using the price of oats. 7.1.2. Meat The component series are beef, cattle, mutton, sheep, pork, pigs, poultry, and eggs. Meat is sold by the pound in later years. Earlier to infer meat prices I have to use the prices of live animals. This will only accurately represent meat prices if animal weights did not change. Since the live animal series are used in the years 1209–1600, where there is no sign of any yield increases in arable crops, this seems a reasonable assumption. The weights are given in Table 17.

89


1.3

Price/Cost

1.2

1.1

1

0.9 1650

1675

1700

Fig. 7.


1725

1750

1775

1800

1825

1850

The Bread Price/Cost Ratio.

The Weights in the Meat and Dairy Indices. 1620–1869

1209–1619

Meat Beef/cattle Mutton/sheep Pork/pigs Poultry Eggs

0.400 0.300 0.200 – 0.100

0.400 0.300 0.100 0.100 0.100

Dairy Milk Butter Cheese

0.300 0.300 0.400

0.300 0.300 0.400

7.1.3. Dairy This series is relatively simple, with just milk, butter, and cheese, and relatively unchanging weights throughout. The weights are also shown in Table 17. 7.1.4. Fish The fish series is a weighted average of three components – herring, salt cod, and salt salmon. The weights are given in Table 18.

90

GREGORY CLARK

Table 18. Commodities Herring Salt cod Salt salmon

1831–1869

1642–1830

1584–1641

1415–1583

1209–1414

1.00 – –

– 1.00 –

0.50 0.50 –

0.50 0.25 0.25

1.00 – –

Table 19. Commodities Cider Beer Wine Tea Coffee

Sugar Treacle Honey Currents/raisins

The Weights in the Drinks Price Index.

1820–1869

1780–1819

1760–1779

1704–1759

1486–1703

1209–1485

– 0.50 0.15 0.30 0.05

– 0.60 0.15 0.20 0.05

– 0.70 0.15 0.10 0.05

– 0.80 0.10 0.10 –

– 0.75 0.25 – –

0.25 0.50 0.25 – –


The Weights in the Fish Price Index.

The Weights in the Sugar Price Index.

1770–1869

1539–1769

1480–1538

1275–1479

0.65 0.10 – 0.25

0.50 – – 0.50

0.333 – 0.333 0.333

0.10 – 0.60 0.30

7.1.5. Drink This series incorporates cider, beer, wine, tea, and coffee. Here the weights change greatly over time as is shown in Table 19. Over time the favored drinks of the population changed greatly, in part as a result of substantial changes in the relative prices of the different beverages. 7.1.6. Sugars: Honey, Currents/Raisins, Sugar, and Treacle Currents and raisins were mainly used as sweeteners in English cooking. The weights are given in Table 20. 7.1.7. Fuel The fuel index has three components – wood and peat, charcoal, and coal. Charcoal was a smokeless version of wood used by the richer. Coal was the

91


Table 21. Decade 1815–1819 1820s 1830s 1840s 1850s 1860s

The Cost of Illumination through Candles and Coal Gas. 1 lb Tallow Candles (d.)

19 ft3 Gas (d.)

10.5 7.1 6.2 5.9 6.3 6.4

3.5 3.0 2.4 1.7 1.1 1.0

smokiest fuel, and hence least desirable. Because of the high cost of transporting fuel, the use of each was dictated by local supply and transport conditions. By the eighteenth century coal dominated in big cities like London, but wood fuel supplies still dominated in countryside locations without good water transport connections. Table 21 shows the weights assigned over time to each fuel type. 7.1.8. Lodging The method for forming the rental values of housing of constant quality is described in Clark (2002b). For this estimation I have 5,125 observations in total, 757 for the years before 1500. Over the Industrial Revolution, with greater urbanization, the rental value of housing increased greatly relative to the general price level. Consequently, the weight given to housing in Table 14 is increased in this period. Greater weights are also assigned in the years before 1375 when the return on capital invested in housing was much greater than in subsequent years, more than 10 percent, implying that corresponding rental values would be greater. Since house rent estimates only go back as far as the 1290s, for earlier years house rent is estimated as the average of 1290–1349, but indexed by the relative price level in the earlier year compared to the average price level of 1290–1349. 7.1.9. Light The component series here are tallow candles (in the thirteenth century tallow itself), wax candles, sperm oil, and coal gas. An issue is what weight to give gaslight in this index. After 1815 the price of gas illumination was dramatically below that of candles. It was reckoned that 19 ft3 of gas had by 1832 the illumination equivalent to a pound of tallow candles (Matthews, 1832, p. 326). Table 22 shows the cost by decade from 1815 to 1869 of a pound of tallow candles compared to the equivalent amount of 19 ft3 of gas.

92

GREGORY CLARK

Table 22. Commodity

Tallow candles Wax candles Lamp oil Coal gas Tallow

The Weights in the Light Price Index.

DE, 1850–1869

DE, 1815–1849

DE, 1281–1814

DE, 1261–1280

DE, 1209–1260

0.600 0.050 0.100 0.250 –

0.750 0.075 0.075 0.100 –

0.750 0.150 0.100 – –

– 0.250 – – 0.750

– – – – 1.000

When gas was being first introduced, it cost only about 40 percent that of candles. But by the 1850s it cost only about one-sixth that of candles. It has been argued, however, that before 1870 gas illumination was found only in middle and upper class homes (Matthews, 1986). However, the poor as well as the rich benefited from the use of gaslight for street illumination, pubs, and shops. By 1876 there were 54,000 street lamps lit by gas in London alone (Chubb, 1876, p. 350). It seems thus that the transformation of public spaces by gaslight in the years 1815–1869 should get some weight in the COL of even the poor. In terms of the weight in the domestic expenditure price index, by the 1860s gas consumption measured just in the price of gas was about 1 percent of GNI. In the 1840s it was about 0.5 percent of GNI (Matthews, p. 247). Table 23 shows the weights in the lighting index over time. To allow for the poor having less access to the benefits of gaslight, in the workers’ COL index gas is counted with only half this national weight. The other weight within the COL light series is all devoted to tallow candles. 7.1.10. Clothing and Bedding The sources for prices here are varied – wool cloth, woolen blankets, linen cloth, cotton cloth, silk thread, stockings, complete suits of clothing (other than stockings), boots and shoes, and leather gloves. Table 24 shows the weights assigned over time to each item of clothing or bedding. 7.1.11. Services The pre-industrial economy had a vast array of domestic servants: cooks, house maids, grooms, and coachmen. This shows even in the 1851 census. At that date, weighting men and women, and boys and girls by their earnings, 13.1 percent of the labor force was engaged in some type of personal service (Parliamentary Papers, 1852–1853, Table 25, pp. 222–227). Wages were 64 percent of national income in 1851, so that this implies that

93


Table 23. Commodity

Wood Charcoal Coal

DE, 1820–1869

DE, 1750–1819

DE, 1690–1749

DE, 1590–1689

DE, 1209–1589

0.16 0.04 0.80

0.32 0.08 0.60

0.48 0.12 0.40

0.60 0.15 0.25

0.64 0.16 0.20

Table 24. Commodity

Wool cloth Linen cloth Cotton cloth Silk thread Stockings Suit of clothes Shoes Leather gloves Labor

The Weights in the Fuel Price Index.

The Weights in the Clothing Price Index.

DE, DE, DE, DE, DE, DE, DE, 1856–1869 1820–1855 1790–1819 1765–1789 1633–1764 1576–1632 1549–1575 0.200 0.075 0.075 – 0.050 0.500 0.100 – –

0.190 0.071 0.071 0.050 0.048 0.475 0.095 – –

0.190 0.095 0.048 0.050 0.048 0.475 0.095 – –

0.190 0.124 0.019 0.050 0.048 0.475 0.095 – –

0.190 0.142 – 0.050 0.048 0.475 0.095 – –

0.285 0.142 – 0.050 – 0.333 0.095 – 0.095

0.500 0.250 – – – – 0.100 – 0.15

8.4 percent of expenditure was on service of some kind: domestic servants, barbers, doctors, nurses, gardeners, and teachers. Thus, the share of expenditure devoted to personal service is assumed at 8 percent in 1840–1869, and somewhat lower in the earlier years (Table 14). 7.1.12. Manufactures Certainly by the end of our period, 1869, the average person was consuming a quantity of manufactured goods aside from clothing and bedding. There were wooden utensils, furniture, brooms, hairbrushes, glasswares, cutlery, pottery, pewter, cooking implements, garden tools, haberdashery, and spectacles. An estimate of the potentially substantial share of these goods in expenditure comes from insurance policies from the years 1750 to 1850 analyzed by Pollard (1988). The policy value of the average house insured in the years 1801–1850 was d449. At the same time the value of the contents insured averaged d242, more than half the structure value (Pollard, 1988, pp. 250, 256). The contents consisted of clothing, bedding, plate, jewelry, housewares, and furniture. Unfortunately Pollard does not sub-divide this

94

GREGORY CLARK

category. But if even just half the value of housewares was for items other than clothing and bedding, then annual purchases of housewares and furniture must have been substantial in nineteenth-century England. Using primarily the copious records of the Founding Hospital in London from 1759 to 1856, I am able to derive price series for many of these items in the Industrial Revolution period, which I class under ‘‘manufactures.’’ As Table 14 shows these are given a very modest weight in the overall expenditure index, but were included as a potential area of significant declines in relative prices as a consequence of the Industrial Revolution. 7.1.13. Investment Goods Under this heading are included construction materials (bricks, timber, and manufactured iron), as well as implement prices (spades and shovels), and window glass. Table A1 shows the decadal level of each of the individual price series and the resulting 11 major component price indices.

8. EXPORT PRICE INDEX Table 25 gives total calculated exports and imports from England from 1784 to 1856 taken from Davis (1979), and the implied average share of exports and imports of net national income. The total value of English exports is inferred from UK exports for 1834–1836 and later, from British exports for 1784–1786 to 1824–1826. This is done by assuming that England was 84 percent of British exports, and that Ireland received the same share of British exports in later years as in 1824–1826. It was assumed throughout that all cotton goods, wool cloth, manufactures, iron, coal, and sugar exports from the United Kingdom were from Britain, with England supplying 84 percent of each. All linen exports from the United Kingdom were assumed to come from Ireland. Table 24 also lists the major exports for which I have price series their shares of total English exports on these assumptions. The price index for exports was based on a weighted average of these prices, with the weights changing each 10 years. Table 26 shows the total value of exports and imports for England for 1699–1774, where the export and import data come from Davis (1962), and refer to England. Table 26 also shows the share of exports for the commodities for which I have prices over these years.

95


Table 25.

Exports and Imports, 1784–1856.

Item

1784– 1786

1794– 1796

1804– 1806

1814– 1816

1824– 1826

1834– 1836

1844– 1846

1854– 1856

Exports (million d) Imports (million d) Ratio to NNI

11.4 16.1 0.125

20.2 24.8 0.158

34.6 38.4 0.164

40.3 45.4 0.146

33.5 47.7 0.133

43.9 50.4 0.144

55.5 59.8 0.148

97.3 109.7 0.219

Export share Cotton goods Woolen goods Ironwares Iron Manufactures Coal Refined sugar

0.059 0.285 0.090 0.000 0.278 0.017 0.026

0.158 0.240 0.093 0.000 0.258 0.012 0.050

0.396 0.165 0.066 0.002 0.177 0.009 0.040

0.396 0.182 0.018 0.006 0.160 0.009 0.061

0.436 0.173 0.021 0.010 0.177 0.011 0.019

0.441 0.168 0.023 0.024 0.157 0.018 0.018

0.403 0.173 0.038 0.045 0.155 0.046 0.007

0.310 0.132 0.061 0.076 0.185 0.074 0.004

Import shares Cotton Grains Sugar Timber Wool Oils Silk Tea Flax Wine Butter Hemp, jute Copper Indigo Meat Rice Spirits Tobacco Cheese Coffee Tallow Linens

0.093 0.038 0.124 0.048 0.014 0.016 0.059 0.114 0.029 0.045 0.020 0.020 0.000 0.019 0.017 0.008 0.019 0.001 0.000 0.001 0.027 0.082

0.089 0.079 0.156 0.055 0.020 0.015 0.035 0.080 0.022 0.047 0.021 0.040 0.001 0.042 0.024 0.009 0.032 0.004 0.003 0.013 0.024 0.058

0.121 0.061 0.132 0.070 0.039 0.020 0.036 0.073 0.030 0.036 0.021 0.042 0.001 0.033 0.022 0.004 0.017 0.006 0.007 0.003 0.034 0.049

0.142 0.056 0.156 0.070 0.072 0.026 0.041 0.068 0.024 0.031 0.025 0.024 0.000 0.043 0.025 0.004 0.007 0.008 0.007 0.014 0.030 0.037

0.117 0.072 0.107 0.070 0.074 0.019 0.057 0.062 0.033 0.049 0.027 0.017 0.001 0.053 0.031 0.003 0.007 0.007 0.010 0.007 0.021 0

0.221 0.054 0.081 0.049 0.110 0.021 0.070 0.046 0.035 0.028 0.028 0.012 0.001 0.023 0.029 0.003 0.010 0.006 0.007 0.008 0.025 0

0.147 0.117 0.081 0.074 0.075 0.021 0.053 0.030 0.029 0.015 0.027 0.015 0.005 0.030 0.035 0.008 0.007 0.006 0.011 0.007 0.023 0

0.151 0.159 0.068 0.064 0.041 0.029 0.040 0.031 0.023 0.017 0.025 0.016 0.015 0.015 0.026 0.010 0.008 0.007 0.007 0.006 0.017 0

The price index for exports was thus composed of indices with the indicated weights for cotton cloth, woolen cloth, manufactured iron, pig iron, manufactures, coal, sugar, and wheat. These should be FOB prices, but I use, as the nearest approximation, domestic retail prices.

96

GREGORY CLARK

Table 26. Item

Exports and Imports, 1699–1774.

1699–1701

1722–1724

1752–1754

1772–1774

Exports (million d) Imports (million d) Ratio to NNI

4.43 3.86 0.072

5.04 4.04 0.066

8.42 4.71 0.086

9.85 6.92 0.088

Export shares Cotton goods Woolen goods Ironwares Manufactures Coal Grains

0.000 0.687 0.026 0.063 0.008 0.033

0.000 0.592 0.036 0.074 0.019 0.117

0.000 0.467 0.070 0.134 0.021 0.107

0.022 0.425 0.122 0.187 0.034 0.004

Import shares Cotton Grains Sugar Timber Wool Silk Tea Wine Tallow Linens

– 0.000 0.089 0.036 0.052 0.073 0.000 0.139 0.022 0.187

– 0.000 0.177 0.039 0.028 0.162 0.000 0.142 0.004 0.193

– 0.000 0.185 0.050 0.016 0.133 0.117 0.087 0.001 0.136

0.020 0.043 0.280 0.046 0.015 0.091 0.080 0.059 0.019 0.022

9. IMPORT PRICE INDEX The main imports of England at various periods are also listed in Tables 25 and 26, taken also from Davis (1962, 1979). The total value of English imports is inferred from UK imports for 1834–1836 and later, from British imports for 1784–1786 to 1824–1826, and from English imports for 1699–1701 to 1772–1774. This is done by assuming that England was 84 percent of British imports. For Ireland after 1824–1826 I assume that Irish exports to England were the same in nominal terms in 1834–1836, 1844–1846, and 1854–1856 as in 1824–1826. I also assume that Ireland took an amount of the imports of foods to the United Kingdom as equaled its exports to England in these years. After deducting the assumed share of Ireland, I assume for each good that 84 percent of the imports went to England. In later years the dominant imports were raw materials or processed farm products – cotton, grains, sugar, timber, wool, tea, silk, tallow, oils, flax, hemp, indigo, wine, butter, meat, spirits, and copper.


97

Earlier there are substantial manufactured imports into England in the shape of linens. The wholesale prices of many of these imports are available from the work of Thomas Tooke and William Newmarch for 1782–1859, and for 1854–1869 from the average prices of imports to the United Kingdom recorded in government trade statistics (Mulhall, 1899, pp. 471–477).

10. REAL NDI AND REAL NDP Table 27 shows the resulting decadal price indices, PDE, PNDP, PX, and PM. All the indices have 1860–1869 set at 100. Because of the more rapid decline of export compared to import prices in the years 1760–1869, PDE rises more than PNDP. Thus, real NDI grows more slowly in the Industrial Revolution era than real NDP. Table 27 also shows estimates of the average of import and export values to NDI for each decade where this is available. Table 28 shows decadal estimates of real NDI, real NDP, real NDI per capita, and real NDP per capita. Fig. 8 shows the decadal estimates of NDI per capita from the 1200s to the 1860s. What is remarkable here is first the high implied pre-industrial level of income per capita in some periods: 1209–1239, and 1380–1509. Estimated incomes in some decades here exceeded average incomes in England for all decades before the 1820s. There were also periods of quite low incomes per person, as in 1270–1329. The high incomes of the 1200s–1240s may be a statistical aberration. The data for the 1200s, for example, are for 1209 only and are based on a couple of wage quotes, with the prices of only a handful of commodities. But the data for the 1440s and 1450s, when estimated income per person exceeded that of the 1840s, are rich and detailed. The implied estimated growth rates of NDP per capita in the Industrial Revolution era are low, relative even to the relatively pessimistic estimates of Harley and Crafts from primal sources. Thus, over the 100 years from the 1760s to the 1860s, real NDP per capita increased by 60 percent, at an average annual rate of 0.47 percent. Crafts and Harley estimate an annual growth rate of GDP per person in this interval of 0.55 percent (1992). This in turn would imply an overall increase of GDP per person of 73 percent in these years. Growth measured in terms of national income was even slower because of the decline in the terms of trade. Thus, income per person (NNI) increased by only 48 percent over the 100 years between the 1760s and the 1860s, implying an annual growth rate of only 0.39 percent. Fig. 8, showing income

98

GREGORY CLARK

Table 27. Decade

1200 1210 1220 1230 1240 1250 1260 1270 1280 1290 1300 1310 1320 1330 1340 1350 1360 1370 1380 1390 1400 1410 1420 1430 1440 1450 1460 1470 1480 1490 1500 1510 1520 1530 1540 1550 1560 1570 1580 1590 1600 1610 1620

Price Indices, 1200s–1860s.

Domestic Expenditures

Imports

Exports

Imports/NDI

National Product

Cost of Living

6.59 7.47 8.32 8.23 8.72 8.83 9.33 11.09 10.01 11.26 11.07 13.59 12.58 10.98 10.37 12.99 13.46 13.68 11.57 11.81 12.33 12.52 11.72 12.77 11.43 11.57 11.73 11.82 12.48 11.79 12.09 12.44 14.32 15.09 17.01 25.3 27.1 29.8 33.1 41.4 43.0 48.7 49.0

– – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – –

– – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – –

– – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – –

7.13 8.08 9.00 5.56 9.43 9.56 10.10 12.00 10.83 12.19 11.97 14.70 13.61 11.88 11.22 14.05 14.57 14.80 12.52 12.78 13.34 13.55 12.68 13.82 12.37 12.52 12.70 12.79 13.50 12.75 13.08 13.46 15.50 16.33 18.40 27.4 29.3 32.2 35.9 44.8 46.5 52.7 53.0

6.54 7.57 8.53 8.39 8.87 9.05 9.46 11.63 10.40 11.61 11.20 14.14 12.97 11.06 10.47 13.09 13.52 13.78 11.40 11.68 12.22 12.40 11.47 12.73 11.14 11.35 11.51 11.60 12.37 11.54 11.94 12.24 14.37 15.21 17.19 26.1 27.5 30.4 34.1 43.8 45.1 51.4 51.7

99



1630 1640 1650 1660 1670 1680 1690 1700 1710 1720 1730 1740 1750 1760 1770 1780 1790 1800 1810 1820 1830 1840 1850 1860

Domestic Expenditures

Imports

Exports

Imports/NDI

National Product

Cost of Living

57.0 59.1 59.3 59.7 59.0 56.6 63.0 58.4 62.8 62.6 57.8 59.3 62.3 65.9 71.6 74.3 85.2 114.6 130.2 108.5 100.9 96.9 93.3 100.0

– – – – – 85.0 102.5 112.0 104.8 101.4 96.7 99.3 94.7 91.1 95.6 98.1 110.0 137.3 146.8 104.3 97.2 80.0 82.8 100.0

– – – – – 151.7 163.3 168.8 168.6 168.4 158.0 158.1 157.4 156.8 157.7 162.5 174.4 167.3 179.5 125.8 108.6 83.0 81.8 100.0

– – – – – – 0.066 0.066 – 0.060 – – 0.078 – 0.079 0.112 0.141 0.145 0.129 0.118 0.127 0.130 0.193 0.250

61.7 63.9 64.2 64.6 63.8 61.3 68.0 62.8 67.8 67.8 62.5 64.0 67.5 71.6 77.4 80.5 91.6 118.6 135.0 112.1 103.0 97.8 93.2 100.0

60.7 62.7 62.2 63.0 62.0 58.9 66.4 59.8 65.6 64.9 58.8 60.6 65.1 69.1 76.4 78.3 89.8 122.5 137.1 110.2 101.3 98.8 95.1 100.0

per person in England from 1200s to 1860s, implies that this makes the discontinuity of the Industrial Revolution less clear. Was the Industrial Revolution just the acceleration of a period of slow growth beginning around 1600?

11. COST OF LIVING INDEX AND REAL WAGES One big issue in the Industrial Revolution era is the standard of living of workers. That requires a COL index which has different weights from the national price index. The COL index aimed for here is one that applies to the average wage earner, not the poorest such as agricultural workers (such an index is reported for agricultural workers in Clark, 2001a, 2001b, 2007b).

100

GREGORY CLARK

Table 28. Decade

1200 1210 1220 1230 1240 1250 1260 1270 1280 1290 1300 1310 1320 1330 1340 1350 1360 1370 1380 1390 1400 1410 1420 1430 1440 1450 1460 1470 1480 1490 1500 1510 1520 1530 1540 1550 1560 1570 1580 1590 1600 1610 1620

Real National Income, 1200s–1860s.

Real National Income (PDE)

Real National Income (PNDP)

Real National Income/N (PDE)

Real National Income/N (PNDP)

Real Wage

14.0 13.1 12.6 12.0 13.1 12.5 13.4 11.2 12.4 11.8 12.0 11.2 12.4 13.0 11.6 11.0 10.7 10.5 11.3 10.5 10.6 10.1 10.4 10.2 10.5 10.5 10.2 9.8 9.5 9.6 10.0 10.6 10.2 10.4 10.2 8.9 9.8 10.8 11.1 10.3 13.2 13.4 14.4

12.9 12.1 11.7 11.1 12.1 11.5 12.4 10.3 11.5 10.9 11.1 10.3 11.5 12.0 10.8 10.2 9.9 9.7 10.4 9.7 9.8 9.3 9.7 9.5 9.7 9.7 9.4 9.1 8.8 8.9 9.3 9.8 9.4 9.6 9.4 8.3 9.0 10.0 10.3 9.5 12.2 12.4 13.3

81.6 76.3 67.0 60.8 66.9 64.2 61.7 45.3 50.4 43.9 44.7 39.6 49.5 54.8 51.4 61.5 66.9 65.8 79.2 73.9 79.6 78.5 83.7 80.6 91.3 91.2 87.0 81.5 78.2 82.4 77.6 74.6 68.3 68.1 67.2 54.4 60.0 60.7 61.8 57.3 59.3 55.9 56.8

75.4 70.5 61.9 56.2 61.9 59.4 57.0 41.9 46.6 40.6 41.3 36.6 45.7 50.6 47.5 56.8 61.9 60.9 73.2 68.4 73.6 72.6 77.4 74.5 84.4 84.3 80.4 75.3 72.3 76.2 71.7 69.0 63.1 63.0 62.1 50.3 55.4 56.1 57.1 52.9 54.8 51.7 52.5

68.8 58.6 53.0 49.8 52.0 52.9 50.7 37.5 43.8 37.9 41.9 36.6 40.7 47.4 47.1 61.8 64.2 69.7 82.3 77.7 81.6 81.5 89.1 84.6 97.1 96.6 91.2 90.6 83.6 90.9 84.3 83.4 72.8 68.0 69.4 59.1 65.4 63.7 58.1 48.0 49.5 46.2 47.8

101



1630 1640 1650 1660 1670 1680 1690 1700 1710 1720 1730 1740 1750 1760 1770 1780 1790 1800 1810 1820 1830 1840 1850 1860

Real National Income (PDE)

Real National Income (PNDP)

Real National Income/N (PDE)

Real National Income/N (PNDP)

Real Wage

13.7 14.6 16.0 16.6 15.9 16.9 15.2 17.1 17.2 18.1 20.1 20.3 21.6 22.8 23.6 25.3 28.7 32.2 38.6 47.9 57.8 68.5 83.5 100.0

12.6 13.5 14.8 15.4 14.7 15.6 14.0 15.9 15.9 16.7 18.6 18.8 19.9 21.0 21.8 23.3 26.7 31.2 37.2 46.4 56.7 67.8 83.6 100.0

51.8 53.3 56.3 58.9 57.4 61.7 55.5 61.2 59.7 61.3 69.3 66.2 68.0 67.8 66.4 65.7 68.4 69.9 73.9 78.9 83.0 86.4 93.7 100.0

47.9 49.2 52.1 54.4 53.1 56.9 51.5 56.9 55.3 56.7 64.1 61.3 62.8 62.4 61.4 60.7 63.6 67.6 71.3 76.4 81.3 85.6 93.7 100.0

43.8 46.3 51.1 52.5 51.6 55.0 49.4 55.7 52.3 53.2 61.9 60.7 57.6 57.4 55.2 56.6 57.5 55.5 60.3 69.4 77.4 82.9 91.1 100.0

Table 29 shows the weights in the principal components of the workers’ COL index for various periods. The principal difference here is the larger weight given to food, and the much smaller weight given to services. These weights are derived from Clark (2001a, 2001b, 2005), which report contemporary surveys of worker consumption patterns in England, mainly for the years 1789–1869. Within the sub-indices the weights also differ. Within the food category, starches are given greater weight and meat and dairy products less (Table 30). In drinks, beer is much more important than wine in the COL index (Table 31). In clothing, silk is excluded from the COL index. The resulting COL index by decade is shown in Table 27. Table 28 shows the implied decadal real wage. Real wages were also surprisingly high in the pre-industrial era, as Fig. 9 shows. What is also evident is that real wages rise more than national income per person in the Industrial Revolution era.

102

GREGORY CLARK 120

Real NNI/N (1860s =100)

100

80

60

40

20

0 1200

1300

1400

Fig. 8.

Table 29. Commodities Food Fuel Lodging Light Soap Clothing Services Tobacco Books Manufactures

1500

1600

1700

1800

NDI/N, 1200s–1860s.

The Weights in the Workers’ Cost of Living Price Index.

1840–1869 1820–1839 1730–1819 1375–1729 1280–1374 1245–1279 1209–1244 0.620 0.050 0.100 0.040 0.010 0.120 0.020 0.010 0.015 0.015

0.670 0.050 0.075 0.040 0.005 0.120 0.010 0.010 0.010 0.010

0.720 0.050 0.045 0.040 0.005 0.120 0.010 – – 0.010

0.720 0.050 0.045 0.040 0.005 0.120 0.020 – – –

0.720 0.040 0.075 0.030 0.005 0.120 0.010 – – –

0.750 0.050 – 0.040 – 0.150 0.010 – – –

0.800 – – – – 0.180 0.020 – – –

Real wages rose 74 percent between the 1760s and the 1860s, compared to an only 48 percent increase in real income per person (Fig. 10). Real wages rose faster than real income per person mainly because of the increased share of wages in national income over these years (caused by the decline of land rents as a share of national income). As Table 13 shows, labor income was 59 percent of all income in the 1760s, but 65 percent by the 1860s. Land income fell over the same years from 20 to 8 percent. A small

103



The Weights in the Cost of Living Food Price Index. 1840–1869

1820–1839

1730–1819

1209–1729

0.50 0.12 0.12 0.03 0.03 0.11 0.07 0.01 0.01

0.55 0.11 0.11 0.03 0.03 0.10 0.05 0.01 0.01

0.60 0.10 0.10 0.03 0.03 0.09 0.03 0.01 0.01

0.60 0.11 0.11 0.04 0.03 0.10 – 0.01 –

Starches Meat Dairy Fats Fish Drinks Sugars Salt Spices


The Weights in the Cost of Living Drinks Price Index.

1820–1869

1780–1819

1760–1779

1704–1759

1486–1703

1209–1485

– 0.65 0.00 0.30 0.05 –

– 0.75 0.00 0.20 0.05 –

– 0.85 0.00 0.10 0.05 –

– 0.85 0.05 0.10 – –

– 0.75 0.25 – – –

0.25 – 0.25 – – 0.50

Cider Beer Wine Tea Coffee Barley

120

Real Wage (1860s =100)

100

80

60

40

20

0 1200

1300

Fig. 9.

1400

1500

1600

1700

Real Wages by Decade, 1200s–1860s.

1800

104

GREGORY CLARK 120

Real GNI/N, Wage (1860s =100)

100 Real Income per Capita 80

60 Real Wage 40

20

0 1760

Fig. 10.

1780

1800

1820

1840

1860

Real Wages and Real Income per Capita by Decade, 1760–1869.

part of the reason for the faster growth of real wages than of incomes comes from the different movement of the COL of workers, compared to the cost of expenditures as a whole. The COL index shows a 45 percent rise in 1760s–1860s, compared to a 52 percent rise in the domestic expenditures price index (and a 40 percent rise in the GDP deflator). The COL rose less than general expenditure prices principally because of the much smaller weight in this index of services – teachers, doctors, nurses, cooks, house maids, gardeners, and coachmen. These were 8.7 percent of all expenditures, but only an estimated 1.3 percent of the expenditures of workers. Because urban wages rose by 148 percent in these years, which was more than any other expenditure price, the expenditure deflator rose by more than the COL of workers. Indeed the higher weight given to services explains most of the difference in the aggregate between these two indices over these years. The weights for other items did differ substantially, as Table 32 shows. The poor ate much more grain and potato products. But here the rate of price increase did not differ much from that applicable to the average good. The workers drank more beer, whose price increase was more than the average, but in compensation, richer consumers were assumed to drink more wine. But since there were so many changes in weights and relative price movements, as Table 32 shows, mostly these effects cancelled out in terms of the rise of the COL versus PDE.

105


Table 32. Commodity

Salt Cotton cloth Sugar Linen cloth Tallow candles Stockings Coal Wool cloth Suit of clothes Firewood Manufactures Starches Wax candles Shoes Investment Lamp oil Beer Meat Dairy Wine Services Lodging

The Weights in the Domestic Expenditure and Cost of Living Indices, 1760–1869. Weight DE

Weigh COL

Price in 1860s/ Price in 1760s

0.005 0.006 0.027 0.011 0.029 0.005 0.035 0.022 0.055 0.015 0.030 0.259 0.005 0.011 0.074 0.004 0.036 0.066 0.066 0.009 0.087 0.068

0.006 0.007 0.030 0.016 0.038 0.006 0.035 0.024 0.057 0.015 0.011 0.387 0.000 0.011 0.000 0.000 0.048 0.073 0.073 0.000 0.013 0.054

0.26 0.38 0.84 0.87 0.94 1.03 1.06 1.10 1.12 1.26 1.30 1.53 1.61 1.66 1.68 1.81 1.89 1.91 2.03 2.27 2.48 2.82

12. EFFICIENCY Since population changed greatly over the years, measures such as real income per person or real wages do not reveal the efficiency level of the economy directly. To calculate that we need to estimate Eq. (1), which looks just at the weighted ratio of input prices (net of indirect taxes) to output prices. We now have all the prices and weights we need to estimate expression (1): r is the real interest rate, PK the index of price of capital, w the index of wages, s the index of farmland rents, P the price index for output, t the share of national income collected in indirect taxes, and a, b, and c the shares in each year of the factor payments of capital, labor, and land, respectively. The shares of labor, land, and capital are calculated based on output values net of indirect taxes, with poor rate collections attributed to either land or capital. Though the index has the Cobb–Douglas form, the

106

GREGORY CLARK

shares of labor, capital, and land are updated annually in calculating the year-to-year movement of efficiency, implying there is no underlying assumption of a Cobb–Douglas technology. In fact, the index is agnostic on the form of the production function. Eq. (1) can be rewritten as:

ðrt þ lÞPKt At ¼ Pt ð1 tt Þ

a

wt Pt ð1 tt Þ

b

st Pt ð1 tt Þ

c (4)

Economic efficiency is thus the geometric weighted average of the real rental of capital, the real price of labor, and the real rental of land. The capital stock earlier consists mainly of a combination of structures (houses, inns, mills, shops, and warehouses) as well as farm working capital (animals and grains) and non-farm working capital. It thus seems reasonable to assume that the price of this capital moves in line with general prices in the economy, so that PKt ¼ Pt. This assumption implies that the capital goods sector experienced as much efficiency growth as the rest of the economy. In implementing this index of productivity, the question arises about what value to give to l, the risk premium on capital. Since early interest rates are high, the greater is l, the smaller will be the capital cost contribution to productivity levels in earlier years. But the capital stock was a combination of houses, with little risk premium on their return, farm capital with an intermediate risk premium, and commercial capital with a significant risk premium. As a compromise, I take l ¼ 3 percent. Fig. 11 shows productivity from 1209 to the 1860s calculated with the expenditure prices from England. It shows the overall efficiency of the economy in translating inputs into real income, whether the purchases were produced domestically or abroad. The decades 1200–1209 to 1280–1289 are shown as a dotted line to emphasize how much more tentative the estimates are for these decades. Fig. 12 shows productivity over the same period calculated based on the prices of English output only. The effect of using the NDP price index in the years before 1870 shows mainly in the greater rise of efficiency in the Industrial Revolution period, the 1780s–1860s. Measured using expenditure prices the productivity growth rate is 0.37 percent per year from the 1760s to 1860s. Measured using output prices the productivity growth rate for the 1760s–1860s is 0.45 percent per year. This 0.45 percent for productivity growth measured on the basis of output prices is similar, but again below, what has been regarded as the pessimistic rate of 0.55 percent per year calculated by Crafts and Harley (1992), which is done on a GDP basis.

107


Efficiency (DE) (1860s =100)

100

80

60

40

20

0 1200

Fig. 11.

1300

1400

1500

1600

1700

1800

Productivity, 1209–1869 – Measured Relative to Expenditure Prices.

120

Efficiency (NNP) (1860s =100)

100

80

60

40

20

0 1200

Fig. 12.

1300

1400

1500

1600

1700

1800

Productivity, 1209–1869 – Measured Relative to GDP Prices.

108

GREGORY CLARK 300

Real factor payment (1860s =100)

250 Capital 200

150

100

Wage

50 Rent 0 1200

Fig. 13.

1300

1400

1500

1600

1700

1800

The Components of English Economic Efficiency Measurement, 1209–1869.

But I will discuss further below why the expenditure-based index seems more appropriate. Table 32 shows both these sets of productivity measures by decade. Fig. 13 shows why the measured efficiency of the economy moves as it does from the 1200s to 1860s, by looking at the components of Eq. (4) for efficiency. The reason that measured efficiency is so unexpectedly high in the years before 1350, despite the relatively low levels of output per person, is the high calculated underlying real interest rate in this epoch. Real interest rates in the thirteenth century were 2.5 times as great as in the eighteenth and nineteenth centuries. Since I have assumed that in both farming and non-farm activities there was a fixed ratio for working capital to output, this raises the weight of capital, a, in the productivity equation (1) in these years, since the share of capital in all payments is assumed to rise. Had I assumed that capital entered the production function in a Cobb–Douglas fashion, with a fixed share of payments to factors going to reproducible capital, then the implied productivity of the years before 1350 would be significantly smaller. Since we cannot observe directly the level of the capital stock in these years, the level of efficiency in the period before 1350 thus remains uncertain, as of course does output per person.


109

12.1. Productivity Fluctuations, 1200–1700 One of the things that makes it very difficult to decide when the transition to modern productivity growth began is the fluctuations we observe in Figs. 11 and 12 in the productivity level of the pre-industrial economy. If these are measurement error, then they suggest the measure of productivity for the pre-industrial era is sufficiently poor that we will never be able to know whether there was a gradual or sudden transition to modern economic growth. If they are not measurement error, then they imply quite inexplicable fluctuations in the performance of the pre-industrial economy. I can say for sure that the observed fluctuations in productivity are not just measurement error, but reflect real changes in the efficiency of the economy. To see this consider Fig. 13. The real rental of capital was clearly falling over time, and that implies higher earlier productivity. The fluctuations in the productivity measure over the pre-industrial period are associated with the very substantial fluctuations in real wages. In 1450 real wages in England were nearly as great as they were at the end of the Industrial Revolution in the 1860s. These fluctuations in turn were highly correlated with movements in the population. There were only about 2.2 million people in England when real wages reached their medieval maximum around 1450. There were nearly 6 million people when they were at their medieval minimum in the 1310s. For measured productivity to remain constant over these huge population fluctuations, there had to be countervailing declines in land rents. But since wages were 44–67 percent of estimated factor payments, even in the medieval period, the countervailing swings in rents would have to be much more substantial than those of wages in order for measured productivity to remain constant. In practice there were instead relatively modest downward movements in real rents in the periods of low population, as Fig. 9 shows. Thus, the fluctuations in medieval productivity do not seem attributable to measurement error, in that the wage series that is causing it is one of the best measured of all the series. Resolving this puzzle is difficult. One possible step toward a resolution would be to re-examine the assumption underlying the generation of the factor shares in the earlier economy, that hours of labor input per person were constant from 1200 to 1869. If hours of labor input were much smaller in the medieval period, then the share of rents in total income would be higher, and the implied productivity fluctuations smaller. But the size of the hours reduction in the period of high measured efficiency necessary to eliminate the measured efficiency gains would have to

110

GREGORY CLARK

be very substantial. And there is little other evidence of any substantial changes in hours worked.

13. OUTPUT AND EFFICIENCY, 1860–2008 In this section I link the earlier estimates of real income, real wages, and factor shares to the modern data for England. The data for 1870 and later are mainly quoted on a UK basis, meaning England, Wales, Scotland, and Ireland (after 1922 Northern Ireland only). UK income is divided into English versus other by, as before, taking the English population divided by British population plus one-half of the Irish population (to allow for the lower income of Ireland). 13.1. Real GDP per Person Feinstein gives for 1860–1920 real GDP per capita for Great Britain (Feinstein, 1972, Tables 118–119). This series is continued for England in the years 1920–1948 by deflating nominal GDP per person for the United Kingdom by Feinstein’s GDP deflator (Feinstein, 1972, Tables 5–9, pp. 132–133). For 1948–2008 the Office of National Statistics (UK) reports real GDP per capita. The decadal averages for this series are reported in the second column of Table 33. This is thus a continuation of the columns in Table 28 giving measures of real NNI per capita and real GDP per capita for the 1860s and earlier. 13.2. Farmland Rent Share NDI for England in the years 1860–2007 is calculated by reducing UK NDI to an English basis. Total farmland rents for England in the years 1860–1914 are from Stamp (1922, p. 49), England and Wales, adjusted to an English basis. This Stamp series is projected onwards to 1944 using the movement of UK farmland rents reported in Feinstein (1972, Table 60). For 1944–2004 the Department of Environment, Food, and Rural Affairs (DEFRA) reports a series ‘‘Agricultural Land Sales and Prices in England’’ which gives prices per acre/hectare. This are converted into rents per acre in 1945–1967 using a return rate of 3 percent, and multiplied by an assumed 28 million acres of farmland in England to give total rents. For 1968–2007 DEFRA reports a series on land tenancy rentals per hectare (normed to the

111


Table 33. Decade

1860–1869 1870–1879 1880–1889 1890–1899 1900–1909 1910–1919 1920–1929 1930–1939 1940–1949 1950–1959 1960–1969 1970–1979 1980–1989 1990–1999 2000–2008

Real Income and Factor Shares, 1860–2008.

Real Income per Person (Index)

Share Land Rents

Share Labor

Share Urban Rents

Share Capital

Real Wage

Skill Premium

Efficiency (Index)

100 110 118 129 141 147 141 168 211 233 292 364 439 548 694

0.075 0.060 0.052 0.034 0.025 0.022 0.007 0.006 0.004 0.004 0.004 0.004 0.004 0.002 0.002

0.656 0.664 0.675 0.682 0.671 0.722 0.759 0.750 0.782 0.802 0.802 0.805 0.772 0.745 0.750

0.026 0.026 0.041 0.045 0.033 0.022 0.021 0.029 0.048 0.039 0.043 0.044 0.049 0.045 0.042

0.243 0.250 0.233 0.239 0.271 0.239 0.209 0.212 0.165 0.154 0.151 0.146 0.174 0.207 0.206

100 128 148 172 182 162 208 236 230 255 303 507 552 731 887

1.53 1.42 1.46 1.42 1.41 1.36 1.22 1.25 1.20 1.10 1.09 1.16 1.17 1.18 1.22

100 119 130 140 146 138 157 169 169 181 204 300 329 367 407

average of Full Annual Tenancy and Farm Business Tenancy leases). Again these are multiplied by an assumed 28 million acres farm area to give total farm rents. The implied share of rents in national income in England is shown in Table 33. 13.3. Share of Wages Feinstein reports both wages and salaries and income from self-employment for the United Kingdom in the years 1860–1947 (1972, Tables 4–9). Income from self-employment is assumed throughout to be half labor income. This is divided by UK NDI minus expenditure taxes (less subsidies). For 1948–2008 the share of wages is calculated in the same way from the series reported by the Office of National Statistics. 13.4. Urban Site Rents The Blue Books (Office of National Statistics) give from 1987 to 2007 the value for the United Kingdom of dwellings and other buildings (including

112

GREGORY CLARK

and excluding the land value). From this can be inferred the capital value of urban site rents. Since site rents exhibit no depreciation, these are converted into site rents by multiplying them by an assumed rate of return (excluding capital gains) of 3 percent over the years 1987–1996 (the rental return on farmland averaged only 2.8 percent between 1968 and 2007). From 1996 to 2005 house prices in England rose by 170 percent, while house rents rose by 34 percent. If the rental return on housing was 3 percent in 1996, then by 2005 it was 1.5 percent. So for these years the capitalized site values are converted into implicit rentals by the current implied rental on housing. These data suggest that urban site rentals represented then about 4 percent of total incomes (and about 45 percent of the value of dwellings and other buildings). Singer (1941) estimates urban site values in England and Wales, 1845–1910, from the movement of construction costs versus tax assessments. The weakness in this estimate, however, is that he has to make an assumption about the share of site rents in property rentals in 1845, which he assumes then is only 8.2 percent of building rents. This gives site rents in the years 1860–1910. Site rents in the years 1947–1986 are estimated from the relative movement of the total average house values (constructed as number of households average house price) compared to the average value of the reproducible housing capital stock reported by the Office of National Statistics. The house price series continues back to 1930. It is used to project site values back to then by assuming the same proportion of house values were site rents throughout the 1930–1947 period.

13.5. Share of Capital This was constructed as a residual so that the shares of land, labor, urban site rents, and capital summed to 1.

13.6. Real Wage and Skill Premium The real wage series here is that of Clark (2005), and covers an average of building workers and craftsmen. Real wages are calculated as the average of nominal wages per 10 h for building tradesmen and laborers, deflated by Feinstein’s COL index (until 1993) (1995), and thereafter by the retail price

113


Table 34. Decade 1200 1210 1220 1230 1240 1250 1260 1270 1280 1290 1300 1310 1320 1330 1340 1350 1360 1370 1380 1390 1400 1410 1420 1430 1440 1450 1460 1470 1480 1490 1500 1510 1520 1530 1540 1550 1560 1570 1580 1590 1600 1610 1620

Efficiency Indices, 1200s–1860s.

Indirect Tax Share

Efficiency (PDE)

Efficiency (PNDP)

0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.003 0.004 0.003 0.003 0.003 0.004 0.006 0.007

108.2 98.7 90.7 85.4 88.5 85.0 81.1 66.0 70.8 66.2 66.2 60.6 73.2 75.4 70.4 73.7 77.4 74.9 85.8 80.2 83.7 82.1 85.8 83.9 91.1 91.0 87.5 82.1 79.1 80.6 77.7 75.8 71.0 70.5 68.6 53.3 57.4 58.5 58.3 48.6 56.6 54.3 55.7

100.0 91.3 83.8 79.0 81.8 78.6 74.9 61.0 65.5 61.2 61.2 56.0 67.7 69.7 65.1 68.1 71.6 69.2 79.3 74.2 77.3 75.9 79.3 77.5 84.2 84.1 80.9 75.9 73.1 74.5 71.8 70.1 65.6 65.1 63.4 49.2 53.0 54.0 53.9 44.9 52.2 50.2 51.5

114

GREGORY CLARK

Table 34. (Continued ) Decade 1630 1640 1650 1660 1670 1680 1690 1700 1710 1720 1730 1740 1750 1760 1770 1780 1790 1800 1810 1820 1830 1840 1850 1860

Indirect Tax Share

Efficiency (PDE)

Efficiency (PNDP)

0.008 0.007 0.008 0.008 0.015 0.015 0.025 0.037 0.042 0.048 0.049 0.046 0.051 0.061 0.059 0.067 0.072 0.095 0.094 0.092 0.072 0.063 0.058 0.048

52.3 54.0 56.9 58.2 56.1 58.9 54.0 58.7 57.8 59.2 65.1 63.0 64.4 64.6 64.4 65.0 67.7 69.6 74.6 79.9 84.7 89.0 95.1 100.0

48.3 49.9 52.6 53.8 51.8 54.4 50.1 54.6 53.5 54.7 60.3 58.5 59.5 59.5 59.6 60.0 62.9 67.4 72.1 77.5 83.0 88.1 95.2 100.0

index from the Office of National Statistics. The skill premium is the differential of the wage of craftsmen to workers.

13.7. Economic Efficiency The decline of land rents to a tiny share of national income since 1860, and the relative constancy of the rate of return on capital, means that the Eq. (1) for calculating economic efficiency since 1860 reduces to the following simple expression: b 1 wt (5) At 1 tt Pt

115


Efficiency (DE) (1860s =100)

400 350 300 250 200 150 100 50 0 1200

1300

Fig. 14.

1400

1500

1600

1700

1800

1900

2000

National Productivity, England, 1209–2004.

The last column of Table 33 gives this simple measure of economic efficiency by decade from the 1860s to the 2000s, with the 1860s set at 100. Because the magnitude of b varies over time between 0.65 and 0.80, the index is constructed by periodically changing the weights and chaining the resulting indices. This is a continuation of the efficiency indices constructed in Table 34. Fig. 14 shows the picture of the efficiency of the English economy from the 1200s to the 2000s. Fig. 14 shows why just with the data on economic growth and economic efficiency it is not possible to assign any definitive moment to the Industrial Revolution. There was a gradual increase in the rate of growth of economic efficiency, and hence also of output per person. But was 1600 the start of this process? Or 1800?

NOTES 1. Coal mining was the other major activity in the primary sector, and was about 10 percent of employment in this sector by 1851. In the years 1830–1869 the wage premium for coal miners compared to farm day laborers averaged 63 percent. However, making explicit allowance for coal miners and their higher wages in the

116

GREGORY CLARK

years 1830–1869 had little effect on the estimated wage trend in the economy in the years 1830–1869. 2. Alternately, employers may pay less per day for annual workers than for day workers because they then have commitments to use workers for a longer period. 3. Voth (2001a) interprets the Exeter Cathedral accounts as suggesting medieval workers put in many few days per year. But on the days when they were not at work on the Cathedral, they may have been engaged elsewhere. If you had used the records of my house remodel to measure the number of days worked by workers in the modern world, you would have concluded that workers put in less than 100 days per year. 4. Feinstein (1972, Table 4) also gives labor income for the United Kingdom in the 1860s that implies much lower levels for England than calculated here. But he also has ‘‘mixed incomes’’ that include some labor income. 5. This was assessed as one-half of land rental values (Stamp, 1922, p. 82). 6. Counting only such incomes d150 or greater. See below. 7. See Stamp (1922, p. 511). 8. See Wratislaw (1861), Tomson (1847), and Squarey (1878). 9. This follows from the fact that, by assumption: Kt Q ¼ t K 0 Q0 ) ðrt þ dÞK t ¼ ðr0 þ dÞK 0

) ðrt þ dÞK t ¼

rt þ d Q t r0 þ d Q 0

ðr0 þ dÞK 0 ½ðrt þ dÞ=ðr0 þ dÞ½ðwagest þ rentst þ taxest Þ=Q0 1 ðr0 þ dÞK 0 ½ðrt þ dÞ=ðr0 þ dÞð1=Q0 Þ

10. Davis (1957, p. 410) estimates the cost 1670–1730 at d11 per ton (d6.5 for the hull and masts, and d4.5 for the rigging). Barney (1999, p. 132) quotes ship prices per ton of d9.7 in the 1760s, and d10.5 in the years 1783–1790. Ville (1990, pp. 47–51) gives ship prices in the coal trade for the years 1792–1825. Harley (1988, p. 872) quotes prices for 1833 and 1852–1858. Graham (1956, p. 80) gives prices for 1825 and 1847. 11. In practice the United Kingdom was a recipient of a positive balance of property income from abroad, but this represented only 2–4 percent of national income in the years 1855–1869 (Feinstein, 1972, Table 4). 12. The London assize called for 240 lb of flour to be made from 6.5 bushels of wheat, or roughly 390 lb of wheat. The other 150 lb was lost as bran in the milling process. By milling less finely to produce a coarser flour, more loaves could be made from a bushel of wheat. 13. This does not seem to be a defect of the wheat price series. That series for the years 1771–1869 is very close to the Gazette series, taken from the whole country, of average wheat prices. Yet in this period there is a nearly 20 percent variation in the price of bread in London relative to wheat prices.


117

REFERENCES Allen, R. C. (2008). Review of Gregory Clark’s a farewell to alms: A brief economic history of the world. Journal of Economic Literature, 46(4), 946–973. Barney, J. M. (1999). Shipping in the port of King’s Lynn, 1702–1800. Journal of Transport History, 20(2), 126–140. Beveridge, W. H. (1939). Prices and wages in England, Vol. 1: The mercantilist era. London: Longmans, Green and Company. Bogart, D. (2005). Did turnpike trusts increase transportation investment in eighteenth-century England. Journal of Economic History, 65(2), 439–468. Bowley, A. L. (1898). The statistics of wages in the United Kingdom during the last hundred years. Part I. Agricultural wages. Journal of the Royal Statistical Society, 61, 702–722. Broadberry, S., Campbell, B., Klein, A., Overton, M., & van Leeuwen, B. (2009). British economic growth, 1300–1850: Some preliminary estimates. University of Warwick. Chubb, H. (1876). The supply of gas to the metropolis. Journal of the Statistical Society of London, 39(2), 350–380. Clark, G. (1988). The cost of capital and medieval agricultural technique. Explorations in Economic History, 25, 265–294. Clark, G. (1996). The political foundations of modern economic growth: England, 1540–1800. Journal of Interdisciplinary History, 26, 563–588. Clark, G. (1998a). Land hunger: Land as a commodity and as a status good in England, 1500–1910. Explorations in Economic History, 35(1), 59–82. Clark, G. (1998b). The reports of the Charity Commission as a source in economic history. Research in Economic History, 18, 1–52. Clark, G. (2001a). Farm wages and living standards in the Industrial Revolution: England, 1670–1869. Economic History Review, 54(3), 477–505. Clark, G. (2001b). Markets and economic growth: The grain market of medieval England. Working Paper. UC Davis. Clark, G. (2002a). Farmland rental values and agrarian history: England and Wales, 1500–1912. European Review of Economic History, 6(3), 281–309. Clark, G. (2002b). Shelter from the storm: Housing and the Industrial Revolution, 1550–1912. Journal of Economic History, 62(2), 489–511. Clark, G. (2005). The condition of the working-class in England, 1209–2004. Journal of Political Economy, 113(6), 1307–1340. Clark, G. (2007a). The long march of history: Farm wages, population and economic growth, England 1209–1869. Economic History Review, 60(1), 97–135. Clark, G. (2007b). A farewell to alms: A brief economic history of the world. Princeton: Princeton University Press. Clark, G., Huberman, M., & Lindert, P. H. (1995). A British food puzzle, 1770–1850. Economic History Review, 48(2), 215–237. Clark, G., & Jacks, D. (2007). Coal and the industrial revolution, 1700–1869. European Review of Economic History, 11, 39–72. Clark, G., & Levin, P. (2001). How different was the Industrial Revolution? The revolution in printing, 1350–1869. Working Paper. University of California, Davis. Clark, G., & Van Der Werf, Y. (1998). Work in progress? The industrious revolution. Journal of Economic History, 58(3), 830–843.

118

GREGORY CLARK

Crafts, N. F. R., & Harley, C. K. (1992). Output growth and the Industrial Revolution: A restatement of the Crafts–Harley view. Economic History Review, 45, 703–730. Davis, R. (1956). Seamen’s sixpences: An index of commercial activity, 1697–1828. Economica, 23(92), 328–343. Davis, R. (1957). Earnings of capital in the English shipping industry, 1670–1730. Journal of Economic History, 17(3), 409–425. Davis, R. (1962). English foreign trade, 1700–1774. Economic History Review, 15(2), 285–303. Davis, R. (1979). The Industrial Revolution and British overseas trade. Leicester: Leicester University Press. De Vries, J. (1994). The Industrial Revolution and the industrious revolution. Journal of Economic History, 54, 249–270. De Vries, J. (2008). The industrious revolution: Consumer behavior and the household economy, 1650 to the present. Cambridge: Cambridge University Press. Deane, P., & Cole, W. A. (1967). British economic growth, 1688–1959 (2nd ed.). Cambridge: Cambridge University Press. Feinstein, C. H. (1972). National income, expenditure and output of the United Kingdom, 1855–1965. Cambridge: Cambridge University Press. Feinstein, C. H. (1988). National statistics, 1760–1920. In: C. H. Feinstein & S. Pollard (Eds), Studies in capital formation in the United Kingdom, 1750–1920 (pp. 258–404). Oxford: Clarendon Press. Feinstein, C. H. (1995). Changes in nominal wages, the cost of living and real wages in the United Kingdom over two centuries. In: P. Scholliers & V. Zamagni (Eds), Labour’s reward: Real wages and economic change in 19th and 20th-century Europe. Aldershot, Hampshire: Edward Elgar. Feinstein, C. H. (1998a). Pessimism perpetuated: Real wages and the standard of living in Britain during and after the Industrial Revolution. Journal of Economic History, 58(3), 625–658. Feinstein, C. H. (1998b). Wage-earnings in Great Britain during the industrial revolution. In: I. Begg & S. G. B. Henry (Eds), Applied economics and public policy. Cambridge: Cambridge University Press. Ginarlis, J., & Pollard, S. (1988). Roads and waterways, 1750–1850. In: C. H. Feinstein & S. Pollard (Eds), Studies in capital formation in the United Kingdom, 1750–1920 (pp. 182–224). Oxford: Clarendon Press. Graham, G. S. (1956). The ascendancy of the sailing ship 1850–85. Economic History Review, 9(1), 74–88 [new series]. Harley, K. (1988). Ocean freight rates and productivity, 1740–1913: The primacy of mechanical invention reaffirmed. Journal of Economic History, 48(4), 851–876. Harley, K. (2001). Cotton textiles and the industrial revolution. Working Paper. University of Western Ontario. Jastram, R. W. (1981). Silver: The restless metal. New York: Wiley. Levi, L. (1867). Wages and earnings of the working classes. London: John Murray. Matthews, D. (1986). Laissez-faire and the London gas industry in the nineteenth century: Another look. The Economic History Review, 39(2), 244–263 [new series]. Matthews, W. (1832). An historical sketch of the origin and progress of gas-lighting. London: Simpkin and Marshall.


119

Mitchell, B. R., & Deane, P. (1962). Abstract of British historical statistics. Cambridge: Cambridge University Press. Mulhall, M. G. (1899). The dictionary of statistics. London: Routledge. Parliamentary Papers. (1803–1804). Report from the Committee on the London Bakers Petition. Parliamentary Papers. (1812–1813). Report from the Committee on the Petition of Certain County Bakers. Parliamentary Papers. (1852–1853). 1851 census: Ages, civil conditions occupations and birthplaces. Part I. Vol. 88, p. 1. Pawson, E. (1977). Transport and economy: The turnpike roads of eighteenth century Britain. London: Academic Press. Pollard, S. (1988). The insurance policies. In: C. H. Feinstein & S. Pollard (Eds), Studies in capital formation in the United Kingdom, 1750–1920 (pp. 225–257). Oxford: Clarendon Press. Shaw-Taylor, L., & Wrigley, E. A. (2008). The occupational structure of England c. 1750–1871: A preliminary report. Cambridge Group for the History of Population and Social Structure, Cambridge, England. Singer, H. W. (1941). An index of urban land rents and house rents in England and Wales, 1845–1913. Econometrica, 9(3/4), 221–230. Squarey, E. P. (1878). Farm capital. Journal of the Royal Agricultural Society of England, 14, 425–444 [second series]. Stamp, J. (1922). British incomes and property. London: King and Son. Tomson, J. (1847). Account of hall farm, near Sevenoaks, Kent. Journal of the Royal Agricultural Society of England, 8, 33–46. Ville, S. P. (1990). English shipowning during the Industrial Revolution: Michael Henley and Son, London shipowners, 1770–1830. Manchester: Manchester University Press. Voth, H. (2001a). Time and work in England, 1750–1830. Oxford: Oxford University Press. Voth, H. (2001b). The longest years: New estimates of labor input in England, 1760–1830. Journal of Economic History, 61(4), 1065–1082. Webb, S., & Webb, B. (1904). The assize of bread. Economic Journal, 14, 196–218. Wratislaw, C. (1861). The amount of capital required for the profitable operation of a mixed arable and pasture farm in a midland county. Journal of the Royal Agricultural Society of England, 2(14), 425–444. Wrigley, E. A., Davies, R. S., Oeppen, J. E., & Schofield, R. S. (1997). English population history from family reconstruction: 1580–1837. Cambridge, NY: Cambridge University Press.

120

GREGORY CLARK

APPENDIX Table A1 gives the decadal level of each of the individual price series and the resulting major components of the price indices. Table A1.

Individual Price Indices, Decadal, 1209–1869.

Decade

Bread (d./lb)

Barley (s./Bushel)

Oatmeal (d./lb)

Peas (s./Bushel)

Potatoes (s./cwt)

Rice (d./lb)

Starches (1860s ¼ 100)

1200 1210 1220 1230 1240 1250 1260 1270 1280 1290 1300 1310 1320 1330 1340 1350 1360 1370 1380 1390 1400 1410 1420 1430 1440 1450 1460 1470 1480 1490 1500 1510 1520 1530 1540 1550

0.091 0.122 0.139 0.130 0.137 0.148 0.145 0.208 0.177 0.224 0.182 0.273 0.231 0.181 0.176 0.240 0.241 0.259 0.183 0.193 0.210 0.216 0.181 0.236 0.182 0.201 0.201 0.210 0.235 0.205 0.229 0.222 0.295 0.323 0.355 0.575

0.206 0.247 0.309 0.277 0.300 0.353 0.330 0.466 0.383 0.502 0.400 0.606 0.488 0.384 0.388 0.587 0.685 0.580 0.376 0.434 0.437 0.439 0.370 0.407 0.313 0.318 0.362 0.319 0.409 0.318 0.348 0.404 0.488 0.468 0.569 1.082

0.076 0.093 0.114 0.133 0.132 0.148 0.153 0.186 0.178 0.188 0.169 0.229 0.222 0.181 0.160 0.235 0.209 0.203 0.216 0.226 0.215 0.202 0.192 0.192 0.165 0.155 0.161 0.162 0.160 0.156 0.172 0.179 0.251 0.279 0.284 0.523

0.222 0.325 0.428 0.385 0.423 0.491 0.411 0.635 0.488 0.675 0.502 0.793 0.621 0.563 0.477 0.647 0.646 0.645 0.477 0.540 0.508 0.556 0.452 0.644 0.446 0.372 0.446 0.452 0.597 0.417 0.448 0.504 0.794 0.644 0.748 1.596

– – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – –

– – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – –

4.35 5.87 6.50 6.33 6.68 7.41 7.10 10.11 8.62 10.84 8.81 13.18 11.19 8.86 8.51 11.68 11.75 12.37 8.92 9.49 10.14 10.39 8.77 11.21 8.64 9.29 9.43 9.78 11.04 9.50 10.57 10.45 13.99 15.02 16.54 27.5

121


Table A1. (Continued ) Decade

Bread (d./lb)

Barley (s./Bushel)

Oatmeal (d./lb)

Peas (s./Bushel)

Potatoes (s./cwt)

Rice (d./lb)

Starches (1860s ¼ 100)

1560 1570 1580 1590 1600 1610 1620 1630 1640 1650 1660 1670 1680 1690 1700 1710 1720 1730 1740 1750 1760 1770 1780 1790 1800 1810 1820 1830 1840 1850 1860

0.526 0.593 0.693 1.036 1.006 1.188 1.177 1.424 1.448 1.374 1.365 1.301 1.128 1.461 1.119 1.343 1.288 1.067 1.100 1.281 1.405 1.661 1.693 2.062 3.010 3.341 2.226 2.080 2.057 1.976 1.951

0.997 1.021 1.137 1.752 1.635 1.967 1.886 2.627 2.410 2.235 2.259 2.091 1.992 2.113 1.968 2.268 2.416 2.035 1.964 2.195 2.470 2.824 2.846 3.760 5.436 5.650 3.962 3.931 3.973 4.125 4.490

0.561 0.544 0.609 0.890 0.760 0.863 0.812 1.240 1.033 1.030 1.069 1.026 1.013 1.055 0.947 1.050 1.047 0.896 1.048 1.082 1.226 1.390 1.412 1.674 2.719 2.782 2.280 2.198 2.106 1.863 2.179

1.458 1.427 1.781 2.442 2.260 2.696 2.441 3.462 2.842 3.423 2.727 2.801 2.794 2.903 2.639 3.021 2.940 2.410 2.396 2.661 2.876 3.636 3.589 4.538 6.866 6.578 4.652 4.521 4.534 4.453 4.472

– – – – – – – – – – – – – – – – 6.847 5.804 4.108 8.616 4.573 3.186 3.607 3.690 4.944 5.301 4.828 5.158 6.322 6.964 7.498

– – – – – – – – – – 2.111 2.427 2.832 3.046 2.920 2.597 2.406 2.456 2.639 2.322 1.875 1.643 2.618 2.118 2.935 3.495 3.185 2.462 2.457 1.721 1.379

25.4 28.2 32.9 48.9 47.0 55.4 54.8 67.7 67.1 64.4 63.6 60.9 53.9 67.7 53.2 63.1 61.0 50.7 51.9 60.5 65.4 75.0 77.3 92.8 134.7 147.9 104.1 99.3 102.3 99.5 100.0

Decade

Beef (d./lb)

Cattle (s.)

Mutton (d./lb)

Pork (d./lb)

1200 1210 1220 1230 1240 1250 1260 1270

0.170 0.190 0.225 0.238 0.277 0.251 0.291 0.336

5.40 6.03 7.17 7.56 8.80 7.99 9.26 10.70

0.175 0.179 0.239 0.272 0.248 0.218 0.242 0.301

– – – – – – – –

Eggs Hens (d.) Meat (d./Dozen) (1860s ¼ 100) 0.260 0.260 0.320 0.493 0.402 0.404 0.424 0.506

1.84 2.62 1.83 2.31 2.38 2.31 2.58 2.60

2.58 2.79 3.32 3.79 3.85 3.50 3.91 4.63

122

GREGORY CLARK


Beef (d./lb)

Cattle (s.)

Mutton (d./lb)

Pork (d./lb)

1280 1290 1300 1310 1320 1330 1340 1350 1360 1370 1380 1390 1400 1410 1420 1430 1440 1450 1460 1470 1480 1490 1500 1510 1520 1530 1540 1550 1560 1570 1580 1590 1600 1610 1620 1630 1640 1650 1660 1670 1680 1690 1700

0.305 0.298 0.336 0.449 0.462 0.397 0.336 0.362 0.455 0.503 0.419 0.428 0.442 0.410 0.407 0.403 0.375 0.379 0.426 0.428 0.406 0.341 0.440 0.521 0.549 0.615 0.663 1.189 1.247 1.359 1.595 2.023 2.110 2.539 2.498 2.692 2.893 2.907 3.000 2.911 2.953 3.003 2.927

9.69 9.48 10.70 14.27 14.69 12.61 10.70 11.50 14.47 15.99 13.32 13.62 14.07 13.03 12.94 12.81 11.93 12.05 13.56 13.61 12.86 10.84 13.38 16.22 17.68 19.13 21.82 41.60 45.22 63.56 66.12 84.81 83.57 – – – – – – – – – –

0.291 0.286 0.286 0.361 0.411 0.290 0.259 0.300 0.382 0.445 0.367 0.335 0.398 0.370 0.318 0.385 0.356 0.318 0.421 0.395 0.414 0.375 0.395 0.446 0.539 0.596 0.717 1.017 1.315 1.475 1.821 2.356 2.594 2.744 2.855 2.905 3.236 3.588 3.545 3.341 3.435 3.481 3.199

– – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – 3.198 3.390 2.776 3.707 3.241 3.170 3.713 3.532

Eggs Hens (d.) Meat (d./Dozen) (1860s ¼ 100) 0.489 0.469 0.523 0.628 0.604 0.554 0.563 0.584 0.580 0.735 0.700 0.684 0.671 0.656 0.649 0.699 0.682 0.703 0.668 0.700 0.707 0.832 1.000 0.865 1.121 1.198 2.200 3.35 3.67 3.70 3.21 3.87 3.93 4.20 3.90 4.76 4.13 4.62 4.85 5.03 5.60 6.05 6.11

2.74 3.01 3.16 3.57 3.81 3.54 3.21 3.65 4.10 4.81 4.73 4.78 4.54 4.71 4.56 4.64 4.65 4.86 4.81 4.83 4.63 5.23 5.77 5.26 6.36 7.32 7.99 10.63 11.98 13.37 12.17 12.99 12.08 – – – – – – – – – –

4.41 4.35 4.63 5.89 6.28 5.08 4.51 5.01 6.12 7.05 6.03 5.86 6.30 5.95 5.56 6.03 5.67 5.48 6.38 6.24 6.23 5.71 6.65 7.32 8.41 9.24 11.14 17.4 20.1 22.1 25.0 31.2 33.1 37.7 37.7 40.3 43.3 43.9 47.2 44.8 45.7 48.0 45.9

123



Beef (d./lb)

Cattle (s.)

Mutton (d./lb)

Pork (d./lb)

1710 1720 1730 1740 1750 1760 1770 1780 1790 1800 1810 1820 1830 1840 1850 1860

3.082 3.063 2.910 3.194 3.198 3.474 4.019 4.157 5.240 7.882 8.306 7.065 6.290 6.001 6.019 6.815

– – – – – – – – – – – – – – – –

3.269 3.192 3.042 3.281 3.303 3.501 4.000 4.066 5.118 7.518 7.943 6.869 6.408 6.245 6.058 7.062

3.668 3.567 3.194 3.466 3.617 4.031 4.555 4.462 5.380 6.959 7.804 6.266 4.722 6.119 6.310 7.420

Decade

1200 1210 1220 1230 1240 1250 1260 1270 1280 1290 1300 1310 1320 1330 1340 1350 1360 1370 1380 1390 1400 1410 1420 1430

Cheese (d./lb) 0.388 0.400 0.420 0.499 0.481 0.469 0.524 0.552 0.508 0.552 0.572 0.688 0.650 0.583 0.505 0.625 0.597 0.520 0.519 0.560 0.523 0.546 0.537 0.607

Eggs Hens (d.) Meat (d./Dozen) (1860s ¼ 100) 6.39 6.36 6.42 6.25 6.55 6.68 7.60 8.39 8.60 11.60 14.99 11.68 11.26 8.61 8.43 10.65

– – – – – – – – – – – – – – – –

Butter Milk Dairy Fats Herring (d./lb) (d./Gallon) (1860s ¼ 100) (1860s ¼ 100) (s./100) – – – – – – 0.877 0.874 0.824 0.863 0.891 1.234 1.200 1.008 1.071 1.075 1.261 1.285 1.122 1.105 1.006 1.149 1.080 0.953

– – – – – – – – 1.01 1.32 1.43 1.54 1.54 1.38 1.21 1.24 1.23 1.07 1.05 1.01 2.22

6.08 6.27 6.56 7.82 7.54 7.40 8.16 8.44 7.84 8.40 8.82 10.85 10.82 9.82 9.23 9.45 10.12 9.77 9.10 9.13 8.54 9.01 9.36 9.07

8.2 7.4 15.5 12.7 10.9 11.1 11.2 11.9 11.2 10.5 12.7 16.4 15.1 13.0 12.1 13.6 14.1 15.4 13.1 12.2 14.0 15.4 14.6 14.0

0.280 0.499 – – 0.522 0.503 0.519 0.457 0.651 0.650 0.726 0.855 0.884 0.832 0.899 1.339 1.180 1.163 1.141 1.440 1.254 1.432 1.431 1.245

47.8 47.0 44.5 47.9 48.6 52.4 60.0 61.5 75.1 107.7 117.3 98.3 86.5 86.1 86.1 100.0 Salt Cod (s/Fish) – – – – – – – – – – – – – – – – – 1.00 – – – 0.95 1.12 0.97

124

GREGORY CLARK


1440 1450 1460 1470 1480 1490 1500 1510 1520 1530 1540 1550 1560 1570 1580 1590 1600 1610 1620 1630 1640 1650 1660 1670 1680 1690 1700 1710 1720 1730 1740 1750 1760 1770 1780 1790 1800 1810 1820 1830 1840 1850 1860

Cheese (d./lb) 0.530 0.524 0.475 0.469 0.533 0.476 0.501 0.772 0.740 0.707 1.699 2.745 2.551 2.582 2.702 3.897 3.464 3.792 3.796 4.255 4.663 3.701 3.922 3.989 3.081 3.683 2.962 3.124 3.094 2.870 3.113 3.095 3.256 3.871 3.741 4.505 6.901 7.411 6.370 5.977 6.228 5.933 6.851

Butter Milk Dairy Fats Herring (d./lb) (d./Gallon) (1860s ¼ 100) (1860s ¼ 100) (s./100) 1.098 1.193 1.190 1.186 1.181 1.331 1.220 1.304 1.172 1.139 1.779 3.50 3.12 3.29 3.31 3.70 4.20 4.47 4.38 4.95 5.28 5.72 5.72 5.64 5.37 5.54 5.12 4.86 4.96 5.22 5.87 6.31 6.37 6.79 6.73 7.85 10.99 12.95 10.41 10.20 10.32 10.20 11.88

1.00 1.00 1.47

1.48 1.07 1.09 1.43 1.59 1.94 3.24 3.77 3.30 3.26 3.64 3.35 4.21 4.63 4.64 5.05 6.42 5.97 6.29 8.65 5.59 5.43 5.10 5.54 5.24 5.17 5.11 5.34 6.70 7.51 8.73 12.27 16.34 15.89 13.36 11.87 9.79 11.86

8.86 9.08 9.05 8.90 9.48 10.03 9.11 10.00 10.73 10.79 16.13 29.2 30.3 29.3 30.0 35.1 34.9 40.1 41.0 44.5 48.3 50.2 50.7 52.9 53.3 48.4 44.1 42.5 44.2 43.6 46.5 47.7 49.2 56.8 58.5 68.6 98.0 118.1 103.0 94.6 92.2 84.9 100.0

11.4 11.7 10.8 10.3 11.8 10.6 9.7 9.9 10.2 11.4 14.8 27.4 29.9 33.4 38.1 48.2 51.3 54.7 51.9 56.0 58.2 61.5 63.3 58.5 54.9 56.6 54.6 54.5 56.2 55.4 57.6 55.6 56.0 57.8 58.2 53.9 78.9 87.6 82.1 80.7 84.9 89.6 100.0

1.342 1.314 1.835 1.491 1.294 1.389 1.187 1.229 1.244 1.614 2.242 2.72 3.03 3.12 2.56 3.12 3.32 3.82 4.43 3.79 5.77 3.39 – – 3.75 4.43 – – – – 2.67 – 2.50 5.00 – – – – – 4.43 4.03 4.60 5.45

Salt Cod (s/Fish) 1.30 1.05 0.95 1.10 1.07 0.96 0.58 0.74 0.79 0.84 0.94 1.23 0.94 1.27 1.68 2.10 2.01 2.06 2.16 3.07 3.07 2.87 3.09 3.08 3.25 4.19 5.06 5.19 4.81 4.04 3.84 3.88 4.72 5.09 4.96 5.12 6.59 7.77 7.02 6.05 – 5.25 –

125



1200 1210 1220 1230 1240 1250 1260 1270 1280 1290 1300 1310 1320 1330 1340 1350 1360 1370 1380 1390 1400 1410 1420 1430 1440 1450 1460 1470 1480 1490 1500 1510 1520 1530 1540 1550 1560 1570 1580 1590 1600 1610

Salt Fish Salmon (1860s ¼ (d. Each) 100) – – – – – – – – – – – – – – – – – – – – – 27.9 31.8 23.6 18.7 21.7 23.2 21.3 26.4 27.1 26.9 27.1 29.0 26.3 33.8 48.1 58.5 70.5 65.8 – – –

3.36 5.99 – – 6.28 6.11 6.36 5.49 7.82 7.81 8.72 10.27 10.63 10.00 10.81 16.09 14.18 14.37 13.71 17.31 15.06 16.59 17.49 14.09 15.47 16.71 17.64 17.53 15.19 14.97 12.56 13.74 14.19 16.01 20.17 27.8 27.5 31.8 32.5 40.2 40.9 44.5

Cider Beer (d./Gallon) (d./Gallon)

– 0.438 0.505 0.486 0.545 0.405 0.465 0.718 0.651 0.622 0.827 0.881 0.897 0.808 0.644 0.713 0.902 0.855 0.652 0.702 0.592 0.691 0.596 0.781 0.520 0.601 0.647 0.474 0.620 – – – – – – – – – – – – –

– – – – – – – – – – – – – – – – – – – – 2.58 2.81 3.08 4.00 2.98 2.57 2.50 2.42 2.68 2.85 2.68 2.74 2.76 2.56 2.70 3.14 3.58 3.71 3.80 4.84 5.51 5.85

Wine (s./Gallon)

Tea (s./lb)

Coffee (d./lb)

0.184 0.174 0.257 0.319 0.213 0.152 0.306 0.327 0.263 0.317 0.360 0.415 0.386 0.479 0.615 0.765 0.760 0.774 0.607 0.558 0.610 0.632 0.633 0.632 0.675 0.707 0.701 0.702 0.812 0.862 0.741 0.817 0.934 0.881 0.958 1.19 1.35 1.58 1.99 2.45 2.35 2.62

– – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – –

– – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – –

126

GREGORY CLARK


1620 1630 1640 1650 1660 1670 1680 1690 1700 1710 1720 1730 1740 1750 1760 1770 1780 1790 1800 1810 1820 1830 1840 1850 1860 Decade

1200 1210 1220 1230 1240 1250 1260 1270 1280 1290 1300

Salt Fish Salmon (1860s ¼ (d. Each) 100) – – – – – – – – – – – – – – – – – – – – – – – – – Drink (1860s ¼ 100) 7.54 8.27 10.40 10.74 9.66 9.61 11.38 14.56 12.33 14.64 14.44

49.2 53.5 54.0 48.5 51.8 51.6 54.2 66.3 78.6 80.6 74.6 62.7 57.9 55.5 67.5 74.4 70.8 73.0 94.0 110.7 100.1 82.3 73.9 84.5 100.0 Honey (d./ Gallon) – – 6.07 – – – 6.59 7.88 6.74 7.92 6.99

Cider Beer (d./Gallon) (d./Gallon)

– – – – – – – – – – – – – – – – – – – – – – – – –

5.37 7.03 7.50 7.95 8.09 8.12 8.22 10.00 9.75 10.16 10.87 10.72 10.91 10.94 10.56 11.62 12.17 12.45 17.81 20.71 19.72 20.17 19.15 18.35 20.00

Wine (s./Gallon)

Tea (s./lb)

Coffee (d./lb)

3.20 3.09 3.48 5.73 5.09 5.85 6.34 6.89 8.67 7.63 6.86 7.94 7.33 7.59 9.03 9.66 11.88 13.58 20.49 26.29 21.16 17.32 17.16 18.60 20.45

– – – – – 37.24 22.33 45.77 22.14 26.40 17.43 14.21 12.68 10.73 8.93 10.03 8.07 6.76 7.94 8.00 7.69 5.35 4.71 4.16 3.57

– – – – – – 53.34 – 89.50 58.68 61.07 62.77 66.16 – 64.34 61.42 65.24 60.22 70.86 51.30 50.81 33.93 25.59 20.28 22.13

Currents/ Raisins (d./lb)

Sugar (d./lb)

Treacle (d./ Gallon)

– – – – – – 6.48 – 1.75 – 5.29

– – – – – – 16.49 – 8.67 17.26 11.25

– – – – – – – – – – –

Sugars (1860s ¼ 100) – – – – – – 39.2 46.8 37.8 43.8 42.7

Salt (1860s ¼ 100) – 13.5 13.3 12.4 15.7 16.8 17.8 19.4 17.6 22.6 18.5

127



1310 1320 1330 1340 1350 1360 1370 1380 1390 1400 1410 1420 1430 1440 1450 1460 1470 1480 1490 1500 1510 1520 1530 1540 1550 1560 1570 1580 1590 1600 1610 1620 1630 1640 1650 1660 1670 1680 1690 1700 1710 1720

Drink (1860s ¼ 100) 18.43 16.52 14.70 15.18 20.35 23.05 21.00 15.07 16.07 16.11 16.65 16.99 18.94 17.10 16.46 16.01 15.38 17.11 18.23 16.76 17.42 18.14 16.96 17.98 21.3 24.2 25.8 27.9 34.7 38.4 41.7 40.9 48.6 53.0 63.3 62.0 64.7 66.5 78.2 80.5 84.7 85.8

Honey (d./ Gallon)


Sugar (d./lb)


Sugars (1860s ¼ 100)

7.95 7.70 6.93 8.97 11.60 11.93 13.60 10.55 11.73 13.09 12.56 11.61 11.32 11.71 12.54 12.63 15.25 15.17 14.41 16.58 17.63 17.28 18.74 12.43 – – 40.4 61.0 53.6 60.7 70.8 53.1

– – 3.14 0.80 – – 2.71 3.04 2.69 3.90 2.68 3.15 3.48 2.76 1.94 2.32 2.50 2.17 1.63 1.55 1.27 1.87 1.80 2.42 3.29 3.04 3.92 3.58 4.13 4.50 4.73 4.29 4.38 5.00 6.77 5.56 5.53 5.18 5.69 5.70 5.52 5.33

6.02 9.36 9.11 11.44 29.93 13.23 14.21 11.84 19.43 15.55 11.99 40.00 22.50 19.82 16.10 13.12 10.08 4.99 4.88 3.00 5.10 5.91 7.21 10.01 12.81 11.24 13.25 16.01 12.24 16.16 15.01 13.77 17.96 16.51 15.58 9.14 8.01 7.46 8.11 7.77 6.93 6.60

– – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – –

46.7 42.9 39.3 45.1 63.4 62.3 74.3 62.2 62.7 75.1 65.3 65.3 65.6 63.1 59.7 62.7 69.1 63.1 53.3 49.6 55.0 66.2 73.4 97.8 132.3 118.9 145.0 153.7 144.7 173.8 171.6 155.7 180.9 186.0 206.9 144.3 135.8 126.8 138.3 135.7 126.2 121.0

86.0 – – – – 159.1 73.8 – –

Salt (1860s ¼ 100) 43.4 29.3 24.6 22.4 53.4 46.4 52.6 44.5 38.5 52.0 40.1 39.7 44.7 39.9 39.3 33.2 32.8 44.8 39.6 38.9 45.2 53.0 53.1 60.9 78.3 84.2 120.1 111.0 138.0 120.2 115.1 139.8 188.7 204.5 192.8 174.7 184.7 178.1 258.6 472.9 447.1 434.4

128

GREGORY CLARK


1730 1740 1750 1760 1770 1780 1790 1800 1810 1820 1830 1840 1850 1860 Decade

1200 1210 1220 1230 1240 1250 1260 1270 1280 1290 1300 1310 1320 1330 1340 1350 1360 1370 1380 1390 1400 1410

Drink (1860s ¼ 100)

Honey (d./ Gallon)


Sugar (d./lb)


Sugars (1860s ¼ 100)

Salt (1860s ¼ 100)

82.6 82.9 81.9 79.5 86.9 91.7 91.3 125.6 140.4 130.9 112.7 104.4 98.4 100.0

– 44.8 10.1 47.7 63.7 – – – – – – – – –

5.11 5.36 4.95 5.24 5.35 5.72 7.16 9.46 9.59 8.80 6.61 5.72 6.65 5.06

6.24 6.70 6.66 6.52 6.70 7.19 9.40 10.15 10.65 8.36 7.96 7.25 5.31 4.98

– – – – 1.25 1.29 1.28 1.26 1.28 1.28 1.28 1.28 1.29 1.33

115.3 122.2 117.2 119.2 122.3 130.4 163.8 184.8 191.4 160.1 144.4 130.8 111.0 100.0

359.4 392.1 383.9 385.4 385.4 486.0 622.3 1,340.9 1,617.6 624.4 155.0 140.5 83.2 100.0

Spices (1860s ¼ 100)

Food DE (1860s ¼ 100)

Food COL (1860s ¼ 100)

5.05 6.05 7.09 7.20 7.30 7.62 7.87 10.10 8.93 10.47 9.54 13.03 11.88 9.87 9.51 12.25 12.84 13.20 10.27 10.67 11.13 11.43

4.78 5.79 6.82 6.85 6.97 7.33 7.50 9.71 8.59 10.10 9.13 12.62 11.42 9.43 9.04 11.68 12.22 12.60 9.76 10.13 10.62 10.88

Ginger (d./lb)

Mace (s./lb)

Cinnamon (s./lb)

Pepper (d./lb)

– – – – – – 24.0 – 18.7 – – 14.0 16.0 19.4 30.4 18.0 16.0 18.2 17.0 36.5 23.6 30.2

– – – – – – – – 1.69 2.50 – – – 2.18 – – – 4.17 – – 2.44 2.31

– – – – – – 0.75 – – – – – 1.61 1.79 1.39 – – – – 2.00 1.13 –

– 5.93 9.08 – 14.30 6.64 9.07 10.53 9.14 12.69 10.96 11.26 12.52 12.40 13.59 22.81 13.87 16.76 10.96 13.92 10.77 21.26

– 21.7 33.3 – 52.4 24.3 42.8 47.2 43.1 56.8 51.1 51.7 55.5 57.2 63.2 97.5 70.8 79.2 49.6 63.8 52.7 84.5

129



1420 1430 1440 1450 1460 1470 1480 1490 1500 1510 1520 1530 1540 1550 1560 1570 1580 1590 1600 1610 1620 1630 1640 1650 1660 1670 1680 1690 1700 1710 1720 1730 1740 1750 1760 1770 1780 1790 1800 1810

Ginger (d./lb)

Mace (s./lb)

Cinnamon (s./lb)

Pepper (d./lb)

Spices (1860s ¼ 100)

Food DE (1860s ¼ 100)

Food COL (1860s ¼ 100)

32.0 24.4 14.8 14.9 16.1 20.0 45.3 27.5 20.8 46.0 33.2 27.4 37.0 52.1 48.9 53.7 44.8 32.4 22.7 22.6 18.5 20.7 24.0 17.3 13.7 11.8 12.3 15.7 15.3 14.2 11.1 12.1 10.8 12.9 12.5 13.0 14.7 21.8 30.8 42.2

2.11 2.00 2.38 2.37 2.18 2.32 2.06 2.68 2.97 3.16 3.95 4.50 2.99 8.27 6.40 7.05 6.81 6.26 5.19 4.80 5.61 6.27 6.86 7.34 6.51 6.29 7.11 9.57 9.61 11.06 10.22 9.17 9.36 8.92 8.80 8.59 13.20 22.39 25.32 21.80

– 2.33 1.50 2.50 1.46 1.63 3.31 3.45 2.29 3.19 5.21 5.50 5.53 5.04 14.05 6.24 7.62 5.84 4.28 4.51 4.04 5.35 9.37 4.00 12.29 9.23 11.34 9.89 9.31 9.11 9.36 9.47 9.97 12.56 15.08 18.58 22.09 24.00 19.63 19.77

16.40 13.48 8.71 10.50 12.77 13.27 15.12 11.97 16.68 13.30 20.21 20.88 21.98 28.73 36.89 30.70 37.39 40.68 30.67 25.58 22.12 21.06 25.04 19.14 15.62 15.30 16.53 23.33 18.33 33.67 26.66 20.20 21.79 22.25 22.75 23.92 23.95 27.25 32.82 40.59

72.8 59.4 45.3 50.1 55.2 60.7 69.4 66.8 74.7 78.1 102.4 105.9 116.7 142.5 184.9 155.2 169.0 159.7 117.5 109.2 99.5 107.9 136.3 108.5 96.1 89.2 98.5 124.6 109.4 149.0 125.6 109.9 113.4 121.4 125.4 133.5 149.5 184.7 214.8 247.6

10.37 11.95 10.09 10.50 10.71 10.84 11.84 10.90 11.40 11.72 14.24 15.07 17.36 27.6 27.7 30.1 33.8 45.6 45.8 52.5 52.2 61.7 63.3 63.2 62.5 61.3 57.8 66.9 58.5 64.6 63.5 56.3 58.1 63.0 66.6 75.2 77.9 90.4 127.9 142.7

9.80 11.42 9.56 9.98 10.16 10.28 11.28 10.31 10.84 11.09 13.58 14.42 16.56 26.6 26.5 28.9 32.7 44.7 44.8 51.7 51.4 60.9 62.2 61.9 62.0 60.7 56.7 66.2 56.9 63.8 62.8 55.1 56.7 62.2 66.1 74.8 77.2 89.9 127.8 142.1

130

GREGORY CLARK


1820 1830 1840 1850 1860 Decade

1200 1210 1220 1230 1240 1250 1260 1270 1280 1290 1300 1310 1320 1330 1340 1350 1360 1370 1380 1390 1400 1410 1420 1430 1440 1450 1460 1470 1480 1490 1500 1510 1520

Ginger (d./lb)

Mace (s./lb)

Cinnamon (s./lb)

Pepper (d./lb)

Spices (1860s ¼ 100)

Food DE (1860s ¼ 100)

Food COL (1860s ¼ 100)

46.3 29.8 30.8 31.0

9.07 8.49 10.66 7.30 5.28

16.00 13.27 11.48 9.36 4.68

38.11 23.18 19.59 17.46 15.18

205.7 142.4 134.8 116.0 100.0

110.8 100.6 99.5 95.1 100.0

109.8 100.6 100.3 95.8 100.0

Fuel (1860s ¼ 100)

House Rent (Index)

Tallow Candles (d./lb)

Wax Candles (d./lb)

– – – – – – 6.60 – 8.35 9.40 11.04 9.63 8.49 7.49 7.25 3.90 3.82 3.65 3.99 4.42 4.99 5.09 4.79 4.33 3.90 3.84 3.89 4.04 4.16 4.13 4.23 4.33 4.71

– – – – – – – – 1.26 1.54 1.94 2.04 2.15 1.85 1.71 2.10 2.21 2.14 2.02 1.81 1.85 1.72 1.63 1.62 1.58 1.33 1.37 1.33 1.32 1.17 1.14 1.24 1.25

– – 3.57 – 4.15 – 4.03 3.78 3.36 4.24 3.93 3.90 3.85 4.04 3.48 3.59 4.17 2.88 4.02 3.43 3.18 3.88 3.73 3.31 4.04 3.82 4.35 4.56 4.56 4.57 4.10 4.54 4.95

Firewood (s./Ton)

Charcoal (s./Bushel)

Coal (s./Ton)

– – – – – 2.29 2.63 3.91 2.79 2.99 3.05 3.41 2.98 3.05 2.53 4.64 4.20 4.38 4.06 4.37 4.36 4.17 4.79 4.50 4.29 4.44 4.21 4.08 3.63 3.77 3.41 3.61 3.68

– – – – – 0.29 – 0.61 0.45 0.59 0.55 0.67 0.72 0.65 0.64 1.90 1.40 1.42 1.06 1.07 1.19 1.06 1.04 1.11 1.10 1.20 1.11 0.97 1.03 1.01 1.03 1.16 1.21

– – – – – – – – 1.42 1.47 1.41 1.58 1.52 1.88 1.44 2.23 1.72 1.65 1.15 1.58 3.56 2.83 2.21 2.45 2.49 2.28 2.39 2.46 2.07 1.97 2.14 1.90 2.84

– – – – – 10.21 12.19 18.38 12.78 13.84 14.10 15.83 14.62 15.06 12.73 23.65 20.51 20.77 17.92 20.00 23.79 21.68 22.57 22.38 21.71 22.14 21.32 20.45 18.70 18.84 18.02 18.58 20.50

131



1530 1540 1550 1560 1570 1580 1590 1600 1610 1620 1630 1640 1650 1660 1670 1680 1690 1700 1710 1720 1730 1740 1750 1760 1770 1780 1790 1800 1810 1820 1830 1840 1850 1860 Decade

1200 1210 1220 1230 1240

Firewood (s./Ton)

Charcoal (s./Bushel)

Coal (s./Ton)

Fuel (1860s ¼ 100)

House Rent (Index)

Tallow Candles (d./lb)

Wax Candles (d./lb)

3.74 4.05 5.14 5.29 6.45 7.65 8.35 10.63 11.49 11.76 13.10 15.36 13.77 14.85 15.96 16.13 15.82 15.46 15.84 15.14 14.90 14.34 13.91 14.43 17.41 14.55 14.37 16.88 20.91 20.36 19.63 17.56 16.73 18.28

1.19 1.39 2.08 2.40 2.59 2.98 3.03 3.13 4.20 4.28 4.21 5.33 6.22 6.94 7.11 7.51 7.61 7.86 7.88 7.55 8.38 9.28 9.39 10.04 10.15 9.92 9.83 11.06 11.58 12.50 13.82 12.63 10.66 11.26

2.36 2.65 4.08 4.35 4.92 5.84 6.33 6.95 7.19 7.49 8.73 11.63 10.48 10.61 10.45 9.05 11.92 12.39 11.75 11.24 11.67 12.24 12.69 13.20 13.53 14.02 16.78 23.33 26.18 22.71 18.52 16.30 13.73 13.93

19.85 21.98 29.7 31.4 37.0 43.8 47.3 56.2 62.0 63.7 70.5 85.9 80.6 85.9 89.7 87.9 93.3 94.1 93.3 89.2 91.0 92.2 92.2 96.1 103.7 99.9 110.5 143.4 165.1 151.3 128.2 113.4 97.4 100.0

4.80 5.52 6.9 8.8 11.0 13.5 15.7 17.0 20.6 20.6 24.2 19.3 19.5 22.5 27.0 29.6 26.6 32.2 28.7 32.3 31.5 29.6 30.5 34.7 38.3 38.7 49.5 70.9 85.3 86.8 85.2 83.8 89.5 100.0

1.38 1.57 2.24 2.88 2.96 3.22 4.02 4.20 4.56 4.64 4.89 5.47 5.39 5.62 5.24 4.88 5.48 4.95 6.08 5.76 5.42 6.54 6.33 6.82 7.15 7.54 8.36 10.73 11.46 7.13 6.18 5.93 6.27 6.40

4.36 3.38 6.83 6.27 6.52 6.35 6.79 7.92 – – 10.76 9.93 13.06 14.59 15.45 12.26 14.54 14.44 18.39 19.61 20.53 20.69 21.53 22.26 24.05 26.19 31.37 49.47 54.14 46.38 34.27 29.79 29.04 35.99

Light (1860s ¼ 100)

Soap (d./lb)

Shoes (s./Pair)

Gloves (d./Pair)

Leather Goods (1860s ¼ 100)

– – – – –

– – – – –

– – – – –

– – – – –

Lamp Coal gas Oil (d./ (s./100 ft3) Gallon) – – – – –

– – – – –

14.09 12.60 26.55 21.69 18.98

132

GREGORY CLARK


1250 1260 1270 1280 1290 1300 1310 1320 1330 1340 1350 1360 1370 1380 1390 1400 1410 1420 1430 1440 1450 1460 1470 1480 1490 1500 1510 1520 1530 1540 1550 1560 1570 1580 1590 1600 1610 1620 1630 1640 1650 1660 1670

Lamp Coal gas Oil (d./ (s./100 ft3) Gallon) – – 5.49 – 10.57 5.74 8.67 8.44 6.99 9.06 11.50 11.11 12.03 11.49 10.71 11.04 10.67 10.27 10.13 11.03 9.53 9.08 9.97 9.48 10.68 10.72 10.82 10.90 11.14 11.81 22.7 38.4 39.9 47.5 47.4 52.1 55.5 48.5 49.0 48.9 64.0 51.9 46.8

– – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – –

Light (1860s ¼ 100) 18.59 21.44 22.13 18.84 22.03 25.52 27.08 27.76 24.56 23.37 28.08 29.50 27.62 27.50 24.61 25.05 24.23 23.00 22.38 22.75 19.67 20.28 20.22 20.03 18.57 17.87 19.45 19.71 20.89 22.54 34.4 43.2 44.8 48.1 60.0 63.3 70.5 68.8 72.1 77.3 81.5 83.9 79.1

Soap (d./lb)

Shoes (s./Pair)

Gloves (d./Pair)


– – 0.57 0.79 1.19 1.08 0.85 1.20 1.21 1.16 1.02 – 1.14 – – – – 1.32 1.52 1.78 1.84 1.65 1.49 1.48 1.74 1.41 1.74 2.01 2.59 1.91 4.06 4.02 3.60 3.49 4.10 3.68 3.79 3.78 4.75 4.94 4.71 4.21 3.71

– – – – – – – – – – – – – – – – – – – – 0.445 – 0.445 0.334 – 0.383 0.358 – 0.452 0.914 1.00 1.30 1.84 1.83 1.68 1.98 2.27 2.30 2.54 2.99 3.33 3.30 2.88

1.67 – – 1.70 1.50 1.50 1.49 1.57 1.63 2.00 2.06 2.03 1.99 1.91 1.92 1.25 1.26 1.75 1.67 1.60 1.40 – 1.00 – 1.75 1.47 1.00 1.00

14.7 – – 15.0 11.9 13.2 13.1 13.9 14.4 17.6 18.2 17.9 17.5 16.8 17.0 11.0 11.1 15.4 14.7 14.1 11.8 – 7.8 5.0 10.3 9.0 7.1 8.8 6.8 13.5 16.4 19.5 27.7 27.6 25.3 29.7 34.1 34.6 38.3 45.0 50.2 49.6 43.4

1.50 3.50 – – – – – – – – – – – –

133



Lamp Coal gas Oil (d./ (s./100 ft3) Gallon)

1680 1690 1700 1710 1720 1730 1740 1750 1760 1770 1780 1790 1800 1810 1820 1830 1840 1850 1860

48.0 66.6 65.2 70.3 54.2 53.1 60.3 58.7 53.5 46.1 56.3 69.1 92.0 106.7 75.1 79.6 83.2 73.4 97.6

Decade

1200 1210 1220 1230 1240 1250 1260 1270 1280 1290 1300 1310 1320 1330 1340 1350 1360 1370 1380

– – – – – – – – – – – – – 15.18 13.31 10.51 7.60 4.76 4.32

Light (1860s ¼ 100)

Soap (d./lb)

Shoes (s./Pair)

Gloves (d./Pair)


74.0 84.2 77.8 95.0 89.8 86.2 100.8 98.5 103.8 107.0 115.2 130.7 173.8 186.6 124.4 118.5 107.5 97.7 100.0

3.97 5.80 4.38 6.03 6.25 5.83 6.81 6.41 6.99 7.12 7.58 8.94 11.09 12.54 9.53 6.87 6.11 5.09 4.91

2.84 3.12 3.06 3.35 3.60 3.63 3.67 4.04 4.01 3.95 3.86 4.20 5.50 5.80 5.96 5.67 5.42 5.42 6.64

– – 6.00 – – – 6.50 7.00 – – – – – – – – – – 10.00

42.8 46.9 46.1 50.5 54.2 54.7 55.2 60.8 60.4 59.5 58.0 63.2 82.7 87.3 89.8 85.4 81.6 81.6 100.0

Silk Thread (d./lb)

Stockings (s./Pair)

Suit of Clothes (s.)

Clothing (1860s ¼ 100)

– – – – – – – – 12.24 – – – – – – – 20.00 – –

– – – – – – – – – – – – – – – – – – –

– – – – – – – – – – – – – – – – – – –

17.30 17.19 15.96 14.79 18.17 16.62 17.96 17.36 17.75 17.18 19.42 21.65 19.97 18.17 16.51 25.33 25.90 26.66 25.48

Linen Cotton Cloth Wool (s./Yard) Cloth Cloth (s./Yard) (d./Yard) – – – – 2.42 2.55 2.44 2.41 2.76 2.11 2.72 3.28 2.74 2.37 2.17 2.89 2.74 2.91 2.73

3.59 3.57 3.32 3.04 3.91 3.46 3.80 3.74 3.47 4.18 4.39 4.69 4.60 4.64 4.09 8.34 9.21 8.83 8.69

– – – – – – – – – – – – – – – – – – –

134

GREGORY CLARK


1390 1400 1410 1420 1430 1440 1450 1460 1470 1480 1490 1500 1510 1520 1530 1540 1550 1560 1570 1580 1590 1600 1610 1620 1630 1640 1650 1660 1670 1680 1690 1700 1710 1720 1730 1740 1750 1760 1770 1780 1790 1800 1810

Wool Linen Cotton Cloth Cloth Cloth (s./Yard) (s./Yard) (d./Yard) 2.52 2.58 2.73 2.59 2.55 2.52 2.33 2.55 2.55 2.69 2.60 2.76 2.86 3.00 3.23 3.48 4.16 5.66 6.03 6.41 6.73 7.63 7.67 7.71 8.31 8.97 9.08 8.72 8.26 7.87 8.36 8.78 8.72 8.66 8.25 8.39 8.04 7.76 7.72 7.67 7.99 8.86 10.03

8.20 7.55 7.03 7.22 7.38 7.12 7.12 7.06 7.43 7.05 6.70 7.44 6.39 6.89 7.33 7.47 10.59 11.50 14.05 15.48 15.91 16.17 16.76 16.84 17.21 17.81 18.49 17.65 16.81 16.64 19.82 21.50 21.89 21.95 21.88 22.42 22.16 20.77 21.26 20.20 20.17 23.37 23.56

– – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – 51.25 43.56 44.02 47.18 53.12 48.46 39.15 38.75

Silk Thread (d./lb)

Stockings (s./Pair)



– – – – – – – – – – 12.50 13.78 10.33 8.65 10.29 – – – – 24.00 24.15 25.90 26.64 29.61 – – 30.00 – – 25.85 29.63 34.52 32.99 32.09 28.10 30.78 26.80 26.95 24.95 21.36 23.53 31.57 40.22

– – – – – – – – – – – – – – – – – – – – – – – – 24.84 25.50 30.20 29.33 21.99 22.53 21.96 20.82 20.54 21.79 20.15 20.23 21.06 21.18 20.54 21.46 22.71 25.23 30.62

– – – – – – – – – – – – – – – – – 9.49 15.06 15.89 17.56 18.43 21.18 25.62 35.19 40.22 36.88 37.71 35.75 34.51 33.82 32.46 34.55 33.33 33.17 34.98 38.53 42.70 41.56 42.19 42.99 49.17 57.11

24.14 23.34 23.73 23.52 23.52 23.34 22.06 22.57 22.79 22.98 22.71 23.45 23.36 24.59 25.98 27.91 35.0 42.8 50.0 53.0 55.4 60.4 65.1 70.1 82.0 90.7 89.1 88.6 82.3 80.1 83.0 82.8 86.2 85.5 84.2 87.1 91.5 94.8 93.6 93.7 96.2 108.8 122.2

135



1820 1830 1840 1850 1860 Decade

1200 1210 1220 1230 1240 1250 1260 1270 1280 1290 1300 1310 1320 1330 1340 1350 1360 1370 1380 1390 1400 1410 1420 1430 1440 1450 1460 1470 1480 1490 1500 1510 1520

Wool Linen Cotton Cloth Cloth Cloth (s./Yard) (s./Yard) (d./Yard) 8.99 8.68 7.45 7.29 8.50

20.72 20.00 15.34 14.13 18.00

Tobacco Books (d./lb) (s./Book)

– – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – –

– – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – –

21.54 16.81 11.59 11.25 16.80

Silk Thread (d./lb)

Stockings (s./Pair)



34.29 22.86 18.84 14.44 –

25.51 23.11 20.77 19.70 21.81

57.29 56.03 56.88 52.77 47.84

114.0 108.6 100.1 95.0 99.9

Silver (d./oz)

Pewter (d./lb)

Brass Goods (d./lb)

Woodwares (Index)

Pottery (d./Plate)

23.8 23.8 23.8 23.8 23.8 23.8 23.8 23.8 23.9 23.9 23.9 23.9 23.9 24.3 25.6 29.2 29.5 29.5 29.5 29.5 29.5 34.2 35.4 35.4 35.4 35.4 39.8 44.3 44.3 44.3 44.3 44.3 47.6

– – 1.85 – – – 1.79 1.70 2.31 2.17 2.20 2.89 2.64 2.64 2.51 3.34 3.55 3.48 3.43 3.18 3.47 3.30 2.87 2.75 3.12 3.12 3.17 3.69 3.58 3.88 4.24 5.13 4.97

– – – – – – – 2.50 2.29 2.28 2.00 3.00 2.53 3.05 2.46 2.87 3.01 3.43 3.84 3.77 3.93 3.61 3.00 3.10 4.02 3.46 3.79 3.99 3.19 4.13 3.87 4.06 3.74

– – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – –

– – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – –

136

GREGORY CLARK


1530 1540 1550 1560 1570 1580 1590 1600 1610 1620 1630 1640 1650 1660 1670 1680 1690 1700 1710 1720 1730 1740 1750 1760 1770 1780 1790 1800 1810 1820 1830 1840 1850 1860 Decade

1200 1210 1220 1230

Tobacco Books (d./lb) (s./Book)

– – – – – – – – – – – – – – – – – – – – – 13.5 13.0 14.8 21.5 34.2 36.9 46.7 56.8 56.9 50.0 48.8 52.1 52.6

– – – – – – – – – – – – – – – – 6.43 4.30 4.50 10.82 5.84 6.05 7.02 7.39 7.58 – 9.66 13.97 14.52 15.14 14.62 13.25 12.26 12.38

Silver (d./oz)

Pewter (d./lb)

Brass Goods (d./lb)

Woodwares (Index)

Pottery (d./Plate)

49.8 52.5 63.8 66.4 66.4 66.4 66.4 68.4 68.6 68.6 68.6 68.6 68.6 68.6 68.6 68.6 69.0 68.3 68.9 69.5 69.2 69.1 70.9 71.6 70.8 70.2 68.0 74.5 76.4 64.5 64.3 64.4 66.2 66.0

4.87 5.26 8.21 8.64 7.68 7.57 8.20 9.52 10.84 11.60 13.62 14.70 14.22 13.81 13.40 12.24 12.06 12.61 12.35 12.19 12.42 11.89 12.36 11.48 12.38 11.57 13.58 16.31 18.25 18.35 15.58 11.42 14.42 –

3.61 4.51 5.14 7.28 7.23 6.42 7.48 8.32 10.46 11.94 13.32 11.17 11.58 11.64 11.64 11.00 11.21 12.95 11.57 13.68 11.29 10.98 11.02 9.39 9.38 – 9.69 5.92 15.59 13.16 11.85 9.87 9.09 11.96

– – – – – – – – – – – – – – – – – – – – – 48.5 56.9 61.8 65.1 65.0 81.7 96.0 128.3 115.6 111.3 122.2 112.8 100.0

– – – – – – – – – – – – – – – – – – – – – 2.08 – 3.05 3.51 2.86 4.52 4.11 4.89 5.17 4.94 3.43 5.28 6.45

Glasswares (d./Quart Battle)

Nails (d./lb)

Manufactured Iron (d./lb)

Spades Shovels (d. Each)

– – – –

1.83 – 3.11 3.40

– – – –

– – 1.60 1.00

Scissors Cutlery (d. Each) (d./Knife)

– – – –

– – – –

Screws (d./ Dozen) – – – –

137



1240 1250 1260 1270 1280 1290 1300 1310 1320 1330 1340 1350 1360 1370 1380 1390 1400 1410 1420 1430 1440 1450 1460 1470 1480 1490 1500 1510 1520 1530 1540 1550 1560 1570 1580 1590 1600 1610 1620 1630 1640 1650


Nails (d./lb)



– – – – – – – – – – – – – – – – – – – – – – – – – – – – – 8.08 – – – 7.00 4.94 2.62 2.62 1.56 3.86 7.09 – –

3.21 2.74 2.65 2.23 2.33 2.26 2.53 2.72 2.61 2.55 2.49 5.11 5.32 4.64 4.10 4.23 3.67 3.63 4.10 3.96 4.18 3.83 3.93 3.43 3.69 3.60 3.31 3.95 2.87 3.08 3.16 3.98 4.37 4.56 5.02 4.77 4.49 4.71 4.13 4.62 4.58 5.12

– – 1.26 1.11 0.95 1.08 0.99 1.21 1.07 1.19 1.31 1.75 1.81 1.97 2.35 1.61 1.74 1.49 1.61 1.41 1.47 1.30 1.28 1.36 1.29 1.38 1.20 1.27 1.60 1.83 2.27 2.79 2.90 2.87 3.14 3.31 4.38 4.52 4.27 4.59 4.86 4.52

1.50 1.75 1.73 1.58 2.02 2.18 2.55 3.37 2.60 8.70 2.60 4.95 5.50 6.70 4.97 3.40 5.03 5.83 3.50 5.70 5.95 5.98 5.73 4.50 5.93 4.32 3.44 5.55 4.37 4.27 6.44 9.07 9.14 11.89 8.06 8.67 12.92 14.36 17.87 21.08 15.80 22.33


– – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – –

– – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – –

Screws (d./ Dozen) – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – –

138

GREGORY CLARK


1660 1670 1680 1690 1700 1710 1720 1730 1740 1750 1760 1770 1780 1790 1800 1810 1820 1830 1840 1850 1860 Decade

1200 1210 1220 1230 1240 1250 1260 1270 1280 1290 1300 1310 1320 1330 1340 1350 1360


Nails (d./lb)



6.36 4.45 3.87 4.88 3.11 – 2.51 2.50 2.84 2.90 3.79 3.70 7.11 2.81 3.23 3.83 4.28 4.57 2.60 2.85 2.80

4.12 4.40 4.84 3.92 4.40 3.81 4.14 3.96 3.76 3.45 3.80 3.98 3.79 4.36 5.05 5.71 3.77 4.33 3.66 3.19 3.72

4.46 4.28 3.84 4.10 4.34 4.16 3.83 3.95 3.50 3.36 3.83 3.76 3.83 4.59 5.05 4.67 3.37 2.92 2.25 2.54 2.23

17.73 21.33 15.72 19.60 21.68 30.83 26.00 36.24 26.03 27.87 37.03 27.37 37.40 54.70 51.43 48.35 41.16 36.26 35.66 30.65 40.83


– – – – – – – – – – 15.36 14.81 17.57 16.56 12.26 26.20 22.04 27.04 21.48 27.02 30.00

– – – – – – – – 4.20 – 4.01 4.30 5.95 6.40 9.16 9.29 8.33 8.29 8.68 9.19 13.35

Screws (d./ Dozen) – – – – – – – – 6.89 – 6.00 5.47 4.81 7.68 8.04 7.37 8.50 8.03 5.06 4.08 5.58

Rope (d./lb)

Paper (d./Quire)

Paint (d./lb)

Manufactured Goods (Index)

Bricks (s./100)

Timber (d./ft3)

Window Glass (d./ft2)

– – – – – – – – – 1.110 0.490 0.950 0.660 0.850 0.500 0.985 1.663

– – – – – – – – – – – – – – – 7.97 10.26

– – – – – – – 2.65 3.52 2.31 3.08 3.42 3.92 4.75 2.48 4.23 2.81

– – 17.7 – 19.6 19.4 19.0 17.5 18.5 18.8 19.3 23.2 21.3 24.4 21.5 30.3 34.3

– – – – – – – – 4.24 4.53 4.82 4.38 3.62 3.88 3.83 6.82 8.57

– – – – – – – – – – – – – – – – –

– – – – – – – – – – 4.97 7.32 4.17 – 2.53 – 13.00

139



1370 1380 1390 1400 1410 1420 1430 1440 1450 1460 1470 1480 1490 1500 1510 1520 1530 1540 1550 1560 1570 1580 1590 1600 1610 1620 1630 1640 1650 1660 1670 1680 1690 1700 1710 1720 1730 1740 1750 1760 1770 1780 1790

Rope (d./lb)

Paper (d./Quire)

Paint (d./lb)


Bricks (s./100)

Timber (d./ft3)


1.375 1.875 – 1.700 1.363 1.536 1.420 1.413 1.500 1.556 1.293 1.260 1.276 1.160 1.262 1.447 1.753 1.596 3.412 3.671 3.156 2.804 3.224 3.337 4.103 4.494 6.407 5.734 6.437 7.320 7.645 7.166 7.371 6.208 6.765 7.476 6.833 5.973 6.767 6.073 6.784 8.082 9.077

9.71 – 7.44 6.75 5.58 4.92 4.61 4.26 4.38 3.92 4.41 4.10 3.69 3.52 3.43 3.75 3.55 4.17 5.36 5.98 6.25 5.74 5.67 6.02 5.84 6.59 6.47 7.78 11.11 12.07 7.91 9.40 13.47 13.07 13.36 12.89 12.88 13.45 10.35 11.13 10.48 11.84 12.67

5.03 4.93 4.64 – – 10.54 – – 1.55 1.38 – 3.49 3.75 3.10 4.70 6.25 4.08

34.9 33.0 30.9 32.1 31.1 29.7 29.8 31.7 30.6 30.9 30.6 30.4 30.1 28.3 31.6 31.5 33.4 38.1 51.6 56.7 57.9 53.9 56.2 63.4 69.1 73.4 84.0 81.0 89.0 86.6 83.3 77.9 83.3 83.3 83.7 85.1 82.4 77.6 77.0 80.1 78.3 87.1 96.4

7.92 7.26 7.97 9.00 8.15 9.57 8.36 9.04 8.45 7.60 7.03 7.43 7.93 7.91 7.91 7.66 7.39 9.04 10.98 17.46 17.43 15.57 15.60 16.31 17.27 16.76 17.33 17.89 22.92 23.76 22.41 23.89 22.77 22.27 24.61 26.36 23.27 28.00 26.15 28.86 28.72 31.99 44.35

– – – – – – – 1.53 1.80 0.53 1.82 1.83 1.81 1.43 1.55 1.28 1.73 1.92 2.97 3.02 3.17 2.85 3.18 4.65 6.45 6.50 7.37 8.04 8.73 10.72 9.59 10.67 10.15 10.00 9.51 8.81 7.95 7.76 8.88 9.34 9.91 9.63 12.98

– 8.00 9.00 11.00 9.38 8.00 9.63 10.00 6.05 6.75 4.39 6.33 – 6.31 4.61 4.96 4.46 4.51 6.92 6.64 6.50 6.34 5.96 5.84 5.83 5.07 5.52 6.30 6.42 5.78 5.32 6.60 6.91 6.55 6.11 6.53 5.88 6.11 8.30 7.43 10.01 12.70 10.18

2.73 4.01 6.78 1.86 – 9.09 10.53 6.50 6.08 5.42 6.97 – – 6.24 3.34 6.41 16.30 21.17 – 3.18 6.63 6.04 3.53 5.45 5.43

140

GREGORY CLARK


Rope (d./lb)

Paper (d./Quire)

Paint (d./lb)


Bricks (s./100)

Timber (d./ft3)


1800 1810 1820 1830 1840 1850 1860

11.315 19.020 15.020 5.780 12.000 8.095 9.510

19.95 21.81 22.59 19.91 16.59 14.35 12.14

8.68 13.00 6.84 4.06 4.37 3.93 5.05

106.1 128.6 114.4 102.9 87.6 88.9 100.0

60.94 68.88 64.10 58.26 53.44 45.62 48.35

22.17 25.31 17.11 15.53 12.42 9.07 9.43

14.21 17.42 19.89 17.32 14.40 7.66 11.90

Note: In addition to the sources listed in the paper, silver prices were derived from Jastram (1981), and book prices by the method detailed in Clark and Levin (2001).

CAPITAL ACCUMULATION IN THE LONG RUN: THE CASE OF SPAIN, 1850–2000 Leandro Prados de la Escosura and Joan R. Rose´s ABSTRACT New series of Spain’s capital stock and input are provided for the last one and a half centuries for the first time. Capital stock and input grew at average rates of 3.5 and 3.7 percent per year, respectively, but not at a steady pace since rates accelerated dramatically during the ‘‘Golden Age.’’ Two major structural changes accompanied this process. The composition of the capital stock and returns to it changed gradually as the contribution of producer durables rose while that of structures declined. Capital deepening took place in the long run. Although the capital–output ratio increased over time, in phases of accelerated growth the productivity of capital rose. Economic historians have long been interested in establishing the role of capital accumulation in long-term development. For example, was the shift to a higher rate of investment a feature of early economic growth (Rostow, 1956)? How strategic a factor has machinery investment been for economic growth (De Long & Summers, 1991)? A major obstacle to providing a satisfactory answer to these queries is the lack of long-run series for capital across countries. Research in Economic History, Volume 27, 141–200 Copyright r 2010 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 0363-3268/doi:10.1108/S0363-3268(2010)0000027005

141

142

LEANDRO PRADOS DE LA ESCOSURA AND JOAN R. ROSE´S

In this paper we present, for the first time, homogeneous series of capital stock and input series for Spain over the last one and a half centuries. In Methods and Data Sources section we discuss the methods and data sources used to derive the new series of capital stock and input, while in Sensitivity Tests section the sensitivity tests for alternative methods of computing capital stock are provided including a comparison between our new estimates and earlier ones. Historical trends in capital stock and input are offered in Trends in Capital Stock and Input section. Finally, in Capital Deepening and Productivity section, capital deepening and productivity in Spain are discussed in international perspective. We conclude that capital stock and input grew at a substantial rate but not at a steady pace, and its pace accelerated dramatically during the ‘‘Golden Age.’’ Major structural changes accompanied this process. The composition of the capital stock and the returns to it changed gradually as the contribution of producer durables rose, while that of structures declined. The Spanish economy experienced capital deepening in the long run. However, although the capital–output ratio increased over time, in phases of accelerated growth it declined, that is, the productivity of capital rose.

METHODS AND DATA SOURCES Our approach to measuring physical capital accumulation in Spain follows the method developed by Jorgenson (1989, 1990) and Hulten (1990) and, hence our measure, capital input, is not necessarily identical to the one usually employed in national accounting.1 Conventionally the stock of capital is defined as all tangible goods that can be used during more than one period to produce other goods and services. More specifically, the capital stock comprises residential and nonresidential structures, transport equipment, and other producer durables (machinery and equipment).2 The input of capital, which is an index number of the flow of services provided by the stock of capital, is determined, then, by the size of the capital stock, and the returns to it, which is the capital remuneration (property compensation) in production outlay. We have proceeded as follows: first, we constructed the stock of capital; then, we estimated the rental price of capital (or price of capital services) and the total returns to capital (the value of capital services). Finally, we weighted the quantity of each asset by its share in the total returns to capital in order to derive a single capital input index.

143

Capital Accumulation in Spain

Capital Stock National accounts document flows of new capital to be added to the actual stock in a year (It), but do not record the actual amount of capital stock in use (Ct). Since capital stocks result from the accumulation of investment flows, social accountants developed the perpetual inventory method (PIM) to infer the capital stock from past years’ additions to capital assets.3 Thus, the stock of capital, Ct, evolves according to the value, at constant prices, of the new investments during that year and depreciation of the existing stock. C t ¼ ð1 dt ÞC t1 þ I t

(1)

where the capital stock C in the year t equals the amount of existing capital in the year t1 multiplied by 1 minus the depreciation rate (d) in the year t, plus gross fixed capital formation, I, in the year t. The use of the PIM method to compute capital stock series requires: (a) historical series of gross fixed capital formation (GFCF) by type of asset, at constant prices; (b) an initial benchmark for the stock of capital; and (c) the efficiency of each vintage of capital. (a) Disaggregated volume and price series of GFCF by asset type (residential buildings; other structures; transport equipment; and machinery and equipment) from 1850 to 2000 are available (Prados de la Escosura, 2003). (b) However, there is no capital stock (C) for each type of asset j at the initial year 1850. In order to derive the initial stock we require information on investment levels and growth rates for the previous years, as well as depreciation rates.4 Specifically, defining investment at the year t as: I jt ¼ ðd þ gÞC jt1

(2)

where d is the depreciation rate, g the rate of increase of the capital stock, and Ct1 the capital stock in the previous year. Then, solving for the capital stock, the following equation is obtained: C jt1 ¼

I jt ðd þ gÞ

(3)

In order to compute C for the year 1849, investment levels (It) for 1850 and depreciation rates (d) for each asset type (j) (see Table 2) are available, while the yearly rates of increase for the capital stock (g) could be proxied by the average growth rate of GFCF in the 1850s.5

144


A caveat is necessary though. It seems plausible enough that the growth of investment was significantly slower in the early nineteenth century than in the 1850s (the decade in which railway construction started in Spain) and that, consequently, by following this procedure we may bias downward the initial level of the stock of capital (and, subsequently, bias upward the growth rate of the capital stock).6 As a way to mitigate these biases, we have arbitrarily assumed an initial capital stock twice as high as the one derived with Eq. (3) (see the next section for a sensitivity test). We have also taken into account the destruction of capital assets as a consequence of the Spanish Civil War (1936–1939). Capital series derived through the PIM method capture the decline in their level over 1935–1940 for some assets. This is the case, for example, for machinery and equipment, merchant shipping, buildings, railways, and roads. Unfortunately, no exhaustive census on war destruction exists and, for this reason, we had to resort to many individual studies for the rest of them.7 We started from the available estimates of destruction for specific assets that have been distributed at an annual cumulative rate over 1936–1939.8 The total war destruction was equivalent to 7 percent of the capital stock and, if dwellings are excluded it reached one-fourth of the so-called productive capital stock on the eve of the Civil War (1935) because destruction was disproportionately concentrated in transport equipment (40 percent) and to a lesser extent in machinery and equipment (13 percent), as opposed to buildings and infrastructure that escaped relatively unscathed (with 4 and 6 percent losses). Thus, the Civil War had a deeper impact on capital input (that is, the service provided to production by the stock of capital) than on the capital stock itself. Interestingly, the destruction of capital in Spain during the Civil War would be at the lower bound of World War II destruction in percentage terms: comparable to that of France (8 percent) but much lower than in Germany (16 percent) or Japan (26 percent) (Maddison, 1991, pp. 284–292). (c) Social accountants resort to indirect methods to infer the efficiency of capital units (Hulten, 1990). A widely employed procedure is to assume that all efficiency patterns (F) follow a pattern determined by the observable lives (T) of capital goods. Among different F, the simplest one assumes that capital goods maintain their efficiency at constant levels across their lives. The problem would, then, be reduced to estimating the useful lives of capital assets (T). The depreciation rate (d) is inversely related to the asset life, so d ¼ X/T, where X is a parameter (declining balance) and T is the life of each type of asset.9 In our case, following Jorgenson (1990), we adopted

145


the ‘‘modified’’ geometric depreciation pattern, in between the arithmetic and geometric patterns.10 Hulten and Wykoff (1981) suggest the parameter X gets values of 1.65 for machinery and equipment and 0.91 for buildings and structures.11 The useful lives assumed for each type of asset derive from available information for Spain, the United States, and Britain (Table 1), are in line with those used in major historical works (Feinstein, 1988; Jorgenson, 1990), and tend to be on the conservative (high) side when compared with available studies for late-twentieth-century Spain. In the case of ‘‘productive’’ capital (namely, nonresidential structures, transport equipment, and other machinery and equipment), as asset lives tend to shorten as one gets closer to the present, different service lives have been attributed to assets during three distinct epochs (1850–1919, 1920–1959, and 1960–2000).12 In the years 1920–1959, which include the interwar and the autarchic periods, the renewal (by replacement) of old capital vintages was hampered by Table 1. Assets Lives Estimates. Dwellings

Myro Hofman Ivie MOISSES Cubel and Palafox Jorgenson (US) Feinstein (UK) Prados de la Escosura and Rose´s

1965–1981 1950–1992 1964–2000 1954–1995 1901–1958 1850–1920 1850–1919 1920–1959 1960–2000

50 50 50 30 50 70 100 70 70 70

Other Transport Constructions Equipment 36 40 20 50 40 80 55.7 54.7 40

10 15 10 10 25 10–20 36.9 27.9 15

Machinery and Equipment 15 15 15 10 25 25–40 30 20 15

Note: Different service lives for different assets within each main type of asset and composition changes reflect in the service lives provided for other constructions 55.7 (54.7) and transport equipment 36.9 (27.9) years in 1850–1919 (1920–1959). For example, the service life of 36.9 years in transport equipment over 1850–1919 captures the composition changes within this broad type of capital asset and the different service lives accepted for railways (40 years), shipping (30), and road (10) equipment. Other constructions include nonresidential buildings for which a service life of 70 years has been assumed, while an average of 50 years has been accepted for the rest of nonresidential constructions. Sources: Myro (1983); Hofman (1993); Ivie; Fundacioń BBV (1995); Mas, Pe´rez, and Uriel (2005a); MOISSES, kindly provided by Antonio Dı´ az Ballesteros; Cubel Montesinos and Palafox Ga´mir (1997); Jorgenson (1989); and Feinstein (1988). See text.

146


Table 2.

Dwellings Other constructions Transport equipment Machinery and equipment

Depreciation Rates. 1850–1919

1920–1959

1960–2000

0.0130 0.0163 0.0447 0.0550

0.0130 0.0166 0.0591 0.0825

0.0130 0.0228 0.1100 0.1100

Note: The depreciation rate, d ¼ X/T, where X is a parameter (declining balance) and T the life of each type of asset. X is 1.65 for transport equipments and machinery and equipment and 0.91 for dwellings and other structures. Depreciation rates for other structures and transport equipment are weighted average of those of their components. Thus, for example, 0.0447 for transport equipment in 1850–1919 represents a weighted average of those for railways (0.0413), shipping (0.0550), and road (0.1650) equipment. See note in Table 1. Sources: X, Hulten and Wykoff (1981); T, Table 1. See text.

restrictions on international trade and factor mobility, and war, and this helps to explain why useful lives were longer than from 1960 onwards, when the growing integration of Spain into the international economy justifies the assumed reduction in the service lives of assets. Using the service lives presented in Table 1 we derived the depreciation rates to be used in our calculations, by asset type and period (Table 2).

Capital Input The input of capital can be defined as the flow of services provided by the stock of capital to production.13 Thus, in addition to the stock of capital, we need estimates of the rental price of capital (or price of capital services) and of total returns to capital (or value of capital services). Then, we will weigh the quantity of each asset by its share in the total returns to capital in order to derive a single capital input index. In competitive equilibrium, the cost of producing a unit of capital is equal to its price which is equal to the discounted present value of its rents during its life. If we assume that old and new vintages of capital are perfect substitutes, the rental price of capital, pk, in year (t), can be estimated as, pk ðtÞ ¼ Pit1 rt þ dpit ½Pit Pit1

(4)

where pi is the investment price of the capital good i, r the nominal rate of return, and d the depreciation rate for the capital good i.14 The rental price


147

of capital asset i is, thus, the sum of return per unit of capital, Pit1rt; depreciation, d pit; and the negative of revaluation, [pit pit1]. We have already established the depreciation rates and the acquisition price of capital, but we do not know the rates of return (r).15 There are two alternative methods to impute the nominal rate of return. The first uses the long-run interest rate as equivalent to the rate of return to capital under perfect competition. The second derives the rate of return from the share of national income received by the owners of capital assets as a compensation for their property.16 In order to maintain the consistency with our perfect competition assumptions we have used the competitive rate of return. The long-run interest rate was used to approximate the rate of return on capital under perfect competition. As a proxy for the long-term interest rate the internal rate of return for private assets that comes from the MOISSES and BDMORES databases (Dabań Sańchez, Dı´ az Ballesteros, Escriba´ Pe´rez, & Murgui Garcı´ a, 1998) was used since 1954, while the corporate rates of return were employed for 1880–1954 (Tafunell, 2001), and the net rate of return on public debt for 1850–1880 (Tafunell, 2005). Multiplying the rental price of capital asset i by the quantity of capital stock i we obtained the returns to capital asset i. Adding up the returns to each type of asset we derived the total returns to capital that equals capital property compensation. The share of each type of asset in total returns to capital was used as weights in the computation of the capital input index. A capital good with a higher amortization rate receives a larger weight in the index of capital input; a million dollar worth of machinery, for example, is allocated a higher weight than a million dollar worth of dwellings (compare Figs. 6 and 8). The implication is that changes in the stock composition from long duration (and low rate of return) to short duration (and high rate of return) capital goods represent an increase in the quality of capital. The final step is to construct a capital input index by combining the quantity of each asset with its share in total returns to capital. To construct yearly indices, we expressed capital input at year t (Kt), as a translogarithmic function of its four components (residential and nonresidential structures, transport equipment, and machinery and equipment). The corresponding translogarithmic capital input index, under the assumption of constant returns to scale, is 2 3 1 K 2 K K K 6 a1 ln K 1 þ a2 ln K 2 þ þ an ln K n þ 2 b11 ðln K 1 Þ 7 7 (5) K ¼ exp6 4 5 1 K K 2 þb12 ðln K 1 Þðln K 2 Þ þ þ bnn ðln K n Þ 2

148


If we take log first differences in Eq. (5), we get the growth of aggregate capital input between two periods as a weighted average of the growth rates of its n quality classes: X yni ðln K it ln K it1 Þ (6) ln K t ln K t1 ¼ i where share values are computed as: 1 yni ¼ ½ynit þ ynit1 ; 2

ði ¼ 1; . . . ; nÞ

(7)

yi denotes the elasticity of aggregate capital input with respect to each asset type and, under the assumption of perfect competition, equals the remuneration of each type of asset in GDP (property compensation). These series, expressed in first differences, can be converted into a yearly index by taking its exponential. The ratio between the capital input and the capital stock provides a measure of capital’s composition changes, or ‘‘quality’’ of capital. However, the idea that technological change embodied in capital is captured by increases in the ‘‘quality’’ of capital lacks consensus and has been often rejected.17

SENSITIVITY TESTS How robust are the new capital estimates to alternative computation methods? In particular, are they sensitive to alternative initial capital values, price indices, and depreciation rates? First, how robust are our capital stock series to alternative assumptions about its initial level? Following Young (1995), we obtained capital stock series with alternative initial values for 1850. Fig. 1 shows the capital stock series resulting from the initial stock yielded by Eq. (3) and the assumptions that it represented double such a level (the one we favor) and simply zero. It can be noticed that the effect derived from choosing these alternative initial capital levels fades away over time and the resulting series converge by 1890. Thus, over the years 1850–1883, the growth rate of capital stock would have been 5.3 percent assuming that its initial level (1850) was the one derived from Eq. (3), while with our favored estimate, that is, assuming twice as much this initial level, the rate of growth becomes 3.6 percent. In the following long swing (1883–1920), the divergence is sharply reduced: with the alternative growth rates being 2.5 and 2.3 percent, respectively. Cautious users are, nonetheless, advised to use the series starting in 1890.

149

Capital Accumulation in Spain 1000000.0

100000.0

10000.0

18

5 18 0 52 18 5 18 4 5 18 6 58 18 6 18 0 6 18 2 64 18 66 18 6 18 8 70 18 7 18 2 7 18 4 7 18 6 78 18 8 18 0 8 18 2 84 18 8 18 6 8 18 8 90 18 9 18 2 94 18 96 18 9 19 8 0 19 0 02 19 0 19 4 0 19 6 0 19 8 10 19 12

1000.0

Zero Initial Capital Stock

Fig. 1.

Double Initial Capital Stock

Initial Capital Stock (expression 3)

Capital Stock Estimates with Alternative Initial Levels, 1850–1913 (1958 Million Pesetas) (Semilog Scale).

Table 3.

Perpetual Inventory Method Versus Direct Estimation in 1965 (000 Million of 1965 Peseta).

Dwellings Other constructions Transport equipment Machinery and equipment Aggregate capital stock

Prados de la Escosura and Rose´s (PIM)

Universidad Comercial de Deusto (UCD)

Ratio (PIM/UCD)

994.1 961.8 201.6 578.9 2,736.6

1,166.0 1,235.7 194.3 633.3 3,229.3

0.85 0.78 1.04 0.91 0.85

Sources: PIM, see text; Universidad Comercial de Deusto (UCD) derived by Myro (1983, Table 2.3).

Since in the construction of yearly capital series through the PIM approach it is quite common to initialize the capital stock series using a benchmark capital stock, we carried out a comparison between the capital stock resulting from a wealth survey for 1965 by Universidad Comercial de Deusto (UCD) (1968–1972), usually employed to anchor annual series of capital, and the estimate we obtained for the same year through the PIM method (Table 3).18 It can be observed that the UCD estimates tend to


150

exaggerate the size of capital stock, both for the economy as a whole, and by type of asset, with the exception of transport equipment.19 How sensitive are capital stock estimates to the degree of disaggregation of the GFCF series used in its construction? Since 1970, Spanish national accounts (CNE70) have distinguished four types of assets (dwellings, other construction – including nonresidential buildings, transport equipment, and machinery and equipment) while previously national accounts (CNE58) allocated residential and nonresidential buildings to the same category. Thus, it was possible to obtain spliced homogeneous series for the four types of assets for the second half of the twentieth century employing the CNE70 criteria. Capital formation for 1850–1958 distinguishes a larger number of assets (Prados de la Escosura, 2003). Therefore, a test of the robustness of the capital stock estimates to alternative degrees of disaggregation in the underlying investment series for 1850–1958 has been carried out. Although their long-run trends do not differ significantly, the capital stock series constructed from more disaggregated GFCF series displays a persistently higher level as a consequence of the longer lives attributed to transport equipment and nonresidential construction (Fig. 2). Another interesting result is that the average service lives for nonresidential construction and for transport equipment differ between these alternative 10000000

1000000

100000

18 5 18 0 5 18 3 5 18 6 5 18 9 6 18 2 6 18 5 6 18 8 7 18 1 7 18 4 7 18 7 8 18 0 8 18 3 8 18 6 8 18 9 9 18 2 9 18 5 9 19 8 0 19 1 0 19 4 0 19 7 1 19 0 1 19 3 1 19 6 1 19 9 2 19 2 2 19 5 2 19 8 3 19 1 3 19 4 3 19 7 4 19 0 4 19 3 4 19 6 4 19 9 5 19 2 5 19 5 58

10000

de-aggregated

Fig. 2.

aggregated CNE58

aggregated CNE70

Capital Stock Estimates Constructed with Alternative Disaggregation of GFCF Series (1958 Million Pesetas).

151


estimates as a result of asset composition changes. Thus, the average useful lives for transport equipment rises to 36.9 and 27.9 years for 1850–1919 and 1920–1959, respectively, from an average of 20 years assumed by Feinstein (1988).20 Likewise the inclusion of nonresidential buildings (for which, as for dwellings, we assume 70 years of useful life) in ‘‘other structures’’ increases average service life in nonresidential structures from 50 to 55.7 and 54.7 years for 1850–1919 and 1920–1959, respectively. As a consequence, the resulting depreciation rates for these two kinds of assets are altered. In our computations we have used the service lives that result from taking into account assets’ composition changes (Tables 1 and 2). We also compared our ‘‘modified’’ geometric depreciation capital stock series with alternative series computed with arithmetic (X ¼ 1) depreciation rates.21 In Fig. 3 and Table 4, we present series constructed with these two alternative methodologies. It can be appreciated that levels are higher in the straight line depreciation series than in those computed with modified geometric depreciation rates. This is not an unexpected outcome given the fact that geometric-type depreciation results in a rapid decline during the early years of asset lives. Furthermore, the modified geometric depreciation series exhibits more intense growth during the phases of acceleration (the 1920s and the Golden Age, 1950–1974), while the arithmetic depreciation 1000000

100000

10000

1850 1854 1858 1862 1866 1870 1874 1878 1882 1886 1890 1894 1898 1902 1906 1910 1914 1918 1922 1926 1930 1934 1938 1942 1946 1950 1954 1958 1962 1966 1970 1974 1978 1982 1986 1990 1994 1998

1000

Stock (Modified Geometric Depreciation)

Fig. 3.

Gross Stock (Arithmetic Depreciation)

Net Stock (Arithmetic Depreciation)

Capital Stock Estimates with Alternative Depreciation Rates (1995 Million Pesetas).

152


Table 4. Alternative Capital Stock Measures, 1850–2000: Growth Rates (%). Modified Geometric Depreciation

Linear Depreciation (Gross Stock)

Linear Depreciation (Net Stock)

1850–2000

3.5

3.6

3.5

Long periods 1850–1950 1951–1974 1975–2000

2.7 6.0 4.5

2.8 5.4 4.6

2.8 5.4 4.6

Long swings 1850–1883 1884–1920 1921–1929 1930–1952 1953–1958 1959–1974 1975–1986 1987–2000

3.6 2.3 3.5 1.6 4.5 7.0 4.5 4.6

4.3 2.4 2.7 1.3 3.7 6.5 5.0 4.3

4.3 2.4 2.7 1.3 3.7 6.4 4.9 4.3

Cycles 1855–1866 1867–1873 1874–1883 1884–1892 1893–1901 1902–1913 1914–1920 1921–1929 1930–1935 1936–1944 1945–1952 1953–1958 1959–1964 1965–1974 1975–1978 1979–1986 1987–1992 1993–2000

5.4 1.6 3.0 2.2 2.3 2.6 1.7 3.5 2.2 0.1 2.7 4.5 5.0 8.2 6.9 3.3 5.2 4.1

5.9 2.8 3.3 2.7 2.6 2.4 1.9 2.7 1.8 0.3 2.1 3.7 4.8 7.5 6.9 4.0 4.6 4.2

5.9 2.8 3.3 2.7 2.6 2.4 1.8 2.7 1.8 0.3 2.1 3.7 4.8 7.4 6.8 4.0 4.6 4.2

Sources: See text.

estimates grew more intensively in phases of slower growth (with the exception of 1850–1883). In the long run, however, capital stocks derived through straight line and modified geometric depreciation grew at the same pace. This fact renders our estimates robust to alternative depreciation rates.


153

Another issue is how sensitive are the capital stock series to the choice of a given benchmark’s relative prices. Usually capital stock estimates are denominated in the currency of a given year, say, in 1995 US dollars, but is this actually the outcome of computing capital stock with a fixed set of prices from a single benchmark (1995, in our example), or is it just a ‘‘numeraire’’ to express in homogeneous units the real value of a stock derived from spliced capital series constructed at different sets of relative prices for different periods? A single weighted index provides a good measure of real capital stock as long as the relative price structure of capital assets over the time span considered does not differ significantly from the one prevalent in base year. However, because of substantial changes over time in the relative prices of capital goods – largely traceable to rapidly declining prices of machinery and equipment – price weights for, say, 1995, would only be appropriate for a short period around this year. For earlier years, the use of fixed 1995 price weights would understate the growth of capital (the so-called Gerschenkron effect), since the most dynamic capital goods grew faster as a consequence (at least, in part) of the more intense decline in their relative prices. Conversely, the growth of capital would be exaggerated if the prices for an early benchmark year were chosen. Thus, we have employed prices from as many benchmarks as possible (1958, 1965, 1970, 1980, 1985, and 1995) and, for the time span between each pair of adjacent benchmark years (say, 1970–1980), we computed alternative capital stock series at the relative prices of each one (say, both at 1970 and 1980 prices) and, then, spliced the two indices into a single one using a variable-weighted geometric average, in which the weight assigned to each benchmark year’s series increases the closer benchmark t is to each of the years considered. However, as can be seen in Fig. 4, the differences between fixed- and variable-weighted series are minimal, because the underlining real GFCF series had already been obtained through splicing volume indices series computed at the relative prices of different benchmark years (Prados de la Escosura, 2003). The comparison with available estimates, mostly for the last decades, provides a final test for the congruence of our results. Although no official capital statistics exist for Spain, scholars have conducted, with disparate methods, independent investigations of the capital stock, including the construction of long-term series.22 Differences in growth rates among different series are not substantial (Table 5). Alternative historical reconstructions share similar trends with our new series. Differences are noticeable, however, and for the first half of the twentieth century, Cubel Montesinos and Palafox Ga´mir (1997) suggest

154


1000000

100000

10000

1850 1854 1858 1862 1866 1870 1874 1878 1882 1886 1890 1894 1898 1902 1906 1910 1914 1918 1922 1926 1930 1934 1938 1942 1946 1950 1954 1958 1962 1966 1970 1974 1978 1982 1986 1990 1994 1998

1000

Spliced Capital Stock (Modified Geometric Depreciation) Capital Stock in 1995 prices (Modified Geometric Depreciation)

Fig. 4.

Single (At Fixed 1995 Prices) and Variable Weighted (Spliced) Capital Stock (1995 Million Pesetas).

a more intense growth in the 1920s, while, in Hofman’s (1993) estimates, the faster growth during the 1950s is compensated by the slower expansion after 1975. This last remark also applies to the rest of the earlier estimates. Only the rates of growth for ‘‘productive’’ capital (that is, excluding residential structures) are similar to our capital input growth since Spain’s admission into the European Union (1986). We can, then, conclude that trends in capital stock (and input) are quite robust to alternative computation methods and assumptions about depreciation rates and service lives of assets.

TRENDS IN CAPITAL STOCK AND INPUT Trends in capital stock and its components are shown in Fig. 5, while their average rates of growth in each of the significant phases and long swings that can be distinguished in Spain’s economic performance (Prados de la Escosura, 2007) are presented in Table 6. Over the last 150 years, the capital stock grew, on average, at 3.5 percent per year, which implies that capital stock doubled every 20 years. Machinery and transport equipment grew faster than the rest of capital stock components and doubled every

6.0 4.5

3.5 1.6 4.5 7.0 4.5 4.6

2.6 1.7 3.5 2.2 0.1 2.7 4.5 5.0 8.2 6.9 3.3 5.2 4.1

1951–1974 1975–2000

1921–1929 1930–1952 1953–1958 1959–1974 1975–1986 1987–2000

1902–1913 1914–1920 1921–1929 1930–1935 1936–1944 1945–1952 1953–1958 1959–1964 1965–1974 1975–1978 1979–1986 1987–1992 1993–2000

Stock (Geometric depreciation)

2.8 2.0 3.9 2.7 0.4 2.7 4.9 5.4 8.6 7.0 3.2 5.5 4.3

3.9 1.5 4.9 7.4 4.5 4.8

6.4 4.7

Input (Geometric depreciation)

2.4 1.9 2.7 1.8 0.3 2.1 3.7 4.8 7.5 6.9 4.0 4.6 4.2

2.7 1.3 3.7 6.5 5.0 4.3

5.4 4.6

Gross stock (Linear depreciation)

2.4 1.8 2.7 1.8 0.3 2.1 3.7 4.8 7.4 6.8 4.0 4.6 4.2

2.7 1.3 3.7 6.4 4.9 4.3

5.4 4.6

Net stock (Linear depreciation)

5.0 7.3 6.0 3.5 4.1

4.4 6.5 4.3 4.1

5.7

(Gross stock) (Linear depreciation)

6.0 8.2 5.7 2.5 3.9

5.4 7.4 3.6 3.9

6.6

(Net stock) (Linear depreciation)

Hofman

Alternative Capital Estimates: Growth Rates (%).

Prados de la Escosura and Rose´s

Table 5.

3.0 1.3 4.8 2.0 0.3 1.3 5.1

4.8 0.9 5.1

(Net stock) (Linear depreciation)

Cubel/ Palafox

5.7

(Geometric depreciation)

Cebriań

Capital Accumulation in Spain 155

(Continued )

5.5 4.4

6.5 4.3

4.8 4.0 2.0 3.3

7.8 9.1 5.2 2.2 4.2

6.3 5.1 3.1 4.0 3.5

7.1 5.1 3.0 4.1 3.3

3.7 3.6

3.8

8.1 6.0 2.9 5.5 4.5

3.9 4.9

4.4

5.6 4.0

4.7

Sources: Prados de la Escosura and Rose´s (Tables 4 and 7); Hofman (1993), Cubel Montesinos and Palafox Ga´mir (1997), Cebriań (2001), Myro (1983), Baiges, Molinas, and Sebastiań (1987), BDMORES, Dabań Sańchez et al. (1998), MOISSES database kindly provided by Antonio Dı´ az Ballesteros, Mas et al. (2005a), Mas, Pe´rez, and Uriel (2005b), Timmer et al. (2005).

1902–1913 1914–1920 1921–1929 1930–1935 1936–1944 1945–1952 1953–1958 1959–1964 1965–1974 1975–1978 1979–1986 1987–1992 1993–2000

8.6 3.2 3.6 3.8 3.7

2.7

1921–1929 1930–1952 1953–1958 1959–1974 1975–1986 1987–2000

2.4

3.9

1951–1974 1975–2000

Myro Baiges et al. BDMORES MOISSES Mas et al. Mas et al. Mas et al. Timmer et al. (Geometric (Linear (Linear (Linear (Gross stock) (Net stock) (Productive K) (Gross stock) depreciation) depreciation) depreciation) depreciation) (Linear (Linear (Linear (Productive K) depreciation) depreciation) depreciation)

Table 5.

156 LEANDRO PRADOS DE LA ESCOSURA AND JOAN R. ROSE´S

157

Capital Accumulation in Spain 1000000

100000

10000

1000

100

Dwellings

Fig. 5.

Other Constructions

Transportation Equipment

Machinery and Equipment

2000

1995

1990

1985

1980

1975

1970

1965

1960

1955

1950

1945

1940

1935

1930

1925

1920

1915

1910

1905

1900

1895

1890

1885

1880

1875

1870

1865

1860

1855

1850

10

Capital Stock

Trends in Capital Stock and its Components (Semilog Scale) (1995 Million Pesetas).

14.5 years, while dwellings’ expansion was, instead, the slowest, doubling only every 22 years. This implies a deep change in the composition of capital stock over the long run with a steady decline in the weight of residential capital and an increasing contribution of infrastructure and equipment (Fig. 6). The relative size of dwellings shrank from two-thirds to over onethird and, altogether, residential and nonresidential structures went from representing nearly all of the capital stock to four-fifths by the end of the twentieth century, while machinery and transport equipment increased their share by more than sixfold over the same period. Capital stock and input did not follow a steady path as Fig. 6 and Table 7 show, with a more intense expansion during the Golden Age but without returning to the pre-1950 path of growth in the last quarter of the twentieth century. Different phases can be distinguished in the evolution of capital during the first hundred years of modern economic growth in Spain: an intense expansion between the 1850s and the early 1880s, followed by a slowdown until World War I; then, growth resumed briskly during the 1920s, was cut short in the early 1930s and remained sluggish until 1950. During the second half of the twentieth century, capital accumulation grew at a faster and steadier pace, with a big spurt in the years 1959–1974. Thus, phases of acceleration and deceleration in the growth of capital correspond to those in GDP growth, especially in the 1920s and 1953–1974, while in 1850–1883 and 1975–1986 the growth rate of output was well behind that of capital.

158


Table 6. Capital Stock and Its Components, 1850–2000: Growth Rates (%). Dwellings

Other Constructions

Transport Equipment


Capital Stock

1850–2000

3.1

3.7

4.6

4.8

3.5

Long periods 1850–1950 1951–1974 1975–2000

2.5 5.5 3.5

2.8 6.1 5.0

3.4 8.6 5.7

3.8 7.9 5.9

2.7 6.0 4.5

Long swings 1850–1883 1884–1920 1921–1929 1930–1952 1953–1958 1959–1974 1975–1986 1987–2000

3.5 2.1 3.1 1.4 4.4 6.2 3.8 3.2

3.7 2.5 3.6 1.8 3.5 7.5 5.2 4.9

7.7 2.2 6.6 2.1 7.6 9.8 4.7 6.5

3.8 4.1 4.7 3.3 7.7 8.3 4.5 7.2

3.6 2.3 3.5 1.6 4.5 7.0 4.5 4.6

Cycles 1855–1866 1867–1873 1874–1883 1884–1892 1893–1901 1902–1913 1914–1920 1921–1929 1930–1935 1936–1944 1945–1952 1953–1958 1959–1964 1965–1974 1975–1978 1979–1986 1987–1992 1993–2000

4.9 1.7 2.8 1.9 2.3 2.4 1.5 3.1 1.3 0.4 2.7 4.4 5.1 6.9 5.8 2.8 3.3 3.1

5.9 1.5 3.0 3.1 2.2 2.6 1.8 3.6 3.2 0.4 2.4 3.5 4.3 9.4 7.6 4.0 5.9 4.1

14.6 1.1 5.2 0.8 2.6 2.6 4.9 6.6 0.7 5.9 0.1 7.6 8.6 10.5 10.6 1.8 5.4 7.4

3.9 1.9 5.7 4.7 4.1 4.9 1.8 4.7 7.1 0.2 4.4 7.7 6.1 9.6 7.4 3.0 9.5 5.5

5.4 1.6 3.0 2.2 2.3 2.6 1.7 3.5 2.2 0.1 2.7 4.5 5.0 8.2 6.9 3.3 5.2 4.1

Source: See text.

Changes in the composition of capital away from residential structures increased the service flow provided by the capital stock to production. This results in a growing gap between growth rates of capital input and stock (Table 7). The relative changes in the capital stock implied modifications in

159


0.60

0.50

0.40

0.30

0.20

0.10

Dwellings

Fig. 6.

Other Constructions


2000

1995

1990

1985

1980

1975

1970

1965

1960

1955

1950

1945

1940

1935

1930

1925

1920

1915

1910

1905

1900

1895

1890

1885

1880

1875

1870

1865

1860

1855

1850

0.00


Composition of Capital Stock (%) (1995 Prices).

the structure of capital compensation (Fig. 7). By 1850, capital compensation (the sum of all capital rents) accrued overwhelmingly to residential and nonresidential structures (87 percent), while capital in transport equipment and machinery received only 13 percent of it. One and a half centuries later, productive capital had increased its share to nearly 30 percent. In particular, the share of machinery and equipment in capital rents trebled. The difference between capital input and stock, that captures composition changes, is sometimes identified with improvements in the ‘‘quality’’ of capital stock.23 Interestingly, the ‘‘quality’’ of capital rose in periods of faster capital growth (Table 7 and Fig. 8). Specifically, three periods in which capital ‘‘quality’’ grew above its long-run trend stand out: from the mid-1850s to the early 1880s, a period of opening up in which foreign capital was invested in railways construction and in mining; the 1920s and early 1930s, that witnessed another episode of capital inflow from abroad and the first phase of Spanish electrification; and the ‘‘Golden Age’’ (1953–1974), in which Spain completed the process of electrification and replaced old vintages capital after two decades of international isolation due to the Great Depression, the Civil War (1936–1939) and the inward looking policies of Franco’s regime. It is worth nothing that in spite of the large influx of European funds since Spanish accession to the European Union (1986), the ‘‘quality’’ of capital did not rise well above the historical trend rate over


160

Table 7. Growth Rates of Capital Stock, Quality, and Input, 1850–2000 (%). Capital Stock

Capital Quality

Capital Input

1850–2000

3.5

0.2

3.7

Long periods 1850–1950 1951–1974 1975–2000

2.7 6.0 4.5

0.1 0.4 0.2

2.8 6.4 4.7

Long swings 1850–1883 1884–1920 1921–1929 1930–1952 1953–1958 1959–1974 1975–1986 1987–2000

3.6 2.3 3.5 1.6 4.5 7.0 4.5 4.6

0.3 0.1 0.4 0.1 0.5 0.4 0.0 0.2

4.0 2.4 3.9 1.5 4.9 7.4 4.5 4.8

Cycles 1855–1866 1867–1873 1874–1883 1884–1892 1893–1901 1902–1913 1914–1920 1921–1929 1930–1935 1936–1944 1945–1952 1953–1958 1959–1964 1965–1974 1975–1978 1979–1986 1987–1992 1993–2000

5.4 1.6 3.0 2.2 2.3 2.6 1.7 3.5 2.2 0.1 2.7 4.5 5.0 8.2 6.9 3.3 5.2 4.1

0.8 0.3 0.4 0.1 0.1 0.2 0.2 0.4 0.5 0.5 0.0 0.5 0.4 0.4 0.2 0.0 0.3 0.2

6.3 1.2 3.4 2.2 2.4 2.8 2.0 3.9 2.7 0.4 2.7 4.9 5.4 8.6 7.0 3.2 5.5 4.3

Source: See text.

1986–2000, which could suggest a delayed impact of ICT technologies on Spain (Timmer, Ypma, & van Ark, 2005). A glance at aggregate capital and its components over different long swings and cycles shows a large variance in their rates of growth (Tables 6 and 7).

161

Capital Accumulation in Spain 0.50 0.45 0.40 0.35 0.30 0.25 0.20 0.15 0.10 0.05

1995

1990

1985

1980

1975

1970

1965


Fig. 7.

Capital Input. Shares of Rental Value.

1894


1960

1955

1950

1945

1940

1935

1930

1925

1920

1915

1910

1905

1900

1895

1890

1885

Other Constructions

1878

Dwellings

1880

1875

1870

1865

1860

1855

1850

0.00

1.4 1.35 1.3 1.25 1.2 1.15 1.1 1.05 1 0.95

Fig. 8.

Index of ‘‘Quality’’ of Capital (1850 ¼ 100).

1998

1994

1990

1986

1982

1978

1974

1970

1962 1966

1958

1954

1950

1946

1942

1938

1934

1930

1926

1922

1918

1914

1910

1906

1902

1898

1886 1890

1882

1874

1870

1866

1862

1858

1854

1850

0.9

162


Institutional reforms and opening up to foreign capital and international trade favored an expansion during the first long swing, 1850–1883. Inflows of foreign capital made it possible to break the close connection between investment and savings and contributed to the economic growth (Prados de la Escosura, 2009). The capital stock grew at an average of 3.6 per annum, well above the nineteenth century’s average but with irregular and pronounced cycles (Table 6, Panel C). The strong start (1855–1866) was led by the railway construction, as evidenced by the dramatic increase in transport equipment, although the expansion affected all types of capital goods. The international crisis of 1866 reduced dramatically the influx of foreign capital that fuelled railways expansion, while political turmoil (with two changes of political regime, social turmoil, coups d’e´tat, and a civil conflict: the Carlist War) made the Spanish economy less attractive for local and foreign investment during the years 1866–1873. The subsequent political stabilization that followed the restoration of the Bourbon dynasty led to a recovery of capital stock growth rates, particularly in transport equipment and machinery, which grew at respectable rates, close to 6 percent per annum. The second long swing, 1883–1920, covers most of the so-called Restauracioń (1875–1923), an era of institutional stability that presumably provided a favorable environment for investment and growth. And yet both permanent and temporary factors worked against it. The growth of capital stock slowed down to 2.3 percent per year with regular and mild cycles. Weak urbanization, resulting from sluggish industrial growth and a delayed demographic transition, slowed down the expansion of the stock of dwellings.24 The closing of a large section of the railway network together with the disappointing financial results for railways companies hindered a further expansion of nonresidential structures (Herranz-Loncań, 2007). Nonetheless, a significant expansion occurred in machinery and equipment as industrialization proceeded steadily during this period. Wars did not have an expansionary effect on capital accumulation. The War of Cuba’s Independence, despite the weak economic flows between the metropolis and the colony, introduced macroeconomic instability that led to a contraction in foreign investment and the depreciation of the peseta after 1891 that, interestingly, had been unaffected by the abandonment of the convertibility of Spain’s peseta into gold in 1883.25 World War I, in spite of Spanish neutrality, witnessed a significant decline in the capital stock growth (by about one-third) and, more prominently, the growth of machinery stock more than halved.26 As Sudria` (1990) emphasized, the slowdown in the replacement of older vintages, associated with an increasing utilization of the installed capacity, led to a rapid obsolescence of the machinery stock in use.


163

The most intense growth of the period 1850–1950 was achieved in the 1920s. In this decade, the growth rates of capital stock were the highest since the 1860s. As such an intense growth took place under the dictatorship of Primo de Rivera (1923–1929), inevitably, economists and historians have tended to assume that state intervention through external protection and regulation, on the one hand, and investments in public infrastructure, on the other hand, made a decisive contribution to capital accumulation, and, subsequently, to growth (Velarde Fuertes, 1968). Against this view it has been argued that (a) government intervention led to resource misallocation because it did not take into account its opportunity cost (Comı´ n, 1987); (b) the increasing power of oligopolies reduced incentives for technological change (Fraile Balbı´ n, 1991); and (c) the expansion of public spending (through the increase in money supply and government debt) fuelled inflation and increased currency volatility (Comı´ n & Martı´ n Acenã, 1984; Palafox, 1991). The emphasis on tariff protectionism has tended to neglect, however, that a significant inflow of foreign capital allowed the purchase of capital goods (Tena Junguito, 1999; Prados de la Escosura, 2009).27 A closer look at the evidence shows that growth rates in transport equipment, stimulated by the Dictatorship’s infrastructure construction policy, exceeded largely those of the rest of capital components. The period 1929–1952 is the fourth, and last, long swing of the 1850–1950 era. The capital stock growth rate fell to 1.6 percent per year with irregular and severe cycles. A deceleration in the capital stock growth between 1929 and 1935 was followed by stagnation during the Civil and World Wars (1935–1944) and, then, a mild recovery up to 1952. The first half of the 1930s represents a fracture in the intense capital stock expansion of the previous decade. Our results suggest a moderate impact of the Depression in capital stock growth. This result is not surprising and is in line with previous research (Comı´ n, 1987; Prados de la Escosura, 2003). However, the effects of the crisis in Spain, although less intense than in European countries until 1932, were most persistent, at least in comparison to those nations which managed to leave the gold standard soon (Eichengreen, 1992). This broad picture is complicated by the disparate evolution of capital stock components. Uncertainty about a new political system, the Second Republic (1931–1936), seems to have been a major cause for the decline in the growth of residential structures. As a consequence of the restrictive budgetary policy and the interruption of public works (Palafox, 1991), transport equipment growth declined to very low levels although, paradoxically, this was not the case of nonresidential structures which remained at levels similar to those observed in the 1920s.28 In sharp

164


contrast, the increase in the stock of machinery and equipment exceeded that of the 1920s, suggesting that the social unrest and political turmoil did not slow down the renewal of industrial machinery, a likely outcome of the electrification process that was taking place since the early 1920s. Hardly any growth of the aggregate capital stock took place during the Civil War and the subsequent post-war years (1935–1944), a period in which the destruction of transport equipment stands out. The impact of war destruction was very uneven, as was the immediate post-war reconstruction. As observed above, while the stock of houses and structures was hardly affected by the Civil War, the stock of machinery and, especially, of transport equipment fell significantly (about one-fourth altogether). In the immediate post-war, a vigorous rise in machinery and equipment contrasted with none in transport equipment. In comparative perspective, western European economies recovered faster from capital destruction during World War II than Spain did from the Civil War (Maddison, 1991). The change in trend which began in the early 1950s ushered in an exceptional phase of rapid growth which lasted until 1974. Despite the fact that the volatility of import capacity rendered investment risky and tended to penalize capital accumulation, while inflows of foreign capital and new technology were restricted (Prados de la Escosura & Sanz, 1996), a dramatic change in trend occurred in the 1950s. Machinery and transport equipment grew at rates above 7 percent per year, while structures did it around 4 percent. It can be hypothesized that the US–Spanish cooperation agreements of 1953 triggered economic agents’ confidence in the viability of Franco’s dictatorship leading to an increase in capital accumulation and to imports of new vintage machinery and equipment (Calvo-Gonza´lez, 2007). The cautious move toward deregulation and opening up initiated in the mid-1950s intensified after the 1959 stabilization and liberalization plan resulting in accelerated capital accumulation during Spain’s Golden Age (1959–1974). Capital stock growth reached peak rates (7 percent on average) and was particularly intense in the case of ‘‘productive’’ capital (that is, excluding residential structures). The adoption of mass production techniques from abroad and the diffusion of road transport appear crucial for this accelerated capital accumulation. A change in trajectory began in the late 1970s, continuing to the end of the twentieth century, in which capital growth rates returned to those of the 1950s and early 1960s. It is worth noting, nonetheless, that capital expansion maintained its Golden Age pace until 1978. This was possibly due to the fact that relative prices did not adjusted immediately to the oil shocks as the government implemented a policy of subsidies to soften the political


165

transition from Franco’s dictatorship to democracy. The severe economic adjustment introduced by Moncloa agreements (1978) led to a deceleration in capital accumulation (Prados de la Escosura & Sanz, 1996). The last quarter of the twentieth century can be split into two periods with Spain’s accession to the European Union as a turning point. The first one (1975–1986) was marked by the transition to democracy and the reorganization of the Spanish economy as capital equipment, largely obsolete and energy intensive, needed to be replaced. Since 1986, European funds largely contributed to the construction of new infrastructures and the renewal of public transport equipment.

CAPITAL DEEPENING AND PRODUCTIVITY How does the long-term rise in the stock and input of capital fit into the wider context of Spain’s economic performance? A possible way to do it is by looking at capital intensity, which relates the amount of capital to other factors of production, especially labor. Since the use of capital makes labor more effective, rising capital intensity (or ‘‘capital deepening’’) pushes up the productivity of labor. In Spain, a process of capital deepening took place over the period 1850–2000 (Table 8): the endowment of capital (stock and input) per hour worked, multiplied by 102 and 140, respectively, while the rates of the second half of the twentieth century practically doubled those prevailing in the previous hundred years. Also noticeable are the significant differences appeared between these two measures of capital intensity. Another measure of capital intensity: the capital to output ratio, multiplied by 4.5 and 6.2 for the stock and input, respectively. Interestingly, and contrary to Kaldor’s (1961, p. 178) stylized fact, the capital–output ratio did not remain stable over the long run in Spain.29 Different phases can be described in its evolution (Fig. 9). The capital–output ratio grew significantly over the first hundred years considered (1850–1950) – with the exception of the 1920s – and, again, during the last quarter of the twentieth century, but decreased during the ‘‘Golden Age’’ (1950–1974), just at the time the growth of GDP was fastest. This exceptional situation in which the productivity of capital (that is, the inverse of the capital–output ratio) increased suggests a significant contribution of total factor productivity to Spanish economic growth over these years. Spanish experience can be better assessed in international perspective. For different world regions in the late twentieth century, Table 9 compares


166

Table 8.

Capital Intensity and Productivity, 1850–2000: Growth Rates (%). Capital Stock/ Hour Worked

Capital Input/ Hour Worked

GDP/Capital Stock

GDP/Capital Input

1850–2000

3.1

3.3

1.0

1.2

Long periods 1850–1950 1951–1974 1975–2000

2.1 5.0 4.9

2.3 5.4 5.1

1.2 0.5 1.5

1.4 0.1 1.7

Long swings 1850–1883 1884–1920 1921–1929 1930–1952 1953–1958 1959–1974 1975–1986 1987–2000

3.0 2.1 1.7 0.7 4.1 6.4 8.1 2.2

3.3 2.2 2.1 0.7 4.6 6.8 8.1 2.5

1.8 1.0 0.3 0.7 0.2 0.1 2.0 1.1

2.2 1.1 0.1 0.7 0.2 0.5 2.0 1.4

Cycles 1855–1866 1867–1873 1874–1883 1884–1892 1893–1901 1902–1913 1914–1920 1921–1929 1930–1935 1936–1944 1945–1952 1953–1958 1959–1964 1965–1974 1975–1978 1979–1986 1987–1992 1993–2000

4.5 0.0 3.6 1.8 1.8 2.4 2.2 1.7 0.5 0.0 1.7 4.1 5.5 7.0 9.9 7.2 2.0 2.4

5.4 0.3 3.9 1.7 1.9 2.6 2.4 2.1 1.0 0.5 1.7 4.6 5.9 7.4 10.1 7.1 2.2 2.7

4.1 1.8 1.9 1.5 1.1 1.3 0.0 0.3 2.0 0.4 0.1 0.2 1.4 1.1 3.1 1.4 1.1 1.2

4.9 2.1 2.3 1.4 1.2 1.5 0.2 0.1 2.5 0.1 0.1 0.2 1.0 1.5 3.3 1.4 1.3 1.4

Sources: Capital stock and input, Table 7; hours worked, Prados de la Escosura and Rose´s (2008); GDP (Divisia), computed from Prados de la Escosura (2003).

167


3.5

3.0

2.5

2.0

1.5

1.0

0.5

18 5 18 0 5 18 4 5 18 8 6 18 2 6 18 6 7 18 0 7 18 4 7 18 8 8 18 2 8 18 6 9 18 0 9 18 4 9 19 8 0 19 2 0 19 6 1 19 0 1 19 4 1 19 8 2 19 2 2 19 6 3 19 0 3 19 4 3 19 8 4 19 2 4 19 6 5 19 0 5 19 4 5 19 8 6 19 2 6 19 6 7 19 0 7 19 4 7 19 8 8 19 2 8 19 6 9 19 0 9 19 4 98

0.0

Fig. 9.

Capital Stock–Output Ratio.

growth rates for capital deepening and capital productivity. In Panels A and B, our measures are constructed with inputs of capital and labor, that is, the service provided to production by these two factors, while in Panel C stocks are employed. The time spans considered are determined by the availability of international evidence, not matching, thus, our favored periodization. Some results emerge from the comparison. In the context of OECD countries during the Golden Age (1950–1973), capital deepening does not appear to have increased particularly fast in Spain: although she was in its upper growth segment during the 1950s, fell behind in the 1960s and early 1970s, only remaining above North America. The productivity of capital increased mildly in Spain, at a much slower pace in the 1950s than in countries that experienced more devastation in World War II. This raises the issue of Spain’s sluggish recovery after the Civil War (1936–1939). For example, why, starting from a lower level of capital, did its productivity grow so slowly? Did it result from a low human capital endowment or from resource misallocation in an overregulated autarchic economy? Conversely, during the years 1960–1973, and after a cautious liberalization and opening up, capital productivity grew in Spain while declining in Western Europe and Japan. Panel B shows some interesting similarities between the East Asian ‘‘tigers’’ and Spain in the late twentieth century. In all of them, intense capital deepening went along with a significant decline in the productivity of

168


Table 9. An International Comparison of Capital Deepening and Productivity: Growth Rates (%). 1950–1960

1960–1973

Capital–labor

GDP–capital

Capital–labor

GDP–capital

5.7 4.4 5.3 1.7 0.3 2.6 4.3 3.5 4.1

1.6 0.2 1.3 2.7 3.6 1.0 1.2 0.8 0.3

2.9 5.9 7.7 6.1 8.8 6.3 4.6 1.8 4.5

0.2 0.4 1.6 0.6 0.6 1.0 0.8 0.3 0.2

Panel Aa Canada France Germany Italy Japan Netherlands United Kingdom United States Spain

1966–1990

Panel Ba Hong Kong Singapore South Korea Taiwan Spain

Capital–Labor

GDP–Capital

5.1 6.3 7.5 7.2 5.6

0.4 2.1 2.6 2.4 1.6 1960–2000

Panel Cb World Industrial countries China East Asia (except China) South Asia Latin America Africa Middle East Spain

Capital–Labor

GDP–Capital

2.9 2.6 4.9 6.6 2.9 1.7 1.4 3.1 5.3

0.6 0.4 0.1 2.7 0.6 0.6 0.8 1.0 0.9

Sources: All countries but Spain; Panel A, Christensen, Cummings, and Jorgenson (1980); Panel B, Young (1995); Panel C, Bosworth and Collins (2003). Spain, as in Table 8. a Capital input–labor input and GDP–capital input ratios. b Capital stock–labor quantity and GDP–capital stock ratios.

169


capital. Finally, since 1960 Spain seems to be close to the top world regions in capital intensity and also in terms of capital productivity decline. Finally, a long-run perspective is provided in Table 10 in which trends in capital deepening and productivity for Spain are compared to those in the Table 10. Long-Run Capital Deepening and Productivity in the United Kingdom, the United States, and Spain: Growth Rates (%). United Kingdom Capital–Labor

Spain

GDP–Capital

Capital–Labor

GDP–Capital

0.3 0.1 1.0 0.4 0.7 0.4

3.1 2.4 2.2 2.9 0.9 5.4

2.1 1.4 0.2 3.1 1.4 0.3

a

Panel A 1856–1873 1873–1913 1913–1924 1924–1937 1937–1951 1951–1973

1.9 1.0 3.2 0.3 1.0 3.7 United States

Spain

a

Panel B 1889–1901 1901–1919 1919–1929 1929–1941 1941–1948 1948–1973 1973–1989 1989–2000

1.7 1.7 1.2 0.2 0.4 2.7 2.6 2.5

0.5 0.0 1.1 2.5 1.3 0.2 1.2 0.6

1.3 2.6 1.2 0.8 1.0 4.8 6.4 3.0

0.7 1.1 0.9 1.9 0.4 0.4 1.5 1.5

Panel Cb 1871–1891 1891–1913 1913–1928 1928–1950 1950–1964 1964–1972 1972–1979 1979–1988 1988–1996

0.4 0.5 0.2 0.2 1.9 3.5 2.3 2.4 0.1

0.1 0.6 0.9 1.6 0.4 0.9 0.9 1.6 0.4

2.1 1.7 1.7 1.1 3.2 6.4 8.1 3.7 3.9

0.8 1.0 0.2 1.2 1.3 1.4 2.9 0.7 2.0

Sources: All countries but Spain, Panel A, Matthews, Feinstein, and Odling-Smee (1982); Panel B, Field (2006); Panel C, Gordon (1999). Spain, for capital, see the text and for labor, Prados de la Escosura and Rose´s (2007). a Capital stock–labor quantity and GDP–capital stock ratios. b Capital input–labor input and GDP–capital input ratios.

170


United Kingdom and the United States. At first glance it seems that, in the three countries, a growth rate of the capital–labor ratio above 2 percent per year brings with it a decline in the productivity of capital. A look at different phases or long swings suggests that capital intensity increased faster in the late nineteenth and early twentieth century in Spain than in both the United Kingdom and the United States and, correspondingly, the efficiency in the use of capital fell more acutely. In the 1920s, Spain performed similarly to the United States (Panel B), with significant capital productivity gains while its intensification kept growing. Electrification has been suggested as a major element underlying capital productivity growth in the United States (David & Wright, 1999) that, according to Field (2006), was concentrated in manufacturing. A similar hypothesis can be entertained for the case of Spain, where the process of electrification, interrupted during the Civil War and its autarchic aftermath but completed in the 1950s, also underlies capital productivity growth (Betrań Pe´rez, 2000; Sanchis, 2006). Moreover, during the 1950s the introduction of new capital vintages under the umbrella of the US–Spanish cooperation agreements, that stimulated investment and the acquisition of foreign technology (Calvo-Gonza´lez, 2007), also contributed to capital efficiency. New capital and organizational improvements, together with increases in utilization rates, provided capital efficiency gains during the 1960s and early 1970s. Nonetheless, once the Golden Age was over, accelerated capital deepening was met again by declining capital efficiency in Spain.

CONCLUDING REMARKS Our main results can be summarized as follows. First, our measurement of capital stock growth over the long run yields only a range of best guess estimates. However, our sensitivity tests indicate that differences in growth rates assuming alternative initial capital stock estimates are fairly small and do not change the overall picture. Capital stock estimation appears, then, much less problematic, and less sensitive to underlining assumptions, than it is commonly believed. Second, we point out that capital input adjustments generate a slightly faster growth rates but do not change significantly the long-run performance of capital. Third, Spanish capital stock grew over the entire period (1850–2000) but not at steady rates. Finally, Spain experienced a process of capital deepening and rising capital–output ratios, although in phases of acceleration (the 1920s and the Golden Age) efficiency gains in capital are found.

171


NOTES 1. OECD (1993) but see OECD (2001a, 2001b). For applications of the Jorgenson/Hulten approach, see Christensen et al. (1980), Jorgenson, Gollop, and Fraumeni (1987), Elı´ as (1990), Young (1995). For the case of Spain, see Myro (1983) and Cebriań (2001). An application of OECD’s (2001a) new methodology to the Spanish case can be found in Mas et al. (2005b). 2. Consequently, intangible goods (like licenses, patents, and property rights), nonreproducible goods (like monuments, pieces of art, and natural resources), consumer durables, military goods, inventories, and intermediate products are not part of our capital stock measure. 3. More specifically, the PIM approach produces an estimate of the stock of fixed assets in existence by estimating how many fixed assets installed have survived to the current period. Cf. Diewert (1980), Jorgenson (1973, 1980), and Hulten (1990). 4. Alternative procedures to derive the initial level of capital stock include the direct computation from the cumulative investment during the past years, surveys or censuses of the capital stock for a given year, and retrospective calculations such as the one proposed by Feinstein (1972, pp. 196–198). 5. Following Young (1995, pp. 651–652), we assumed that, for each type of asset, investment growth in the earlier years for which information is available (that is, the 1850s) are representative of investment growth rates in the pre-1850 period. Thus, we used the investment growth rates over 1850/54–1855/59 in our calculations. Baiges et al. (1987) employed a similar approach. 6. A good example of how much capital accumulation intensified during the early stages of the construction of railways is provided by the case of Britain, where without the railways gross fixed capital formation would have expanded at half its pace. Thus, between 1821–1830 and 1841–1850, GDCF grew in the United Kingdom at an annual average rate of 3.1 percent. If railways investment is excluded, its growth rate falls to 1.5 percent (Feinstein, 1988, p. 444). 7. We tried to follow as closely as possible assessments of war destruction by Ros Hombravella, Clavera, Esteban, Mone´s, and Montserrat (1973), Barciela Lo´pez (1986) and, especially, Catalan (1995), together with specific estimates from a wide variety of sector studies. Thus, Lo´pez Carrillo (1998), Appendix 3, provides estimates of the reduction in motor vehicles between 1935 and 1940. Railways’ rolling stock destroyed can be deduced from Go´mez Mendoza (1989) figures weighted by the prices of each type of asset. Munõz Rubio (1995) presents estimates of structures and rolling stock of the two main railways companies, Norte and MZA. Nelson A´lvarez kindly provided us with figures that allowed us to estimate the destruction of telephone networks (equipment and structures). Moreover, available data on installed electric power (Carreras, 1989) and the number of urban dwellings (Tafunell, 1989) allowed us to complete our crude assessment of capital destruction during the Civil War. 8. An earlier attempt to estimate the destruction of capital stock during the Civil War was carried out by Cubel Montesinos and Palafox Ga´mir (1997) who distributed linearly the destroyed assets during the war years to keep consistency with the arithmetic rate of depreciation they used to compute the capital stock.

172


9. If depreciation is assumed to be arithmetic, X takes the value of 1; if, alternatively, depreciation is assumed to be geometric, X equals 2. Following Jorgenson (1990), we adopted the ‘‘modified’’ geometric depreciation pattern, in between the arithmetic and geometric patterns. 10. See a comparison between alternative capital stock estimates derived by using arithmetic (X ¼ 1) and modified geometric depreciation in the section on sensitivity tests below. 11. The values of the parameter were derived from a careful econometric exercise in which a large database was used. Accepting the X parameter’s values from Hulten and Wykoff (1981) for historical purposes is, nonetheless, arbitrary. It is worth noting that these parameters have been widely employed in empirical studies as they correspond to the technological frontier to which countries tend to converge. In his pioneer contribution, Myro (1983) employed also a modified geometric depreciation rate but assumed X ¼ 1.5. 12. Cf. Feinstein (1988), Blades (1993), and O’Mahony (1996, p. 173). Only in the case of buildings we have assumed a fixed useful life over the entire period considered. We accepted different service lives within two asset types: other constructions (in which we assume a service life of 70 years for nonresidential buildings and an average of 50 years for the rest) and transport equipment (railway (40 years up to 1919 and, then, 30 up to 1959), shipping (30 years), and road (10 years)). 13. It is usually assumed that capital input (K) in year t is proportional to the stock of capital C at the end of the period t1. Thus, Kt ¼ lCt1, where the constant (l) transforms the capital stock into its services, and where the capital stock Ct moves according to the new investments, at constant prices, during the year, and to the depreciation and replacement rates. Cf. Jorgenson (1990). 14. Jorgenson (1990), Hall and Jorgenson (1967). 15. We have derived the aggregate depreciation rate by weighting the specific rates of depreciation for each asset type by the corresponding amount of capital. The GFCF deflators for each category of capital goods (Prados de la Escosura, 2003), smoothed with the Hodrick–Prescott filter, provide the remaining information. 16. Under the assumption that rates of return are identical across all types of capital investment, r could be directly computed employing the same Eq. (4) given that the aggregate rental of capital goods is equivalent to property compensation; that is, the remuneration of capital in aggregate value added. Observable differences between the actual nominal rate of return, computed with property compensation, and competitive nominal rates of return, measured with long-run interest rates provide a direct measure of monopolistic rents gained by proprietors. However, deviations from the constant return to scale assumption and unobserved returns to production factors would also account for the discrepancies. 17. Young (1995, p. 649) argues that as each type of input i is assumed to be identical over time, any increases in the efficiency of such input will appear in the ‘‘residual,’’ as is true in our case. Abramovitz and David (2001, p. 23) point out that by ‘‘Capital Quality, we do not refer to the important changes in the characteristics of capital goods which raise their productivity but are the result of technological progress [and are] y embedded in the TFP residual.’’ For a less skeptical view, see Hulten (1990, p. 134; 1992). Our historical estimates fit in the case exposed by Young (1995).


173

18. The UCD capital estimate for 1965 has provided the initial benchmark for capital stock computations such as those by Myro (1983), Go´mez Villegas (1988), Fundacioń BBV (1995), Dabań Sańchez et al. (1998), and Mas, Pe´rez, and Uriel (2000, 2006), Mas et al. (2005a). 19. See the discussion in Young (1995, pp. 650–651) for similar results in the cases of South Korea and Taiwan. As this author stresses, ‘‘because of scrapping, the wealth survey gross assets should actually be less than the cumulative national accounts investment’’, not the opposite. 20. For transport equipment we accepted Feinstein’s (1988) service lives with slight modifications. Thus, Feinstein (1988) assumes 30 years for railway stock, 25 years for ships, and 10 for cars and trucks, with an average useful life of 10–20 years for the whole sector. In our case, we assume 40 years for railway stock up to 1919, and 30 for 1920–1959, while we chose 30 and 10 for ships, and cars and trucks, respectively, over the entire span 1850–1959. From 1960 onwards, our choice of service lives matches the conventional estimates (see Table 1). 21. In the case of the arithmetic depreciation rate, gross capital stock for the period t is computed as: GKt ¼ GKt1þItRt, where It is real gross domestic capital formation, and Rt, retirements. Net capital stock, with straight line depreciation, can be estimated with Eq. (1), but the depreciation rate will be d ¼ 1/T. 22. It is worth highlighting, in addition to the Universidad Comercial de Deusto (1968–1972) study on the national wealth for 1965, Myro’s (1983) pioneering work – in which the Jorgenson approach was applied to Spain for the first time – and the research conducted by Mas et al. (2005a, 2005b) at the Instituto Valenciano de Investigaciones Econo´micas (Ivie) under the sponsorship of the Fundacioń BBVA, during the last two decades. Historical series produced by Hofman (1993), Cebriań (2001) and, especially, Cubel Montesinos and Palafox Ga´mir (1997) are worth mentioning. 23. Alas, our historical exercise fails to capture all the composition – or ‘‘quality’’ – changes, as we cannot carry out a deeper disaggregation by type of asset, but hints into the direction of composition changes. 24. Restrictions on both internal and external competition, according to Fraile Balbı´ n (1998) counterweighted political and social stability. Cf. Tena Junguito (1999), Palafox (1999), and Pardos (1998) on tariff protection and its effects. On the pace of industrialization and the demographic transition, see Prados de la Escosura (2003) and Pe´rez Moreda (1999), respectively. 25. Cf. Prados de la Escosura (2009) on the balance of payments; and Martı´ n Acenã (1994) and Bordo and Rockoff (1996) on the gold standard. On the consequences of Cuba’s war of independence, see the discussion in Fraile Balbı´ n and Escribano (1998) and Maluquer de Motes Bernet (1999). 26. This result is in stark contradiction with the conventional view that stresses World War I stimulating aggregate effects. Cf. Roldań and Garcı´ a Delgado (1973) for the conventional view on the positive impact of the Great War on Spain. 27. It has also been noted that the positive situation of the current account balance of payments during the World War I contributed to the boom of the 1920s (Sudria`, 1990). 28. An alternative view sustaining that expansionary monetary and anti-cyclical fiscal policies were tried to compensate for the fall in private investment and exports

174


(Comı´ n & Martı´ n Acenã, 1984; Garcı´ a Santos & Martı´ n Acenã, 1990) could help explaining this apparent paradox. 29. Maddison (1995) already observed the variance of the ratio of capital stock to GDP.

ACKNOWLEDGMENTS Financial support from the Spanish Ministry of Education and Sciences (Research Projects SEC2002-01596, and ‘‘Consolidating Economics’’), Comunidad de Madrid (CCG06-UC3M/HUM-0872 and CCG07-UC3M/ HUM-3288), and Fundacioń ICO is gratefully acknowledged. An earlier version was presented at the Universidad Carlos III economic history seminar. We acknowledge Antonio Dı´ az Ballesteros and Knick Harley for their comments and advice and Xavier Tafunell for sharing his unpublished data.

REFERENCES Abramovitz, M., & David, P. A. (2001). Two centuries of American macroeconomic growth from exploitation of resource abundance to knowledge-driven development. SIEPR Discussion paper no. 01-05. Baiges, J., Molinas, C., & Sebastiań, M. (1987). La economıá espanõla 1964–1985. Datos, fuentes y ana´lisis. Madrid: Instituto de Estudios Fiscales. Barciela Lo´pez, C. (1986). Introduccioń. In: R. Garrabou, C. Barciela & J. I. Jimeńez Blanco (Eds), Historia agraria de la Espanã contemporańea. 3. El fin de la agricultura tradicional (1900–1960) (pp. 383–454). Barcelona: Crı´ tica. Betrań Pe´rez, C. (2000). Natural resources, electrification and economic growth from the end of the XIXth century until World War II. Universidad de Valencia Working Papers no. DT 00-03. Blades, D. W. (1993). Comparing capital stocks. In: A. Szirmai, B. van Ark & D. Pilat (Eds), Explaining economic growth. Essays in honour of Angus Maddison (pp. 399–409). Amsterdam: North Holland. Bordo, M., & Rockoff, H. (1996). The gold standard as a ‘good housekeeping seal of approval’. Journal of Economic History, 56, 389–428. Bosworth, B. P., & Collins, S. M. (2003). The empirics of growth: An update. Brookings Papers in Economic Activity, 2, 113–179. Calvo-Gonza´lez, O. (2007). American military interests and economic confidence in Spain under Franco dictatorship. Journal of Economic History, 67, 740–767. Carreras, A. (1989). Industria. In: A. Carreras (Ed.), Estadı´sticas Histo´ricas de Espanã. Siglos XIX–XX (pp. 169–247). Madrid: Fundacioń Banco Exterior. Catalan, J. (1995). La economıá espanõla y la Segunda Guerra Mundial. Barcelona: Ariel.


175

Cebriań, M. (2001). Las fuentes del crecimiento econo´mico espanõl, 1964–1973. In: F. Comı´ n, & B. Sańchez-Alonso (Eds), Los novı´simos en la historia econo´mica de Espanã, Revista de Historia Econo´mica XIX (special issue), pp. 277–299. Christensen, L. R., Cummings, D., & Jorgenson, D. W. (1980). Economic growth, 1947–1973: An international comparison. In: J. W. Kendrick & B. N. Vaccara (Eds), New developments in productivity measurement and analysis (pp. 595–698). Chicago, IL: University of Chicago Press/NBER. Comı´ n, F. (1987). La economı´ a espanõla en el perı´ odo de entreguerras (1919–1935). In: J. Nadal, A. Carreras & C. Sudria` (Eds), La economıá espanõla en el siglo XX. Una perspectiva histo´rica (pp. 105–149). Barcelona: Ariel. Comı´ n, F., & Martı´ n Acenã, P. (1984). La polı´ tica monetaria y fiscal durante la Dictadura y la Segunda Repu´blica. Papeles de Economıá Espanõla, 20, 236–261. Cubel Montesinos, A., & Palafox Ga´mir, J. (1997). El stock de capital de la economı´ a espanõla. 1900–1958. Revista de Historia Industrial, 12, 113–146. Dabań Sańchez, T., Dı´ az Ballesteros, A., Escriba´ Pe´rez, F. J., & Murgui Garcı´ a, M. J. (1998). La base de datos BDMORES. M1 de Economı´ a y Hacienda. Direccioń General de Ana´lisis y Programacioń Presupuestaria. Documento de Trabajo D-98001. David, P. A., & Wright, G. (1999). General purpose technologies and surges in productivity: Historical reflections on the future of the ICT revolution. University of Oxford Discussion papers in Economic and Social History no. 31 (September). De Long, J. B., & Summers, L. H. (1991). Equipment investment and economic growth. Quarterly Journal of Economics, 102(2), 445–502. Diewert, W. E. (1980). Capital and the theory of productivity measurement. American Economic Review, 70(2), 260–267. Eichengreen, B. (1992). Golden fetters: The gold standard and the great depression, 1919–1939. New York: Oxford University Press. Elı´ as, V. (1990). Sources of growth: A study of seven Latin American Economies. San Francisco, CA: Institute for Contemporary Studies Press. Feinstein, C. H. (1972). National income, expenditure, and output of the United Kingdom, 1855–1965. Cambridge: Cambridge University Press. Feinstein, C. H. (1988). Sources and methods of estimation for domestic reproducible fixed assets, stocks and works in progress, overseas assets, and land. In: C. H. Feinstein & S. Pollard (Eds), Studies in capital formation in the United Kingdom 1750–1920 (pp. 257–471). Oxford: Clarendon Press. Field, A. J. (2006). Technical change and US economic growth: The interwar period and the 1990s. In: P. W. Rhode & G. Toniolo (Eds), The global economy in the 1990s a long-run perspective (pp. 89–117). Cambridge: Cambridge University Press. Fraile Balbı´ n, P. (1991). Industrializacioń y grupos de presioń. La economıá polı´tica de la proteccioń en Espanã, 1900–1950. Madrid: Alianza. Fraile Balbı´ n, P. (1998). La reto´rica contra la competencia en Espanã (1875–1975). Madrid: Fundacioń Argentaria/Visor. Fraile Balbı´ n, P., & Escribano, A. (1998). The Spanish 1898 disaster: The drift toward, nationalprotectionism. In: P. K. O’Brien, & L. Prados de la L Escosura (Eds), The Costs and Benefits of European Imperialism from the Conquest of Ceuta, 1415, to the Treaty of Lusaka, 1974. Revista de Historia Econo´mica (special issue) 16, pp. 265–290. Fundacioń BBV. (1995). El stock de capital en Espanã y sus comunidades autońomas (Vol. 4). Madrid: Banco Bilbao Vizcaya.

176


Garcı´ a Santos, N., & Martı´ n Acenã, P. (1990). El comportamiento del gasto pu´blico en Espanã durante la II Repu´blica, 1931–1935. Revista de Historia Econo´mica, 8, 397–415. Go´mez Mendoza, A. (1989). Transportes y Comunicaciones. In: A. Carreras (Ed.), Estadı´sticas Histo´ricas de Espanã. Siglos XIX–XX (pp. 269–325). Madrid: Fundacioń Banco Exterior. Go´mez Villegas, J. (1988). La industria espanõla seguń su stock de capital: 1964–1981. Investigaciones econo´micas, 12(2), 337–398. Gordon, R. J. (1999). U.S. Economic growth since 1870: One big wave? American Economic Review Papers and Proceedings, 89(2), 123:28. Hall, R., & Jorgenson, D. W. (1967). Tax policy and investment behaviour. American Economic Review, 57, 391–414. Herranz-Loncań, A. (2007). Infrastructure investment and Spanish economic growth, 1850–1935. Explorations in Economic History, 44, 452–468. Hofman, A. (1993). The capital stock of Spain in the 20th century. EHES Workshop long-run Economic Growth in the European Periphery. La Corunã: Universidad Meneńdez Pelayo (Mimeo). Hulten, Ch. (1990). The measurement of capital. In: E. Berndt & J. Tripplett (Eds), Fifty years of economic measurement. The Jubilee of the conference on research in income and wealth (pp. 119–152). Chicago, IL: University of Chicago Press/NBER. Hulten, Ch. (1992). Growth accounting when technical change is embodied in capital. American Economic Review, 82(4), 964–980. Hulten, Ch., & Wykoff, F. (1981). The estimation of economic depreciation using vintage asset prices. Journal of Econometrics, l.15, 367–396. Jorgenson, D. (1973). The economic theory of replacement and depreciation. In: W. Sellekaerts (Ed.), Econometrics and economic theory (pp. 189–221). New York: Macmillan. Jorgenson, D. (1980). Accounting for capital. In: G. von Furstenberg (Ed.), Capital efficiency and growth (pp. 251–319). Ballinger, TX: Cambridge. Jorgenson, D. (1989). Capital as a factor of production. In: D. J. Jorgenson & R. Landau (Eds), Technology and capital formation (pp. 1–35). Cambridge, MA: MIT Press. Jorgenson, D. (1990). Productivity and economic growth. In: E. Berndt & J. Tripplett (Eds), Fifty years of economic measurement. The jubilee of the conference on research in income and wealth (pp. 19–118). Chicago, IL: University of Chicago Press/NBER. Jorgenson, D., Gollop, F., & Fraumeni, B. (1987). Productivity and US economic growth. Cambridge, MA: Harvard University Press. Kaldor, N. (1961). Capital accumulation and economic growth. In: F. A. Lutz & D. Hague (Eds), The theory of capital (pp. 177–222). London: Macmillan. Lo´pez Carrillo, J. M. (1998), Autarquıá y automocioń: Evolucioń de la empresa nacional de autocamiones (ENASA) entre 1946 y 1958. Fundacioń Empresa Pu´blica. Programa de Historia Econo´mica. Documento de trabajo 9809. Maddison, A. (1991). Dynamic forces in capitalist development. A long-run comparative view. Oxford: Oxford University Press. Maddison, A. (1995). Standardised estimates of fixed capital stock: A six country comparison. In: A. Maddison (Ed.), Explaining the economic performance of nations. Essays in time and space (pp. 137–166). Aldershot: Edward Elgar. Maluquer de Motes Bernet, J. (1999). El impacto de las guerras coloniales de fin de siglo sobre la economı´ a espanõla. In: P. Tedde (Ed.), Economıá y colonias en la Espanã del 98 (pp. 101–121). Madrid: Sı´ ntesis.


177

Martı´ n Acenã, P. (1994). Spain during the classical gold standard years, 1880–1914. In: M. Bordo & F. H. pie (Eds), Monetary re´gimes in transition (pp. 135–172). Cambridge: Cambridge University Press. Mas, M., Pe´rez, F., & Uriel, E. (2000). Estimation of the stock of capital in Spain. Review of Income and Wealth, 46(1), 103–116. Mas, M., Pe´rez, F., & Uriel, E. (2005a). El stock de capital en Espanã y su distribucioń territorial (1964–2002). Bilbao: Fundacioń BBVA. Mas, M., Pe´rez, F., & Uriel, E. (2005b). El stock de capital y los servicios del capital en Espanã (1964–2002). Nueva metodologıá. Bilbao: Fundacioń BBVA. Mas, M., Pe´rez, F., & Uriel, E. (2006). Capital stock in Spain, 1964–2002: New Estimates. In: M. Mas & P. Schreyer (Eds), Growth, capital, and new technologies (pp. 67–91). Bilbao: Fundacioń BBVA. Matthews, R. C. O., Feinstein, C. H., & Odling-Smee, J. (1982). British economic growth, 1856–1973. Oxford: Oxford University Press. Munõz Rubio, M. (1995). RENFE (1941–1991). Medio siglo de ferrocarril pu´blico. Madrid: Ediciones Luna. Myro, R. (1983). La evolucioń de la productividad global de la economı´ a espanõla en el perı´ odo 1965–1981. Informacioń Comercial Espanõla, 594, 115–127. OECD. (1993). Methods used by OECD countries to measure stocks of fixed capital. National Accounts: Sources and Methods no. 2, Paris: OECD. OECD. (2001a). Measuring capital measuring capital: Measurement of capital stock, consumption of capital and capital services. OECD manual. Paris: OECD. OECD. (2001b). Measuring productivity. Measurement of aggregate and industry-level productivity growth OECD manual. Paris: OECD. O’Mahony, M. (1996). Measures of fixed capital stocks in the post-war period: A five country study. In: B. Van Ark & N. Crafts (Eds), Quantitative aspects of post-war European economic growth (pp. 165–214). Cambridge: Cambridge University Press. Palafox, J. (1991). Atraso y democracia. La Segunda Repu´blica y la economıá espanõla, 1892–1936. Barcelona: Crı´ tica. Palafox, J. (1999). Proteccionismo y librecambio. In: G. Anes (Ed.), Historia econo´mica de espanã (Siglos XIX–XX) (pp. 251–287). Barcelona, Spain: Galaxia Gutemberg/Cı´ rculo de Lectores. Pardos, E. (1998). La incidencia de la proteccioń arancelaria en los mercados espanõles (1870–1913). Madrid: Banco de Espanã. Pe´rez Moreda, V. (1999). Poblacioń y economı´ a en la Espanã de los siglos XIX y XX. In: G. Anes (Ed.), Historia Econo´mica de Espanã. Siglos XIX y XX (pp. 7–62). Barcelona: Galaxia Gutemberg/Cı´ rculo de Lectores. Prados de la Escosura, L. (2003). El progreso de Espanã, 1850–2000. Madrid: Fundacioń BBVA. Prados de la Escosura, L. (2007). Growth and structural change in Spain, 1850–2000: A European perspective. Revista de Historia Econo´mica/Journal of Iberian and Latin American Economic History, 25(1), 147–181. Prados de la Escosura, L. (2009). Spain’s international position, 1850–1935: Evidence from a Balance of Payments Reconstruction. Universidad Carlos III Working Papers in Economic History 09-1. Prados de la Escosura, L., & Rose´s, J. R. (2008). Proximate sources of economic growth in Spain, 1850–2000. Universidad Carlos III Working Papers in Economic History 08-12.

178


Prados de la Escosura, L., & Sanz, J. C. (1996). Growth and macroeconomic performance in Spain (1939–1993). In: N. Crafts & G. Toniolo (Eds), Economic growth in Europe since 1945 (pp. 355–387). Cambridge: Cambridge University Press and CEPR. Roldań, S., & Garcı´ a Delgado, J. L. (1973). La formacioń de la sociedad capitalista en Espanã, 1914–1920 (Vol. 2). Madrid: Cajas de Ahorro Confederadas. Ros Hombravella, J., Clavera, J., Esteban, J. M., Mone´s, M. A., & Montserrat, A. (1973). De la autarquıá a la estabilizacioń (1939–1959) (Vol. 2). Madrid: Edicusa. Rostow, W. W. (1956). The take-off into self-sustained growth. Economic Journal, 66(261), 25–48. Sanchis, T. (2006). Spanish economic ‘miracle’: A disaggregated approach to productivity growth, 1958–1975. Revista de Historia Econo´mica/Journal of Iberian and Latin American Economic History, 24(2), 383–417. Sudria`, C. (1990). Los beneficios de Espanã durante la Gran Guerra. Una aproximacioń a la balanza de pagos espanõla, 1914–1920. Revista de Historia Econo´mica, 8(2), 363–396. Tafunell, X. (1989). Construccioń. In: A. Carreras (Ed.), Estadı´sticas Histo´ricas de Espanã. Siglos XIX–XX (pp. 249–267). Madrid: Fundacioń Banco Exterior. Tafunell, X. (2001). La rentabilidad financiera de la empresa espanõla, 1880–1981: Una estimacioń en perspectiva sectorial. Revista de Historia Industrial, 18, 71–111. Tafunell, X. (2005). Empresa y bolsa. In: A. Carreras & X. Tafunell (Eds), Estadı´sticas Histo´ricas de Espanã, siglos XIX y XX (pp. 707–833). Bilbao: Fundacioń BBBV. Tena Junguito, A. (1999). Un nuevo perfil del proteccionismo espanõl durante la Restauracioń, 1875–1930. Revista de Historia Econo´mica, 17, 579–621. Timmer, M. P., Ypma, G., & van Ark, B. (2005). IT in the European Union: Driving productivity convergence? Research Memorandum GD-67, Groningen Growth and Development Centre, October 2003, Appendix Tables (updated). Universidad Comercial de Deusto. (1968–1972). Riqueza nacional de Espanã (Vol. 5). Bilbao: Universidad Comercial de Deusto. Velarde Fuertes, J. (1968). La polı´tica econo´mica de la Dictadura. Madrid: Guadiana. Young, A. (1995). The Tyranny of numbers: Confronting the statistical realities of the East Asian growth experience. Quarterly Journal of Economics, 90(2), 641–680.

APPENDIX In the Appendix (Tables A1–A7) we provide alternative series of capital stock and its composition by asset type. In addition, indices of capital stock and input and capital quality are presented together with a capital stock/ output ratio.

179


Table A1.

Capital Stock Estimates with Alternative Initial Levels, 1850–1913 (1958 Million Pesetas). Zero Initial Capital Stock

1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889

3,591 7,549 12,050 16,553 20,158 23,683 28,063 33,989 43,287 52,359 63,326 73,178 84,102 94,355 102,790 109,607 116,005 122,713 125,208 127,376 130,272 134,218 138,856 142,149 145,922 149,755 156,458 163,651 173,081 179,585 188,287 195,425 205,107 217,747 228,437 235,485 242,252 248,680 254,112 260,580

Double Initial Capital Stock 78,779 81,345 84,494 87,681 90,006 92,284 95,451 10,0194 108,338 116,284 126,153 134,934 144,810 154,041 161,475 167,315 172,756 178,529 180,108 181,380 183,398 186,484 190,279 192,747 195,710 198,750 204,675 211,105 219,785 225,555 233,536 239,966 248,954 260,912 270,932 277,322 283,443 289,236 294,044 299,900

Initial Capital Stock (Expression 3) 41,185 44,447 48,272 52,117 55,082 57,984 61,757 67,091 75,813 84,322 94,739 104,056 114,456 124,198 132,132 138,461 144,380 150,621 152,658 154,378 156,835 160,351 164,568 167,448 170,816 174,252 180,566 187,378 196,433 202,570 210,912 217,696 227,031 239,330 249,684 256,403 262,848 268,958 274,078 280,240

180


Table A1. (Continued ) Zero Initial Capital Stock 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932

267,911 274,964 281,885 288,138 294,003 301,495 308,320 315,488 321,608 333,993 349,784 361,412 370,370 379,675 389,659 398,813 408,774 420,245 430,860 442,822 456,309 471,011 490,564 511,469 529,826 539,828 549,530 557,978 564,641 574,280 584,386 598,208 612,192 628,229 650,817 671,351 698,741 728,481 768,020 813,500 855,776 875,644 888,555

Double Initial Capital Stock 306,628 313,089 319,429 325,110 330,413 337,352 343,633 350,266 355,862 367,729 383,012 394,140 402,606 411,427 420,935 429,621 439,121 450,139 460,307 471,831 484,885 499,162 518,296 538,789 556,741 566,344 575,654 583,714 589,997 599,262 608,997 622,455 636,080 651,763 674,004 694,196 721,249 750,658 789,871 835,030 876,989 896,547 909,151

Initial Capital Stock (Expression 3) 287,269 294,026 300,657 306,624 312,208 319,424 325,976 332,877 338,735 350,861 366,398 377,776 386,488 395,551 405,297 414,217 423,947 435,192 445,584 457,326 470,597 485,087 504,430 525,129 543,283 553,086 562,592 570,846 577,319 586,771 596,692 610,331 624,136 639,996 662,411 682,773 709,995 739,570 778,946 824,265 866,382 886,095 898,853

181


Table A1. (Continued )

1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958

Zero Initial Capital Stock

Double Initial Capital Stock

Initial Capital Stock (Expression 3)

904,799 919,815 935,764 922,785 901,923 881,283 863,193 869,796 886,228 909,284 930,374 949,477 970,463 991,060 1,014,105 1,051,163 1,087,654 1,121,278 1,146,525 1,182,352 1,225,464 1,271,910 1,337,386 1,408,822 1,474,265 1,550,868

925,093 939,813 955,469 942,067 920,793 899,755 881,277 887,619 903,794 926,596 947,436 966,293 987,037 1,007,394 1,030,204 1,067,030 1,103,292 1,136,692 1,161,717 1,197,326 1,240,222 1,286,456 1,351,724 1,422,954 1,488,194 1,564,598

914,946 929,814 945,616 932,426 911,358 890,519 872,235 878,708 895,011 917,940 938,905 957,885 978,750 999,227 1,022,155 1,059,097 1,095,473 1,128,985 1,154,121 1,189,839 1,232,843 1,279,183 1,344,555 1,415,888 1,481,229 1,557,733

Table A2. Capital Stock Estimates Constructed with Alternative De-aggregation of GFCF Series (1958 Million Pesetas). Disaggregated

1850 1851 1852 1853 1854 1855 1856 1857 1858 1859

79,274 82,399 86,185 90,166 93,095 95,936 99,782 105,602 115,681 125,488

Aggregated (CNE589) 79,179 82,175 85,809 89,465 92,171 94,851 98,473 103,839 112,887 121,754

Aggregated (CNE70) 78,779 81,345 84,494 87,681 90,006 92,284 95,451 100,194 108,338 116,284

182


Table A2. (Continued ) Disaggregated

1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902

137,365 148,344 160,578 172,248 182,156 190,388 198,400 207,041 209,297 211,153 213,953 217,531 222,390 225,385 229,545 233,251 240,942 248,758 261,242 268,418 278,774 286,227 296,253 311,069 324,137 332,336 340,333 347,576 353,424 360,364 368,508 375,679 382,877 389,209 395,165 402,763 409,652 416,814 422,918 434,927 450,638 463,256 473,701

Aggregated (CNE589) 132,854 142,713 153,741 164,040 172,373 178,923 185,048 191,548 193,550 195,199 197,608 201,188 205,521 208,433 211,879 215,444 222,090 229,308 238,855 245,282 254,006 261,085 270,900 283,922 294,887 301,952 308,747 315,239 320,712 327,173 334,539 341,624 348,583 354,930 360,795 368,394 375,235 382,408 388,553 401,314 417,691 429,681 438,833

Aggregated (CNE70) 126,153 134,934 144,810 154,041 161,475 167,315 172,756 178,529 180,108 181,380 183,398 186,484 190,279 192,747 195,710 198,750 204,675 211,105 219,785 225,555 233,536 239,966 248,954 260,912 270,932 277,322 283,443 289,236 294,044 299,900 306,628 313,089 319,429 325,110 330,413 337,352 343,633 350,266 355,862 367,729 383,012 394,140 402,606

183



1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945

Disaggregated

Aggregated (CNE589)

484,483 496,142 507,066 518,643 532,235 543,233 555,862 570,160 586,451 609,507 632,514 654,521 665,647 676,815 687,850 697,559 708,900 719,057 732,833 748,291 765,684 791,757 812,911 842,822 874,904 919,629 968,873 1,011,587 1,030,251 1,041,049 1,056,941 1,071,159 1,086,270 1,067,772 1,042,066 1,017,402 995,819 1,000,696 1,015,656 1,038,267 1,058,240 1,076,276 1,095,922

448,358 458,620 467,949 477,986 489,690 500,485 512,722 526,725 542,068 562,470 584,056 603,245 613,599 623,363 632,065 638,767 648,784 659,427 673,743 688,606 705,091 728,747 749,589 777,944 808,469 849,530 896,480 939,081 958,720 970,917 987,091 1,002,120 1,017,985 1,005,129 984,089 963,317 945,123 951,049 966,715 989,434 1,009,925 1,028,482 1,048,646

Aggregated (CNE70) 411,427 420,935 429,621 439,121 450,139 460,307 471,831 484,885 499,162 518,296 538,789 556,741 566,344 575,654 583,714 589,997 599,262 608,997 622,455 636,080 651,763 674,004 694,196 721,249 750,658 789,871 835,030 876,989 896,547 909,151 925,093 939,813 955,469 942,067 920,793 899,755 881,277 887,619 903,794 926,596 947,436 966,293 987,037

184



1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958

Table A3.

Disaggregated

Aggregated (CNE589)

Aggregated (CNE70)

1,115,216 1,136,726 1,172,210 1,207,062 1,239,097 1,262,778 1,297,149 1,338,746 1,383,824 1,448,050 1,518,148 1,582,271 1,657,683

1,068,645 1,091,021 1,127,582 1,163,557 1,196,704 1,221,495 1,256,825 1,299,377 1,345,411 1,410,579 1,481,783 1,547,056 1,623,551

1,007,394 1,030,204 1,067,030 1,103,292 1,136,692 1,161,717 1,197,326 1,240,222 1,286,456 1,351,724 1,422,954 1,488,194 1,564,598

Capital Stock Estimates with Alternative Depreciation Rates (Billion 1995 Pesetas). Modified Geometric Depreciation Stock

1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871

1,820 1,883 1,959 2,030 2,083 2,137 2,203 2,307 2,474 2,651 2,877 3,063 3,270 3,461 3,615 3,728 3,829 3,935 3,986 4,026 4,077 4,154

Linear Depreciation

Gross stock 1,849 1,941 2,048 2,150 2,237 2,325 2,425 2,566 2,770 2,989 3,261 3,495 3,754 4,002 4,216 4,392 4,559 4,734 4,857 4,968 5,089 5,236

Net stock 1,819 1,910 2,015 2,116 2,200 2,286 2,386 2,524 2,726 2,941 3,208 3,438 3,693 3,935 4,144 4,316 4,480 4,650 4,770 4,879 4,998 5,142

185


Table A3. (Continued ) Modified Geometric Depreciation

1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914

Linear Depreciation

Stock

Gross stock

Net stock

4,244 4,316 4,387 4,472 4,596 4,746 4,900 5,026 5,178 5,326 5,513 5,735 5,934 6,072 6,200 6,327 6,447 6,575 6,724 6,872 7,025 7,161 7,295 7,454 7,601 7,750 7,884 8,102 8,393 8,619 8,813 9,013 9,229 9,414 9,598 9,813 10,041 10,282 10,560 10,871 11,239 11,627 11,981

5,397 5,539 5,682 5,841 6,034 6,256 6,471 6,669 6,900 7,124 7,386 7,684 7,965 8,185 8,395 8,602 8,829 9,064 9,318 9,574 9,833 10,086 10,330 10,610 10,864 11,132 11,355 11,691 12,070 12,410 12,707 12,987 13,302 13,599 13,897 14,220 14,563 14,908 15,248 15,628 16,073 16,547 17,019

5,300 5,439 5,580 5,736 5,927 6,145 6,356 6,550 6,777 6,997 7,255 7,548 7,823 8,038 8,244 8,447 8,671 8,902 9,151 9,402 9,657 9,905 10,145 10,421 10,669 10,933 11,152 11,484 11,856 12,189 12,480 12,754 13,066 13,358 13,651 13,968 14,304 14,643 14,977 15,351 15,788 16,253 16,716

186



1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957

Linear Depreciation

Stock

Gross stock

Net stock

12,212 12,407 12,546 12,637 12,778 12,977 13,240 13,520 13,864 14,279 14,729 15,262 15,838 16,534 17,345 18,144 18,481 18,714 18,986 19,229 19,490 19,333 19,060 18,780 18,515 18,613 18,952 19,383 19,844 20,268 20,683 21,103 21,621 22,427 23,186 23,891 24,409 25,093 25,873 26,784 28,067 29,468 30,710

17,393 17,744 18,037 18,296 18,571 18,857 19,211 19,593 20,058 20,608 21,189 21,835 22,489 23,267 24,141 25,010 25,473 25,778 26,121 26,468 26,927 26,633 26,203 25,769 25,359 25,570 25,991 26,511 27,089 27,593 28,116 28,593 29,139 29,909 30,615 31,281 31,826 32,563 33,380 34,357 35,759 37,362 38,905

17,079 17,424 17,710 17,963 18,231 18,494 18,842 19,217 19,673 20,213 20,781 21,415 22,055 22,819 23,672 24,518 24,960 25,251 25,585 25,922 26,370 26,662 26,219 25,777 25,362 25,057 25,471 25,982 26,548 27,040 27,555 28,019 28,557 29,317 30,012 30,669 31,205 31,931 32,733 33,691 35,071 36,645 38,151

187



1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000

Linear Depreciation

Stock

Gross stock

Net stock

32,191 33,498 34,767 36,403 38,294 40,471 43,085 46,291 50,041 54,084 58,534 63,440 68,465 73,248 79,090 86,059 93,720 100,991 108,033 114,907 121,419 127,460 132,394 136,945 141,549 145,797 149,315 153,250 157,930 163,928 171,353 180,323 190,005 199,683 208,399 215,274 222,467 230,621 238,745 247,305 257,119 268,112 279,756

40,723 42,381 44,087 46,159 48,534 51,203 54,352 58,114 62,505 67,230 72,424 78,158 84,013 89,852 96,915 105,223 114,500 123,591 132,669 141,778 150,712 159,338 166,993 174,057 181,073 187,768 193,765 200,160 207,505 215,912 225,472 236,474 248,580 261,142 273,002 283,351 294,293 306,477 318,973 332,186 347,052 363,847 381,507

39,931 41,545 43,121 45,145 47,458 50,057 53,129 56,805 61,099 65,699 70,758 76,347 82,046 87,719 94,622 102,727 111,758 120,564 129,366 138,203 146,863 155,219 162,618 169,448 176,260 182,756 188,565 194,799 201,972 210,166 219,475 230,187 241,956 254,144 265,615 275,594 286,234 298,082 310,168 322,941 337,332 353,572 370,594

188


Table A4. Alternative Capital Stock: Single and Chain Weighting (Billion 1995 Pesetas) (Modified Geometric Depreciation). Chain Weighting 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893

1,454 1,502 1,560 1,619 1,661 1,703 1,762 1,849 2,000 2,146 2,329 2,491 2,673 2,843 2,981 3,088 3,189 3,295 3,325 3,348 3,385 3,442 3,512 3,558 3,613 3,669 3,778 3,897 4,057 4,164 4,311 4,430 4,595 4,816 5,001 5,119 5,232 5,339 5,428 5,536 5,660 5,779 5,896 6,001

Single Weighting 1,820 1,883 1,959 2,030 2,083 2,137 2,203 2,307 2,474 2,651 2,877 3,063 3,270 3,461 3,615 3,728 3,829 3,935 3,986 4,026 4,077 4,154 4,244 4,316 4,387 4,472 4,596 4,746 4,900 5,026 5,178 5,326 5,513 5,735 5,934 6,072 6,200 6,327 6,447 6,575 6,724 6,872 7,025 7,161

189



1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939

Chain Weighting

Single Weighting

6,099 6,227 6,343 6,466 6,569 6,788 7,070 7,275 7,432 7,594 7,770 7,930 8,106 8,309 8,497 8,709 8,950 9,214 9,567 9,945 10,277 10,454 10,626 10,775 10,891 11,062 11,241 11,490 11,741 12,031 12,441 12,814 13,313 13,856 14,580 15,414 16,188 16,549 16,782 17,076 17,348 17,637 17,390 16,997 16,608 16,267

7,295 7,454 7,601 7,750 7,884 8,102 8,393 8,619 8,813 9,013 9,229 9,414 9,598 9,813 10,041 10,282 10,560 10,871 11,239 11,627 11,981 12,212 12,407 12,546 12,637 12,778 12,977 13,240 13,520 13,864 14,279 14,729 15,262 15,838 16,534 17,345 18,144 18,481 18,714 18,986 19,229 19,490 19,333 19,060 18,780 18,515

190



1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985

Chain Weighting

Single Weighting

16,384 16,683 17,104 17,489 17,837 18,220 18,595 19,016 19,696 20,366 20,982 21,444 22,101 22,893 23,747 24,951 26,266 27,470 28,881 30,118 31,292 32,821 34,594 36,639 39,091 41,951 45,524 49,461 53,803 58,469 63,337 67,885 73,652 80,710 88,600 96,031 103,131 110,054 116,544 122,479 127,225 131,513 135,928 139,892 142,935 146,591

18,613 18,952 19,383 19,844 20,268 20,683 21,103 21,621 22,427 23,186 23,891 24,409 25,093 25,873 26,784 28,067 29,468 30,710 32,191 33,498 34,767 36,403 38,294 40,471 43,085 46,291 50,041 54,084 58,534 63,440 68,465 73,248 79,090 86,059 93,720 100,991 108,033 114,907 121,419 127,460 132,394 136,945 141,549 145,797 149,315 153,250

191



1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000

Chain Weighting

Single Weighting

151,218 157,689 165,993 176,182 186,982 197,672 207,207 214,151 221,629 230,521 239,453 249,075 260,287 273,682 287,661

157,930 163,928 171,353 180,323 190,005 199,683 208,399 215,274 222,467 230,621 238,745 247,305 257,119 268,112 279,756

Table A5. Capital Stock and Its Composition (Billion 1995 Pesetas) (Modified Geometric Depreciation) (Chain Weighting). Dwellings

1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870

961 999 1,045 1,086 1,116 1,145 1,183 1,241 1,340 1,444 1,555 1,638 1,736 1,828 1,905 1,964 2,021 2,080 2,102 2,118 2,147

Other Constructions 446 454 465 479 490 502 517 541 575 607 675 745 817 882 931 964 992 1,022 1,037 1,050 1,059

Transportation Equipment 15 16 17 20 21 21 26 31 46 54 57 64 74 84 93 106 120 137 128 122 121

Machinery and Equipment 32 32 33 34 35 35 36 37 39 41 42 44 47 50 52 54 56 57 57 57 58

Capital Stock 1,454 1,502 1,560 1,619 1,661 1,703 1,762 1,849 2,000 2,146 2,329 2,491 2,673 2,843 2,981 3,088 3,189 3,295 3,325 3,348 3,385

192


Table A5. (Continued ) Dwellings

1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913

2,189 2,239 2,279 2,320 2,364 2,433 2,510 2,603 2,667 2,744 2,813 2,906 3,025 3,126 3,192 3,248 3,299 3,340 3,397 3,460 3,523 3,588 3,653 3,719 3,796 3,868 3,944 4,010 4,133 4,294 4,414 4,513 4,615 4,729 4,827 4,919 5,030 5,142 5,268 5,394 5,542 5,717 5,910

Other Constructions 1,075 1,093 1,104 1,117 1,134 1,165 1,203 1,240 1,275 1,323 1,369 1,427 1,492 1,554 1,597 1,647 1,699 1,754 1,800 1,856 1,910 1,963 2,004 2,039 2,086 2,127 2,167 2,203 2,259 2,334 2,394 2,445 2,497 2,548 2,596 2,653 2,716 2,779 2,843 2,940 3,038 3,167 3,289

Transportation Equipment 118 118 111 112 105 114 114 141 143 157 153 159 187 201 204 206 205 195 194 190 185 175 170 162 159 155 153 148 178 208 221 217 215 214 218 229 242 237 244 245 246 273 304

Machinery and Equipment 60 62 63 64 65 67 70 74 78 86 94 103 112 120 126 132 136 139 146 154 162 170 173 179 186 193 202 207 218 234 246 257 267 278 289 305 321 338 355 372 388 411 443

Capital Stock 3,442 3,512 3,558 3,613 3,669 3,778 3,897 4,057 4,164 4,311 4,430 4,595 4,816 5,001 5,119 5,232 5,339 5,428 5,536 5,660 5,779 5,896 6,001 6,099 6,227 6,343 6,466 6,569 6,788 7,070 7,275 7,432 7,594 7,770 7,930 8,106 8,309 8,497 8,709 8,950 9,214 9,567 9,945

193


Table A5. (Continued ) Dwellings

1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957

6,085 6,192 6,283 6,357 6,410 6,482 6,580 6,691 6,813 6,972 7,183 7,385 7,641 7,923 8,285 8,695 9,075 9,165 9,212 9,267 9,318 9,373 9,312 9,193 9,079 8,983 8,968 9,124 9,319 9,523 9,680 9,853 10,046 10,309 10,741 11,121 11,473 11,728 12,050 12,422 12,902 13,566 14,295 14,932

Other Constructions 3,401 3,474 3,539 3,582 3,609 3,662 3,730 3,839 3,954 4,072 4,204 4,361 4,539 4,723 4,931 5,179 5,436 5,627 5,770 5,951 6,112 6,283 6,200 6,076 5,942 5,800 5,908 6,016 6,173 6,335 6,524 6,679 6,818 6,953 7,145 7,357 7,558 7,707 7,924 8,157 8,380 8,720 9,084 9,420

Transportation Equipment 328 317 317 343 370 407 428 448 470 473 524 517 554 590 682 772 807 805 778 782 793 803 732 637 555 495 484 476 504 492 472 460 460 445 451 457 461 464 475 488 511 559 614 674


Capital Stock

462 471 487 493 503 512 504 512 504 514 530 552 579 620 682 767 871 953 1,022 1,077 1,125 1,177 1,146 1,090 1,032 989 1,025 1,067 1,107 1,138 1,161 1,229 1,272 1,309 1,359 1,430 1,490 1,544 1,653 1,827 1,953 2,108 2,274 2,445

10,277 10,454 10,626 10,775 10,891 11,062 11,241 11,490 11,741 12,031 12,441 12,814 13,313 13,856 14,580 15,414 16,188 16,549 16,782 17,076 17,348 17,637 17,390 16,997 16,608 16,267 16,384 16,683 17,104 17,489 17,837 18,220 18,595 19,016 19,696 20,366 20,982 21,444 22,101 22,893 23,747 24,951 26,266 27,470

194



1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000

Dwellings

Other Constructions



Capital Stock

15,710 16,432 17,138 18,055 19,035 20,104 21,368 22,939 24,684 26,475 28,244 30,271 32,357 34,298 36,665 39,526 42,655 45,556 48,401 51,178 53,727 56,025 57,892 59,680 61,453 63,008 64,279 65,653 67,177 69,118 71,542 74,213 77,014 79,637 82,031 84,002 86,087 88,505 91,261 94,080 97,330 101,282 105,428

9,800 10,099 10,381 10,716 11,204 11,859 12,682 13,671 14,917 16,357 18,242 20,224 22,226 24,190 26,576 29,370 32,443 35,481 38,379 41,224 44,018 46,709 48,922 50,886 52,933 54,916 56,566 58,434 60,765 63,677 67,180 71,697 76,678 81,887 86,514 90,203 93,987 98,053 101,546 105,310 109,537 114,620 119,993

751 841 915 982 1,041 1,131 1,255 1,364 1,508 1,650 1,843 2,026 2,232 2,382 2,645 3,066 3,590 4,120 4,595 5,087 5,480 5,800 5,994 6,134 6,356 6,444 6,327 6,284 6,333 6,603 7,039 7,591 8,078 8,488 8,763 8,775 8,899 9,136 9,316 9,653 10,126 13,003 15,810

2,620 2,745 2,859 3,068 3,315 3,546 3,786 3,977 4,415 4,979 5,474 5,947 6,523 7,016 7,766 8,749 9,911 10,874 11,754 12,565 13,318 13,946 14,417 14,813 15,185 15,524 15,763 16,220 16,943 18,291 20,231 22,682 25,212 27,660 29,900 31,172 32,657 34,827 37,330 40,032 43,294 44,778 46,429

28,881 30,118 31,292 32,821 34,594 36,639 39,091 41,951 45,524 49,461 53,803 58,469 63,337 67,885 73,652 80,710 88,600 96,031 103,131 110,054 116,544 122,479 127,225 131,513 135,928 139,892 142,935 146,591 151,218 157,689 165,993 176,182 186,982 197,672 207,207 214,151 221,629 230,521 239,453 249,075 260,287 273,682 287,661

195


Table A6.

Indices of Capital Stock, Input, and Quality of Capital (1850 ¼ 100). Capital Input

1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893

100 103 107 111 114 117 122 129 143 154 166 179 194 209 221 232 242 254 253 253 255 259 263 264 268 270 279 287 304 312 327 335 350 372 389 398 407 415 419 427 437 445 452 458

Capital Stock 100 103 107 111 114 117 121 127 138 148 160 171 184 196 205 212 219 227 229 230 233 237 242 245 248 252 260 268 279 286 296 305 316 331 344 352 360 367 373 381 389 397 405 413

Capital Quality 1.00 1.00 0.99 1.00 1.00 1.00 1.01 1.01 1.04 1.04 1.04 1.05 1.06 1.07 1.08 1.09 1.10 1.12 1.11 1.10 1.10 1.09 1.09 1.08 1.08 1.07 1.07 1.07 1.09 1.09 1.10 1.10 1.11 1.12 1.13 1.13 1.13 1.13 1.12 1.12 1.12 1.12 1.11 1.11

196



1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939

Capital Input

Capital Stock

Capital Quality

464 472 480 489 496 518 545 563 574 587 600 613 629 648 662 681 700 722 754 791 820 833 847 862 875 894 908 930 950 972 1,010 1,039 1,083 1,132 1,206 1,290 1,365 1,404 1,428 1,459 1,488 1,519 1,483 1,429 1,377 1,336

419 428 436 445 452 467 486 500 511 522 534 545 557 571 584 599 616 634 658 684 707 719 731 741 749 761 773 790 807 827 856 881 916 953 1,003 1,060 1,113 1,138 1,154 1,174 1,193 1,213 1,196 1,169 1,142 1,119

1.11 1.10 1.10 1.10 1.10 1.11 1.12 1.13 1.12 1.12 1.12 1.12 1.13 1.13 1.13 1.14 1.14 1.14 1.15 1.16 1.16 1.16 1.16 1.16 1.17 1.18 1.17 1.18 1.18 1.17 1.18 1.18 1.18 1.19 1.20 1.22 1.23 1.23 1.24 1.24 1.25 1.25 1.24 1.22 1.21 1.19

197



1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985

Capital Input

Capital Stock

Capital Quality

1,349 1,373 1,413 1,442 1,467 1,500 1,532 1,563 1,616 1,672 1,723 1,762 1,822 1,900 1,978 2,087 2,206 2,320 2,451 2,565 2,670 2,812 2,977 3,167 3,392 3,630 3,968 4,352 4,766 5,193 5,647 6,061 6,601 7,270 8,022 8,720 9,381 10,024 10,623 11,166 11,597 11,982 12,381 12,735 12,997 13,324

1,127 1,147 1,176 1,203 1,227 1,253 1,279 1,308 1,354 1,400 1,443 1,475 1,520 1,574 1,633 1,716 1,806 1,889 1,986 2,071 2,152 2,257 2,379 2,520 2,688 2,885 3,131 3,401 3,700 4,021 4,356 4,668 5,065 5,550 6,093 6,604 7,092 7,568 8,014 8,423 8,749 9,044 9,347 9,620 9,829 10,081

1.20 1.20 1.20 1.20 1.20 1.20 1.20 1.20 1.19 1.19 1.19 1.19 1.20 1.21 1.21 1.22 1.22 1.23 1.23 1.24 1.24 1.25 1.25 1.26 1.26 1.26 1.27 1.28 1.29 1.29 1.30 1.30 1.30 1.31 1.32 1.32 1.32 1.32 1.33 1.33 1.33 1.32 1.32 1.32 1.32 1.32

198



1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000

Capital Input

Capital Stock

Capital Quality

13,750 14,371 15,178 16,170 17,211 18,236 19,151 19,795 20,499 21,358 22,233 23,194 24,337 25,684 27,080

10,399 10,844 11,415 12,116 12,858 13,594 14,249 14,727 15,241 15,852 16,467 17,128 17,899 18,821 19,782

1.32 1.33 1.33 1.33 1.34 1.34 1.34 1.34 1.34 1.35 1.35 1.35 1.36 1.36 1.37

Table A7.

Capital Stock–Output Ratio.

Capital Stock/Output 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871

0.8 0.8 0.8 0.8 0.8 0.8 0.9 0.9 1.0 1.0 1.1 1.1 1.2 1.2 1.3 1.4 1.4 1.4 1.6 1.5 1.5 1.4

Capital Stock/Output 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947

2.1 2.0 2.1 2.0 2.2 2.3 2.3 2.4 2.3 2.3 2.9 3.1 3.0 2.7 2.5 2.6 2.5 2.5 2.4 2.6 2.6 2.6

199


Table A7. (Continued ) Capital Stock/Output 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915

1.3 1.2 1.3 1.3 1.3 1.2 1.3 1.4 1.4 1.4 1.4 1.4 1.5 1.6 1.6 1.7 1.7 1.7 1.7 1.7 1.6 1.7 1.7 1.8 2.0 1.9 1.8 1.8 1.9 1.8 1.9 1.9 2.0 2.1 2.0 2.0 2.0 2.0 2.1 2.0 2.1 2.1 2.2 2.2

Capital Stock/Output 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991

2.7 2.8 2.8 2.5 2.4 2.5 2.4 2.4 2.3 2.4 2.4 2.5 2.6 2.4 2.3 2.2 2.2 2.2 2.2 2.2 2.3 2.3 2.4 2.4 2.4 2.4 2.4 2.5 2.6 2.7 2.8 2.9 2.9 3.0 3.0 3.0 3.1 3.1 3.1 3.0 3.0 3.1 3.1 3.2

200


Table A7. (Continued ) Capital Stock/Output 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925

2.1 2.2 2.2 2.2 2.1 2.1 2.0 2.1 2.1 2.0

Capital Stock/Output 1992 1993 1994 1995 1996 1997 1998 1999 2000

3.3 3.4 3.4 3.4 3.5 3.5 3.5 3.6 3.6

U.S. TRADE POLICY AND THE PACIFIC RIM, FROM FORDNEY–MCCUMBER TO THE TRADE EXPANSION ACT OF 1962: A POLITICAL–ECONOMIC ANALYSIS Lei (Sandy) Ye ABSTRACT This paper investigates the extent to which U.S. trade policies during 1922–1962 impacted the Pacific Rim economies differently from the rest of the world. Empirical analysis demonstrates that U.S. trade with the Pacific Rim had consistently higher tariff barriers than U.S. trade with the rest of the world among import-competitive manufactures. This paper then analyzes the reasons behind this phenomenon from both a political economy and a historical perspective. On both fronts, the Pacific Rim was at a disadvantage, and its higher barrier to trade with the United States was by no means historically accidental.

1. INTRODUCTION For much of U.S. history, policy makers largely rejected free trade. Before the twentieth century, protectionist tariffs were seen as crucial in developing Research in Economic History, Volume 27, 201–253 Copyright r 2010 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 0363-3268/doi:10.1108/S0363-3268(2010)0000027007

201

202

LEI (SANDY) YE

the nation’s manufacturing sector.1 However, the ‘‘infant industry’’ argument for protectionism became obsolete in the twentieth century when the United States became a leading industrial power. On economic grounds, a cutback on tariffs seemed timely, and rates did actually fall between 1912 and 1922. Politics, however, made this period short-lived, and trade barriers rose to all-time highs in the 1920s and 1930s. Alarmed by a sharp downward spiral in world trade, the United States again pursued trade liberalization in 1934 through bilateral agreements. In 1947, the General Agreement on Tariffs and Trade (GATT) established a multilateral framework for negotiations, and since then U.S. trade policy has continued to evolve toward lower barriers. Many works have closely studied U.S. trade policies during this period.2 However, they analyze policies from the perspective of the United States vis-a`-vis the rest of the world as a whole, with no particular emphasis on trading relations with particular regions. When these relations are discussed, the primary focus is almost always on Europe. Hardly anything has been written about Pacific trade. The Institute of Pacific Relations published some studies on U.S. trade with the Pacific Rim in the first half of the twentieth century, the most notable of which was by Wright (1935). Such studies provide comprehensive information regarding the region’s trade with the United States, but did not focus on the difference in Pacific Rim tariff barriers compared to those faced by the rest of the world. This paper focuses on U.S. trade policy from 1922 to 1962, and most importantly, its effect on a particular region – the Pacific Rim, defined as the economies of what constitute East Asia, Southeast Asia, Australia, and Oceania today.3 The Pacific Rim is a diverse group of nations that has had a long history of trade with the United States. It was an important supplier of raw materials after World War I. At the same time, its export of finished manufactures, such as textiles and wool, faced heavy competition from U.S. domestic sectors. Does the unique economic position of the Pacific Rim suggest that U.S. trade policy constructed higher trade barriers with that region than with the rest of the world? If so, why was this the case? In this paper, I address these questions through an integrated approach. In Section 2, I discuss the data showing the trends in U.S. imports from the Pacific Rim. Section 3 addresses whether the region encountered higher barriers to trade than those faced by the rest of the world. I present two original sets of estimates of average ad valorem tariff rates of imports from the Pacific Rim using two separate sources of data. Both measures show that average tariff rates of U.S. imports from the Pacific Rim were consistently higher than those from the rest of the world among import-competitive

U.S. Trade Policy and the Pacific Rim

203

high-end manufactures. I then explore the reasons for this first by conducting a political analysis of the voting pattern behind various trade policies using formal econometrics (Section 4). This analysis finds that Congressional support for protection was stronger and more persistent among domestic sectors that competed significantly with the Pacific Rim. Lastly, in Section 5, I move beyond the political economy aspects of trade policies and examine the historical geopolitical dimension of U.S. trade policies in relation to the Pacific Rim. This section shows that U.S. foreign policy priorities and the Pacific Rim’s regional political disarray disadvantaged the region in trade negotiations. This paper demonstrates that the absence of explicit geographic discrimination in international trade does not always preclude the possibility of implicit regional bias. In this case, such bias grew out of the historical circumstances of the two regions. Its form was influenced by the products that the Pacific Rim exported. U.S. domestic politics and foreign policy from 1922 to 1962 combined to produce relatively high tariff rates on U.S. import products supplied by Pacific Rim countries, even though most favored nation (MFN) clauses were standard in bilateral trade treaties. The experience of the Pacific Rim vis-a`-vis the United States during that period parallels that of developing countries vis-a`-vis developed countries more generally during the GATT/WTO era. The least developed countries (LDCs) today face many of the same dilemmas as those of the Pacific Rim in the middle of last century, as their export products are often those that are protected most heavily by the developed economies. Hence, regional bias in trade liberalization may have disproportionate consequences for the economic development of LDCs.

2. TRENDS IN U.S. IMPORT FLOW FROM THE PACIFIC RIM From 1922 to 1962, the Pacific Rim as a whole was a significant trading partner with the United States. As Table 1 shows, relative to all imports to the United States during that period, those from the Pacific Rim held a substantial share, reaching more than 25 percent in the peak years of 1925–1926 and 1940–1941. Even though the share of U.S. imports from the Pacific Rim lagged that of other close trading partners such as Europe, the Pacific Rim supplied a number of goods that were critical to key sectors in the United States. These goods usually consisted of raw materials. For instance, as a critical input in

204

LEI (SANDY) YE

Table 1.

Share of U.S. Imports from the Pacific Rim and its Subregions (in %).

Years

Pacific Rim

China and Japan

Southeast Asiaa

Australia and Oceania

1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962

22.0 21.8 20.3 25.1 25.5 23.8 21.6 22.6 20.5 20.0 18.0 19.2 20.2 21.6 22.1 24.5 19.6 21.9 28.3 27.7 14.0 7.8 3.7 4.4 11.6 11.4 10.8 10.4 12.4 14.3 12.2 9.4 8.7 10.7 10.6 10.5 10.2 13.0 13.9 13.0 14.8

15.8 14.2 12.8 13.2 12.4 13.3 12.9 13.7 12.5 13.1 12.2 11.6 10.0 10.9 10.3 10.1 9.0 9.7 9.7 5.0 0.6 0.4 0.3 0.1 3.6 2.6 2.6 2.9 3.7 2.3 2.4 2.4 2.7 3.8 4.4 4.6 5.2 6.8 7.8 7.2 8.3

4.1 5.5 5.7 9.7 11.4 8.8 7.1 7.3 6.6 5.6 4.9 6.4 9.0 9.0 10.0 11.8 9.6 10.9 17.2 17.9 5.0 0.2 0.1 0.2 4.2 6.0 5.8 5.5 6.2 7.7 7.2 4.7 4.0 5.1 4.2 3.9 2.9 3.2 3.1 2.5 2.3

1.6 1.6 1.4 1.8 1.5 1.3 1.3 1.3 1.1 0.9 0.6 0.9 0.9 1.3 1.5 2.2 0.8 1.2 1.3 4.8 8.4 7.3 3.3 4.1 3.7 2.7 2.3 1.9 2.4 4.1 2.3 1.9 1.6 1.5 1.6 1.7 1.6 2.2 1.8 2.2 2.7

Source: Statistical Abstract of the United States, various years. a Excluding the Philippines due to its special trading relationship with the United States.

205


the silk and rayon manufacturing industry, raw silk was imported almost exclusively from China and Japan. The Pacific Rim also exported some agricultural products. Wool, for example, was heavily imported from Australia and New Zealand. In addition, the Pacific Rim also exported finished manufactures, a high proportion of which were labor-intensive textile products such as fine silk manufactures from China and Japan. These products often competed with U.S. domestic manufactures and hence were subject to high levels of protection. The tariff barriers on finished manufactures will be the main focus of this paper. Between 1922 and 1962, high tariff rates on finished manufactures severely impacted imports from the Pacific Rim. Consider Figs. 1 and 2, which show U.S. imports of silk fabric as well as clothing and combing wool. The total value of imports (adjusted for inflation using an allcommodity wholesale price index from series Cc84 in Carter et al., 2006) shows strong negative correlation with the level of tariff barriers. The Underwood–Simmons Act in 1913 temporarily brought down tariff levels from an earlier era for about eight years, and during this period there were some signs of surges in importation levels. In contrast, the Fordney– McCumber Act of 1922 and later the infamous Smoot–Hawley Act of 1930

20,000,000

0.7

18,000,000

0.6

12,000,000

0.4

10,000,000 0.3

8,000,000 6,000,000

0.2

4,000,000 0.1

2,000,000 0

0

year China

Fig. 1.

Japan

Tariff Rate (% dutiable)

U.S. Import of Silk Fabric from China and Japan, 1900–1940.

tariff rate

0.5

14,000,000

1900 1902 1904 1906 1908 1910 1912 1914 1916 1918 1920 1922 1924 1926 1928 1930 1932 1934 1936 1938 1940

value (in 1926 $)

16,000,000

LEI (SANDY) YE 70,000,000

1.2

60,000,000

1

50,000,000 0.8 40,000,000 0.6 30,000,000

tariff rate

value (in 1926 $)

206

0.4 20,000,000 0.2

10,000,000

1939

1936

1933

1930

1927

1924

1921

1918

1915

1912

1909

1906

1903

0 1900

0

year Clothing Wool

Tariff Rate, Clothing Wool (% dutiable)

Combing Wool

Tariff Rate, Combing Wool (% dutiable)

Fig. 2. U.S. Import of Clothing and Combing Wool from Australia and New Zealand Combined, 1900–1940.

elevated tariff levels to all-time highs and in some cases virtually prohibitory levels, and subsequently the level of imports became dramatically lower. While reduction in imports cannot be attributed exclusively to the onset of the Fordney–McCumber Tariff, the policy nonetheless constituted a major barrier to trade. Subsequent sections consider whether there were any differential barriers to trade between the Pacific Rim and the rest of the world.

3. EMPIRICAL INVESTIGATION OF DIFFERENCES IN TARIFF RATES BETWEEN THE PACIFIC RIM AND THE REST OF THE WORLD In this section, I address these questions by presenting a new set of tariff rates for the Pacific Rim trade in 1926, 1928, 1931, 1935, 1940, and 1946. The first two years were under the tariff structure of the Fordney– McCumber Act; 1931 was one year after the Smoot–Hawley Act; 1935 was one year after the Reciprocal Trade Agreements Act (RTAA); and 1940 and

207


1946 were in the decade leading up to the GATT. While U.S. government publications report tariff rates of all U.S. imports in a given year, they do not provide any measure of the average tariff rate on imports from particular regions or countries. Hence, calculations must be made from available raw data on imports and duties. The average tariff rate is not a comprehensive measure of the level of protection, but in this era it was the most direct, visible, and accessible index available to policy makers.

3.1. Methodology In calculating this new set of tariff rates, my sources for the raw data are the annual volumes of the Foreign Commerce and Navigation of the United States (FCNUS) (U.S. Department of Commerce, Foreign and Domestic Commerce Bureau, 1900–1940, 1946). These volumes contain relatively complete information on import value of goods broken down by country and duties collected broken down by items. I have grouped the data into three categories: all goods, semi-manufactures, and finished manufactures. Semi-manufactures and finished manufactures represent selections of goods approximately as defined in the 1931 edition of FCNUS.4 To compute estimates for the average ad valorem tariff rates of imports from the Pacific Rim in year i, denoted ti, I use the following simple form for each category: P aj Dworld j j

¼ PP tpacific i j

k

I pacific jk

I pacific jk

where denotes the import value of good (or product group)5 j in the duties collected for good j from all country k in the Pacific Rim,6 Dworld j countries, and aj the percentage of good j (that was dutiable) imported from Pacific Rim in year i. In other words: P pacific I jk k aj ¼ world Ij In this estimate, I have assumed that aj is also equal to the percentage of duties collected from the Pacific Rim,7 because data on duties is then compared collected broken down by country are not available. tpacific i , the tariff rate of imports from the rest of the world excluding the to tr:o:w: i

208

LEI (SANDY) YE

Pacific Rim, which is denoted as: P world P Dj aj Dworld j j j r:o:w: ti ¼ P world P pacific Ij Ij j

j

Within each of the three broad categories of goods, both Pacific Rim and rest of the world tariff rates are calculated over the set of good j’s that comprise only dutiable goods as well as dutiable and free goods combined. In short, I assumed that tariffs on any given good are constant across countries. However, I am calculating average tariff levels for a region. Average tariff levels may differ if regions differed in the composition of goods exported, and these estimates explore whether average tariffs were higher for the Pacific Rim because it exported a disproportionate share of the goods that faced high tariffs. Table 2 shows the estimates for each year, with one section each representing all goods, semi-manufactures, and Table 2. Year

Tariff Rates (in %).

Pacific Rim

Rest of the World

Free and dutiable

Dutiable only

Free and dutiable

Dutiable only

4.7 5.4 7.8 7.5 4.8 11.2

37.3 37.2 48.0 40.8 34.9 36.7

16.4 15.5 20.2 20.3 15.6 9.7

39.5 38.9 53.8 43.1 35.7 24.2

3.5 2.8 6.9 2.4 2.9 3.9

21.0 18.5 18.7 20.0 10.6 7.2

8.6 8.4 5.8 9.6 7.3 5.7

26.0 26.0 23.3 25.1 16.5 9.7

Finished manufactures 1926 44.4 1928 44.3 1931 42.6 1935 43.7 1940 35.7 1946 23.8

48.0 48.6 49.0 47.7 40.2 29.8

25.5 25.0 23.9 22.6 15.9 13.5

39.6 40.9 46.6 40.6 31.9 25.1

All goods 1926 1928 1931 1935 1940 1946 Semi-manufactures 1926 1928 1931 1935 1940 1946


209

finished manufactures. This set of results serves to explain the difference in tariff barriers between the Pacific Rim and the rest of the world.

3.2. The Results The results as presented in Table 2 exhibit consistent trends. First, tariff rates for both rest of the world and the Pacific Rim were significantly higher for finished manufactures than semi-manufactures across all years. Hence, the positive relationship between the level of protection and the stage of manufacturing as described by Hawke (1975) for the years 1899 and 1904 persisted through World War II (WWII). The high tariff rates of 1926, 1928, and 1931 reflect the protectionist policies of the Fordney–McCumber and Smoot–Hawley Acts, while the lower tariff rates of 1935, 1940, and 1946 marked the downward trend in U.S. trade barriers beginning with the RTAA of 1934 (Bailey, Goldstein, & Weingast, 1997). When comparing the Pacific Rim and the rest of the world within each category, however, the difference is clear and quite consistent. Among all goods, the tariff rate for Pacific Rim was consistently lower when taking into account both free and dutiable goods. However, among only dutiable goods, the Pacific Rim had roughly the same or only slightly lower tariff rate than the rest of the world. The two disparate trends imply that relative to the rest of the world, the Pacific Rim was exporting proportionally more goods on the duty-free list to the United States. Hence, counting duty-free goods in the calculation of tariff rates dramatically lowers the tariff rate for the Pacific Rim. To conclude that Pacific Rim countries actually enjoyed a lower tariff barrier is nevertheless misleading, because most of the items on the dutyfree list were raw materials in primitive stages of manufacturing that the United States did not produce domestically anyway. For example, even during the protectionist era of 1922–1930, raw silk remained duty-free. Compared to Europe, the Pacific Rim was supplying much more of these raw materials relative to manufactures. Including these free goods in the calculation distorts the picture, because these materials were not targeted by protectionist motives. For semi-manufactures, Pacific Rim imports also show lower tariff rates than the rest of the world. With one exception, this was true for all the years excluding free goods or not. The story with semi-manufactures is largely the same. Many goods that were included in this category in FCNUS were also imported free of duty, including many goods that were in relatively primitive

210

LEI (SANDY) YE

stages of manufacturing. In fact, if we examine the table of the semimanufactures and compare them with the corresponding cells of the allgoods table, the semi-manufactures have a lower tariff rate in every instance. This implies that even though semi-manufactures excluded both raw materials such as metal ores and finished manufactures such as silk fabric, the exclusion of the latter goods had a greater weight. Hence, even more so in this category than the all-goods category, the lower tariff rates of Pacific Rim imports reflected the relatively heavier import of more primitive-stage goods. By excluding the effect of non-import-competitive raw or primitive-stage goods, results for the finished manufactures reveal the impact of U.S. trade policy on the group of goods that was truly targeted by U.S. tariff policy and that was in most direct competition with domestic industries. For every one of the six years, the tariff rates on Pacific Rim finished manufactures have been higher. The tariff differentials were much greater when free imports are included, implying that the Pacific Rim exported proportionally less free finished manufactures (of which there were not much) than the rest of the world. However, even the differences among just the dutiable goods were significant, ranging from 3 to 8 percentage points. This tariff rate differential suggests an implicit bias built in U.S. tariff legislation throughout this period. Among finished manufactures, what the Pacific Rim was supplying in much greater proportion relative to total exports to the United States were products such as textiles. At the same time, it was precisely these goods that received higher levels of protection. Items such as woolen goods and silk products were among the most heavily protected items on the U.S. import list. Hence, the relative weighting of these goods was much higher when calculating tariff rates on Pacific Rim imports. In contrast, finished manufactures that were more heavy-industry and technology-oriented such as machinery and metal manufactures were rarely imported from the Pacific Rim. For example, the imports of iron and steel advanced manufactures from the Pacific Rim had been no more than 10 percent of total for any of the years considered (FCNUS, various years). The manufactures of these heavy-industrial goods had a much more established presence in Europe. In addition, these goods also included most of the duty-free finished manufactures, which explains why the tariff differential widens when both free and dutiable imports are included. The results for finished manufactures are consistent with Hayford and Pasurka’s finding (1992) that the Fordney–McCumber and Smoot–Hawley tariff rates were positively correlated with both labor-intensive and


211

agricultural products. The same study finds that capital-intensive goods and those with natural resource abundance tend to have lower rates. Thus, even though U.S. trade policy throughout this period did not officially discriminate against imports from any particular region, the higher level of protection toward industries such as textiles implicitly imposed greater protection against Pacific Rim imports. It is probable that my calculation of Pacific Rim imports’ tariff rate underestimates the level of protection imposed on Pacific Rim countries among finished manufactures, if not all goods. Even within a group of goods like textiles, imports from the Pacific Rim tend to be those that were in higher stages of refinement, such as the hand-made silk laces from China (Wright, 1935). However, while the current estimates of Pacific Rim tariff rates assign more weight to imports of products or product groups that the region was exporting more heavily, they do not take into account tariff level differences within a product group. For example, if the Pacific Rim had to export about 50 percent of all silk products that the United States imports, then the current estimate would assume 50 percent of duties on all U.S. silk product imports were collected from the Pacific Rim. However, because the Pacific Rim exported relatively more silk laces than silk fabric, and silk laces had a higher tariff rate, the actual amount of duties attributed to the Pacific Rim should be higher. Hence, the true difference in the level of protection for Pacific Rim economies versus the rest of the world should actually be greater than presented. In order to obtain a sense of the magnitude of underestimation as well as to extend the analysis to 1960, I construct a similar set of estimates from a different data source. The Annual Report of the Secretary of the Treasury on the State of the Finances presents duties collected on dutiable imports for consumption for various years (U.S. Department of Treasury, 1939, 1941, 1945, 1947, 1950, 1951, 1953, 1955, 1957, 1961). This source lacks the details on import value of goods broken down by country that FCNUS provides. However, its greatest advantage is that the amount of duties collected is recorded according to region. Recognizing this feature, I compute the ad valorem equivalent tariff rates for all dutiable imports for consumption from 1938 to 1960 (see Table 3).8 The results are even more striking. In every year except for 1951, the tariff rate for the Pacific Rim was higher than that for Europe and the rest of the world, implying that higher tariff rates for the Pacific Rim persisted after 1946. Furthermore, comparing these rates for 1940 and 1946 to my previous estimates of the same measures, the latter significantly underestimated the gap in tariff rates. In 1940, the Treasury figures show a 10-percentage-point higher rate for the Pacific Rim than the rest of the world (including Europe),

212

LEI (SANDY) YE

Table 3.

Tariff Rates of Dutiable Imports for Consumption (in %).

Year

Pacific Rima

Europe

Rest of the World (Including Europe)

Rest of the World (Excluding Europe)

1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1959 1960

47.6 48.5 47.0 48.1 55.6 61.0 54.8 61.3 58.5 46.9 28.9 21.2 23.2 17.8 18.8 21.6 20.4 22.2 21.0 20.5 19.5 18.8

38.2 36.4 36.9 40.1 39.5 41.2 40.5 34.8 28.2 28.0 24.4 20.8 21.0 18.0 17.3 17.2 17.0 17.3 16.8 15.9 14.5 14.3

37.2 37.1 35.9 36.6 30.4 28.9 31.5 26.2 23.9 20.3 15.2 12.6 12.6 11.9 11.6 11.7 11.0 11.1 10.7 10.1 10.1 10.7

36.1 37.8 35.1 35.6 28.6 26.3 29.8 24.6 22.4 17.6 12.3 9.8 10.0 9.1 8.7 8.6 7.5 7.4 6.9 6.3 6.4 7.1

Source: Annual Report of the Secretary of the Treasury on the State of the Finances. Department of Treasury, various years. a The original source did not record value of duties in a given year from all countries in the Pacific Rim as defined in my previous estimate. Hence, the Pacific Rim’s scope varied some throughout different years. However, the tariff rates should not be affected greatly, because the source suggests that the countries that were not recorded had low levels of duties.

while my own estimates suggest they were about the same. In 1946, the Treasury figures show a 30-percentage-point difference, compared to my own estimate of 12 percentage points. Hence, we expect the gap in tariff rates among manufactured goods to be even higher than what my estimates show, even though the Treasury reports did not record the duties broken down by manufacturing stage such that precise calculations could be made. It should be noted that the average ad valorem tariff rate as computed in this section is only one possible measure of the level of trade barriers. In fact, there is no ‘‘true’’ measure of the correct tariff rate in economic terms. Using unweighted averages could be misleading, giving high weight to items whose importance for trade was minor. Weighted averages such as my


213

estimates are an improvement on this account. Trade-share weights, however, are nevertheless partly endogenous to the effects of tariff rates on trade flows. The most commonly used alternative is the ‘‘effective’’ tariff rate,9 which takes account of tariff rates on inputs. However, existing studies such as Hawke (1975) suggest that American effective tariff rates were highly correlated with average tariff rates during the historical period we are considering. If this is so, then the additional investment involved in calculating effective tariff rates may not contribute substantial new insights for the purpose of this paper. Overall, the data show that Pacific Rim countries faced consistently higher barriers to trade, primarily on finished goods. Their exports of raw materials or primitive-stage manufactures enjoyed low or no tariffs. However, the Pacific Rim faced implicit discrimination among the finished manufactures that were under direct import competition with the United States. What were the reasons behind this biased trend? In the next section, with a special focus on Congress, we analyze some political economy issues that help explain this difference in level of protection.

4. POLITICAL ECONOMY OF TRADE LEGISLATION: EVIDENCE FROM CONGRESSIONAL ROLL-CALL VOTES From 1922 to 1962, Congress was instrumental in guiding U.S. trade policy. Before the 1934 RTAA, Congress held the power of setting tariff rates on U.S. imports on an item-by-item basis.10 While the executive branch began to play a prominent role in trade policy after the RTAA, Congress remained a strong force. Most importantly, studies have emphasized that Congress could have severely impeded or destroyed trade liberalization efforts after 1934, and its support for trade liberalization was reluctant (see Bailey et al., 1997; Lusztig, 2004). For these reasons, this section investigates whether members of Congress were responsive to economic interests that competed relatively more heavily with the Pacific Rim, the relative importance of these interests compared to party affiliation, and how these trends changed from 1922 to 1962. In doing so, probit regressions (separated for House and Senate) were conducted on Congressional votes on the following major trade bills: 1922 Fordney–McCumber Act; 1930 Smoot–Hawley Act; 1934 RTAA; 1945, 1949, and 1955 RTAA renewals; and the Trade Expansion Act of 1962.

214

LEI (SANDY) YE

Because member of Congress voted for each of these bills in its entirety as opposed to individual tariff rates on products, there is no direct way of discerning each member’s disposition on trade with specific geographic regions. However, what this section will demonstrate is that Congress was persistently more protectionist toward sectors that competed intensively with the Pacific Rim. As discussed in Section 2, the Fordney–McCumber and Smoot–Hawley Acts were two of the most protectionist bills in U.S. trade history. Both bills were initially designed to help the agriculture sector.11 The United States moved toward trade liberalization in 1934, when the RTAA was passed.12 Designed by free-trade advocates in the newly elected Roosevelt Administration, the RTAA gave the President the authority to reduce (or raise) tariffs up to half of Smoot–Hawley levels in exchange for reciprocal reductions on American exports with other nations. The bill, subject to renewal in three years, was passed by a Congress controlled by a Democratic majority. The RTAA was renewed 11 times over the next 30 years or so. This section analyzes the 1945, 1949, and 1955 renewals. The 1945 renewal was the last renewal before the creation of the GATT, which the United States initiated as a framework for multilateral trade negotiations. However, it was not technically an organization, and hence bypassed Congressional approval, to the dismay of many members of Congress (Kaplan, 1996). The 1949 renewal of RTAA came one year after the GATT talks in Geneva were concluded. Hence, votes on these two renewals allow us to compare changes in voting patterns in the pre- and post-GATT years. Furthermore, these two years were also the first instances in which the Republican Party was split on, instead of overwhelmingly against, free-trade legislation. The 1955 renewal was chosen to reflect voting patterns in the 1950s, when more exception clauses and peril points were introduced and attitudes on international trade within parties were shifting.13 The last bill that this section will analyze is the Trade Expansion Act of 1962. This legislation removed the item and country-specific approach in tariff reduction used in earlier years. The Kennedy administration successfully tied this bill to business interests, and the bill easily passed in a Democrat-controlled Congress. The Trade Expansion Act set the scene for modern U.S. trade policy. Overall, the seven bills studied were major trade bills that altogether represented a sequential roadmap of U.S. trade policy during those four decades. In the following probit voting analysis, the dependent variable is a binary (0, 1) variable, where 0 ¼ vote against free trade and 1 ¼ vote in favor of free trade. Voting in favor of free trade means a ‘‘no’’ vote on the 1922 and


215

1930 bills, and a ‘‘yes’’ vote on all the other bills.14 The independent variables are a binary party affiliation variable (0 ¼ Democrat and 1 ¼ Republican), and variables reflecting several state-level economic interests – cotton and silk manufacturing, wool, sugar, wheat, cotton, and an agriculture production variable. These variables are values of state-level per capita production or value added, or per capita land and farm value for the agriculture variable in some years.15 The use of state-level data relies on the assumption that Congressional representatives tended to vote based on statewide economic interests. However, later in this section, a supplementary analysis will show that the results are largely robust if one instead assumes House members voted based on district-wide economic interests. Data for each variable are calculated from the nearest years’ Census of Manufactures or Census of Agriculture.16 For each session of Congress for a particular vote, the regressions are conducted according to the following basic models for 1922 and 1930: V i ¼ a1 þ a2 Party þ u1

(1)

V i ¼ b1 þ b2 CottonManuf: þ b3 SilkManuf: þ u2

(2)

V i ¼ c1 þ c2 CottonManuf: þ c3 SilkManuf: þ c4 Wool þ c5 Cotton þ c6 Sugar þ c7 Wheat þ u3 V i ¼ d 1 þ d 2 CottonManuf: þ d 3 SilkManuf: þ d 4 Wool þ d 5 Cotton þ d 6 Sugar þ d 7 Wheat þ d 8 Agric þ u4 V i ¼ f 1 þ f 2 Party þ f 3 CottonManuf: þ f 4 SilkManuf: þ f 5 Wool þ f 6 Cotton þ f 7 Sugar þ f 8 Wheat þ f 9 Agric þ u5

(3)

(4)

(5)

The model is the same for the other years, except that variables for cotton manufacturing and silk manufacturing are combined into one variable for 1945 and 1949, and a broad variable for all textile products is used for the years 1955 and 1962. These changes are made due to the lack of detailed available data on cotton and silk manufacturing production as separate industries. The full results of this probit analysis are shown in Appendix C, and each equation denotes a column in each regression table.17 Note the number of observations in each regression does not exactly equal the number of members in the House or Senate, because some members did not cast a vote. Table 4 summarizes the direction and statistical significance of

216

LEI (SANDY) YE

Table 4.

Summary of Probit Estimation.

Vote

Variables Party

Cotton Manuf.

Silk Manuf.

Wool

Sugar

Cotton

House of Representatives 1922 Fordney-McCumber Act 1930 Smoot-Hawley Act 1934 RTAA 1945 RTAA Renewal 1949 RTAA Renewal 1955 RTAA Renewal 1962 Trade Expansion Act

þ þ

þ þ

þ þ þ þ þ þ

Senate 1922 Fordney-McCumber Act 1930 Smoot-Hawley Act 1934 RTAA 1945 RTAA Renewal 1949 RTAA Renewal 1955 RTAA Renewal 1962 Trade Expansion Act

þ þ þ

þ þ

þ Dropped þ þ

þ þ þ þ þ þ

þ þ þ

þ þ

Party: binary independent variable, where 0 ¼ Democrats and 1 ¼ Republicans; Vote: binary dependent variable, where 0 ¼ protectionist and 1 ¼ free trade. All production variables are on a per-capita basis. Note: Due to data limitations, variables for cotton manufacturing and silk manufacturing are combined into one variable for the 1945 and 1949 votes, and a broad textile variable was used for the years 1955 and 1962 (see Section 4). Significant at the 10 percent level. Significant at the 5 percent level. Significant at the 1 percent level.

the main coefficients of the regressions. In this summary table, the party variable is taken from the results of Eq. (1), and all other variables from the results of Eq. (3). 4.1. The Results One of the most prominent features of the results is the powerful effect of party affiliation in predicting a Congress member’s vote. The regressions show that votes were strongly partisan – Democrats more likely to vote in favor of free trade and Republicans in favor of protectionism. For House votes, the party variables in both Eqs. (1) and (5) are significant at least at


217

the 5 percent level for all the years studied. The same result is true for Senate votes except for 1922 and 1955, in which the significance, if any, is weaker. In addition, the pseudo-R2 values indicate that party affiliation had by far the strongest predictive power compared to any other variables. Since 1945, party’s predictive power was still high but weakened significantly. This trend is consistent with the Republican’s conversion to trade liberalization in the 1940s as argued by Irwin and Kroszner (1997).18 When all the economic interest variables are taken into account, party still adds significant explanatory power. This effect can be seen from the increase in pseudo-R2 from Eq. (4) to Eq. (5) for each vote. This increase averaged 0.42 for votes from 1922 to 1934, and 0.21 for 1945 to 1962. Because coefficients in a probit regression can only be interpreted for the significance of the direction of voting, not the magnitude, I construct a more informative estimate of the coefficients by calculating the predicted probabilities of voting in favor of free trade given a certain party (with all other variables held at the mean).19 As Table 5 shows, the Fordney–McCumber, Smoot–Hawley, as well as the RTAA votes were strictly partisan. The Senate showed breakdown in partisanship in 1945 and 1949, in which Republicans were much more likely to vote in favor of free trade. The same breakdown in the House was slower Table 5. Predicted Probability of Voting in Favor of Free Trade Given a Certain Party (with all other Variables Held at the Mean). Year

Democrat

Republican

House 1922 1930 1934 1945 1949 1955 1962

1.000 0.943 0.968 0.945 0.975 0.857 0.896

0.003 0.060 0.024 0.216 0.601 0.608 0.407

Senate 1922 1930 1934 1945 1949 1955 1962

1.000 1.000 0.942 0.942 1.000 0.823 1.000

0.000 0.000 0.100 0.657 0.885 0.973 0.119

218

LEI (SANDY) YE

and weaker, as it was not until 1949 that House Republicans’ likelihood to vote in favor of free trade took a major jump. There seems to have been a retreat to stronger partisanship in 1962 as the predicted probability for Republicans to vote for free trade decreased significantly. Despite the breakdown in partisanship, party remained the most reliable single variable predicting a Congress member’s vote. Beyond party affiliation, economic interests also played important roles. Silk manufacturing’s coefficients for 1922, 1930, and 1934 were negative, indicating states with heavier production of silk manufactures were more likely to vote in favor of protectionism. Excluding the party variables, these coefficients are significant at the 1 percent level in the House votes on all three years. Once party is added in Eq. (5), the coefficients became less significant or insignificant. One explanation for this is that in these three years, the Republican Party was the majority party in the top silk manufacturing states, which were heavily concentrated in the Northeastern states of Pennsylvania, New York, and New Jersey. Hence, once taking into account this Republican Party effect, the coefficients are not as significant. The same coefficients for the Senate votes are weaker in significance, which is not surprising, as a Senator is representing a much broader set of interests. For the same three years, coefficients on cotton manufacturing showed little, if any, significance. In fact, their signs were often slightly positive. This can be explained by the shift of cotton manufacturing toward the South since the 1920s.20 The cotton-manufacturing states were more heavily Democratic and hence were more likely to vote in favor of free trade. These regressions suggest that their response to the export interests of farmers (as well as party affiliation) outweighed their response to manufacturing interests. While data limitations do not allow us to see strictly comparable coefficients on cotton and silk manufacturing for the years since 1945, we can still gain insights by examining the combined variable of cotton, rayon, and silk manufacturing or textile products in general. For the 1945 and 1949 renewals, cotton and rayon manufactures (which included silk production) had mostly positive coefficients. They were significant in the House votes for Eq. (2). Because cotton manufacturing is likely weighted heavily in this variable, the coefficients’ trend resembled closely those on cotton manufacturing in the previous years. But turning to 1955 and 1962, textile groups showed significant resistance to free trade, as the coefficients in (5) were negative and significant in the 1955 House votes and 1962 House and Senate votes. Even though this period is well into the phase of trade liberalization in the United States, a resurgence of resistance to free trade by textile groups marked an emerging counter-trend at that time.


219

This new import sensitivity may very well have contributed to the higher trade barrier with the Pacific Rim. Even though these free-trade bills were passed, many exception clauses and special provisions were added, adversely affecting imports from Pacific Rim. Furthermore, the Pacific Rim, in particular textile-producing countries like Japan, was often on the minds of members of Congress. Representatives from textile-producing states in New England and the South argued for non-tariff barriers to protect the domestic textile industry. For example, the first voluntary export restraint (VER) against Japanese textiles was negotiated in 1955 and this authority was expanded later (Sherman, 2002). The regression results are consistent with this new orientation in political interest. In addition to textile products, we also include several major agricultural products in the analysis. Wool is included because it was a major import from the Pacific Rim, especially Australia and New Zealand. To give a comparative perspective between different agricultural interests, cotton, sugar, wheat, and a general agriculture variable are also included. For wool, almost all coefficients for all votes were negative. The coefficients were significant, however, in the beginning and end of the time period. The strong push for protectionism from the wool-producing states in 1922 was due to the Fordney–McCumber Act’s intent of providing relief for an agricultural depression for which wool producers were a major interest (Taussig, 1931). The support of wool states for protectionism was statistically insignificant for the subsequent years until 1955, when it seemed to resurface as an interest in favor of protectionism. This phenomenon was similar to the case of the textile interests. The wool states’ support for protectionism during a late stage of liberalization implied greater harm for Pacific Rim imports. As early as GATT’s Geneva negotiations in 1947, wool was a major issue. Congress attempted to impose trade restrictions on wool in order to support domestic prices, and import fees as well as quotas were passed. Among the other variables, cotton-producing states show mostly positive and significant (when not including party) coefficients. Mainly produced in Democratic states in the South, cotton has always been a significant export item. Thus, cotton states are expected to vote in favor of free trade.21 Meanwhile, sugar shows significant coefficients toward protectionism for the years 1922, 1930, and 1934, but not afterwards.22 This, however, could be due to the increasing importance of the sugar quota system after 1934.23 The wheat variables were also only weakly significant, if at all. Lastly, most of the coefficients for agriculture are close to 0 and weak in significance, suggesting a relatively neutral attitude from agriculture as a whole. The push for higher protectionism in agriculture was sector- and

220

LEI (SANDY) YE

region-specific. Those who produced export-oriented crops did not need protection, as it could do them more harm than good.24 Overall, the explanatory power that all the per capita production variables could add to the party effect is sizable. For all 14 regressions, these economic interest variables increased the pseudo-R2 from column (1) to column (5) by 15 percentage points on average. To a lesser extent than party affiliation, votes on all the trade legislations have clearly been responsive to economic interests. However, the extent of the ability of economic interests to lobby for protection varied. In 1955 and 1962, when neither sugar, wheat, and cotton nor agriculture production was strongly associated with a vote in favor of protectionism, wool and textiles remained significant economic interests in that direction. Constant adjustment and realignment of economic interests were taking place in Congress, and there was considerable movement for negotiations and trade-offs in supporting freer trade. However, this movement happened in a way that disadvantaged the Pacific Rim, as domestic economic interests that had stronger ability to maintain protection were often competitive with Pacific Rim imports. The results of the probit analysis can be linked to existing theoretical and empirical works on the effect of domestic political lobbying on the level of protection. For example, an extension by Grossman and Helpman (1994) of their political economy framework for the setting of tariff rates suggests that finished-good manufactures, such as textiles, tend to be more successful in securing protectionist policies compared to intermediate-good sectors. Furthermore, their model predicts that sectors that are more organized would receive higher protection. In fact, the textile sector constituted one of the most active and strong coalitions in this period pushing for higher trade barriers. For example, when the textile industry supported stricter VERs on Japanese cotton textile in 1956, Congress appended an additional section to the Agricultural Act of 1956 to allow for bilateral export restraint negotiations (Destler, 2005, p. 25). Beyond the level of protection across industries, the persistence of high protection over time within the same industries suggested by the probit results may also be explained by extensions of the Grossman–Helpman model (see Brainard & Verdier, 1997). Empirically, the results of the voting probits are also consistent with previous works that found higher levels of protection in labor-intensive industries, industries with high import penetration, consumer goods industries, and industries that are clustered regionally (see Rodrik, 1995, p. 1481 for a comprehensive summary of the relevant literature for each of these points). U.S. sectors that are import-competitive with the Pacific Rim matched these characteristics well. As mentioned before, among textiles,


221

goods that were import-competitive with the Pacific Rim were often the most labor-intensive ones (e.g., silk laces and embroideries). Calculations based on Irwin’s 1930 sector-level data in 1998 (Table 4) suggest that the textile sector had comparatively high import penetration. The regional concentration of the textile industry is also well documented, though there were some shifts in the regional presence of certain textile products.

4.2. A Supplementary Robustness Check The probit models in the above analysis utilized state-level data for its independent variables (excluding party). However, some would argue that House members would have voted based on district-wide, rather than statewide, economic interests. In order to address this issue, this subsection demonstrates a similar set of probit regressions based on district-level employment data across industries. For this supplementary framework, I use Fishback’s data (2009) on sectoral employment across counties in the 1930 Census (the same data are not available for 1940–1960). The same dataset also matches the county data to Congressional districts in the 72nd, 73rd, 74th, and 75th Congress. The regression framework is essentially the same as above, with some adjustments in the independent variables due to data availability. Specifically, the probit regressions were conducted on the same set of bills according to the following framework: V i ¼ a1 þ a2 Party þ u1

(1)

V i ¼ b1 þ b2 totalCotMill: þ b3 totalSilMill þ u2

(6)

V i ¼ c1 þ c2 totalCotMill: þ c3 totalSilMill þ c4 totalWoolWors þ u3 V i ¼ d 1 þ d 2 totalCotMill: þ d 3 totalSilMill þ d 4 totalWoolWors þ d 5 totalAgr þ u4 V i ¼ f 1 þ f 2 Party þ f 3 totalCotMill: þ f 4 totalSilMill þ f 5 totalWoolWors þ f 6 totalAgr þ u5

(7) (8)

(9)

The variables totalCotMill., totalSilMill, and totalWoolWors represent district-level employment in cotton, silk, and woolen and worsted mills, respectively. The variable totalAgr represents the district-level employment

222

LEI (SANDY) YE

of farm owner and tenants, managers and foreman, and laborers altogether. The results of this supplementary framework, shown in Appendix D for the House votes, showed similar trends compared to the previous analysis. With the exception of 1962, silk mills have been persistently protectionist, with significance in the Fordney–McCumber Act, Smoot–Hawley Act, RTAA, and 1955 RTAA renewal. Woolen and worsted was highly significant in favor of protectionism before 1955. Cotton mills and agriculture also had similar patterns as the previous analysis. The results of this supplementary analysis suggest that using state-level measures in the main probit framework sufficiently captures the relative weighting of economic interests for each vote in the House of Representatives. In addition, this exercise provides confidence that the state-level data are reasonably good proxies for the years after 1930. However, beyond Congress’s relative weighting of domestic economic interests, could there have been a more international and diplomatic aspect that also contributed to the higher barriers to trade with the Pacific Rim? To extend this question, in the next section we examine historical and diplomatic evidence on U.S. trade policy making, now incorporating the important role played by the executive branch.

5. A HISTORICAL INTERPRETATION OF U.S.–PACIFIC RIM TRADE BARRIERS The transformation of U.S. trade policy from 1922 to 1962 was based on broad principles and goals in U.S. foreign policy and was intimately tied to the influential role that the executive branch played after 1934. To initiate trade liberalization, the executive branch had to resort to an incremental approach. As a result, it had to carefully assess trade relations with different regions. In this process, the United States pursued a policy that prioritized freer trade with certain regions over others, impeding trade liberalization with the Pacific Rim. In this section, I examine the historical circumstances that contributed to this trend by focusing on three themes: the reciprocal trade agreements, the GATT framework, and the Pacific Rim’s historical development. 5.1. The Use of Reciprocal Trade Agreements Shortly after the Smoot–Hawley Act and the subsequent downward spiral of world trade, the newly elected Roosevelt administration decided to


223

pursue trade liberalization based on bilateral reciprocity. In 1934, the RTAA shifted trade policy making from Congress to the executive branch, and the President was given power to reduce tariffs by as much as 50 percent of Smoot–Hawley levels. The administration pursued reciprocal trade agreements for two reasons. First, a unilateral tariff reduction would have faced strong resistance from Congress, as protectionist sentiments remained strong. Second, reciprocal trade agreements closely tied export-oriented coalitions into the policy making process and increased the political cost of lobbying for import-competing groups (Brenner, 1977). However, the United States had to select which nations to negotiate trade agreements with, and in this process the Pacific Rim was largely neglected. In fact, among all the nations that signed trade agreements with the United States between 1934 and 1947, none were Pacific Rim nations (see Table 6). Table 6.

Countries that Signed Reciprocal Trade Agreements with the United States, 1934–1947.

Argentina Belgium and Luxembourg Brazil Canada Colombia Costa Rica Cuba Czechoslovakia Ecuador El Salvador Iceland Finland France Guatemala Haiti Honduras Iran Mexico Netherlands Nicaragua Peru Sweden Switzerland Turkey United Kingdom Uruguay Venezuela

Source: Lusztig (2004).

224

LEI (SANDY) YE

The United States was much more interested in opening up trade with nations in Europe or the western hemisphere, as the bulk of U.S. trade was with these two regions. Even more importantly, the main priority of the United States in negotiating agreements was the opening of its export market, and the Pacific Rim was not seen as an important partner for this purpose. Through reciprocal agreements, import liberalization of certain products could only be achieved by simultaneously tying them to export expansion of similar scope for U.S. products. The export-oriented coalition that advocated trade liberalization in the United States was led by large industry groups (Lusztig, 2004), which often represented heavy-industry and technology-based goods such as automobiles. U.S. export of these goods to the Pacific Rim was relatively low, since many Pacific Rim economies lacked the purchasing power of many more developed economies of Europe at the time. The State Department set up special agencies and committees to make decisions on negotiation partners and strategy. For example, in 1934 Assistant Secretary of State Francis Sayre submitted to Congress a list of 29 countries that were chief suppliers of certain products, sometimes known as the ‘‘Sayre’s List’’ (Tasca, 1938, p. 137). Only three economies, China, Japan, and Australia, were in the Pacific Rim. The United States did not end up negotiating agreements with any of these three economies. In addition to the selection of countries, the State Department also had to craft its negotiations by choosing an appropriate range of products, a process that also placed the Pacific Rim at a disadvantage. In order to maintain its newly delegated powers in trade negotiations, the executive branch attempted to minimize the domestic political costs of reciprocal trade agreements by choosing tariff reduction in products that avoided stiff domestic resistance. In fact, evidence presented by Tasca (1938) shows that duty reductions were most substantial within five schedules – earths, earthenware, and glassware; agricultural products; metals and manufactures; chemicals, oils, and paint; and sundries. The Pacific Rim exported little of these goods. Of the 447 reductions achieved by the negotiated trade agreements, a total of 356 were in these five schedules. In contrast, among the five textile schedules in which the Pacific Rim did have substantial specialization, only 56 reductions were negotiated. The executive branch tried to mollify protectionist interests such as the textile industry and prevent them from influencing the general direction of trade liberalization. Some might point out that these discriminatory effects were mitigated by the unconditional MFN principle of the reciprocal agreements, under which trade reductions negotiated with one country would apply to all other


225

countries. Thus, the Pacific Rim could have benefited as external participants. However, evidence suggests that the extent of the benefit was limited. First, the executive branch negotiated agreements by adhering to the ‘‘chief-supplier principle’’: a reciprocal trade agreement on a certain product was negotiated with the country that was the chief supplier of that product (Lusztig, 2004). Thus, the external benefits to third parties were minimized. Second, a reclassification of products occurred after the RTAA was passed. As a result, products defined in the Smoot–Hawley Act were classified into finer categories, providing an additional channel to minimize third-party benefits. Take an example given by Tasca (1938, p. 143): in U.S. negotiations with Belgium, tariff reductions were included on ‘‘woven green billiard fabrics.’’ In this case, even though the reduction was in a textile product, the East Asian economies could not expect to reap much benefit because the defined product was very specific. In fact, Brenner (1977) noted that the reclassification helped U.S. negotiate textile reduction with Great Britain without affecting the more competitive textiles from Japan. Lastly, trade agreements often contained escape clauses that were specifically designed to prevent benefits to competitive third-party countries and ease domestic industries’ concern about flooding of cheap imports. For example, in a notable case, the then governor of Puerto Rico voiced a strong complaint regarding Japanese and Chinese handkerchiefs imports, which later caused the Committee on Trade Agreements (the agency partly in charge of administering the escape clauses) to grant an escape clause in 1939. President Roosevelt reportedly remarked that he did not wish to grant tariff concessions that would ‘‘seriously hurt our difficult task in Puerto Rico, the Virgin Islands, and Hawaii’’ (Eckes, 1995, p. 222). To be sure, the implementation of the reciprocal trade agreements was instrumental in moving the United States toward trade liberalization. Bilateral reciprocity was perhaps the most politically feasible approach at the time. However, the flexibility of reciprocal agreements also induced geographic bias. Negotiating parties’ interests, rather than broad principles of trade liberalization, prevailed in determining the trajectory of tariff reductions. The tendency for bilateral agreements to take such a course has historical precedent that dates back at least to the nineteenth century. In a study of the bilateral agreements during 1860–1875 known as the Cobden– Chevalier Network, Lampe (2009) shows that tariff cuts were highly uneven across commodities. Furthermore, similar to the findings of this paper, the presence of MFN contributed little to mitigating this effect. The reciprocal agreements after 1934 produced the same type of effect as Lampe found,

226

LEI (SANDY) YE

and by extension, the Pacific Rim benefited less from tariff reduction because of the nature of its exports. Within the international context, it was natural that the United States targeted its negotiations with European and South American countries. The former represented a long-standing partnership as well as a major market for the expansion of U.S. exports of heavy-industry goods. The latter are nearby neighbors in the western hemisphere, and maintaining stability in the region with sound trade relations was equally important. In the run-up to WWII (and afterwards), reciprocal trade agreements were a major tool to achieve foreign policy objectives in these regions, and President Roosevelt deemed reciprocal trade as ‘‘an indispensable part of the foundation of any stable and durable peace’’ (Eckes, 1995). The Pacific Rim, far and beyond, was given relatively little attention. The United States saw little opportunity in export expansion there, and it was constrained in freely negotiating import-sensitive goods that could have benefited the Pacific Rim. Indeed, evidence by Tasca (1938) suggests that trade barrier reductions were higher for agreement countries than non-agreement countries, consistent with the empirical estimates in this paper. Because the Pacific Rim also benefited little from external effects of the unconditional MFN clause, the RTAA affected the Pacific Rim very little, despite its large success in helping the United States jump toward a course of trade liberalization.

5.2. Multilateral Agreements under the GATT In 1947, the United States moved toward a new stage in trade liberalization. For the first time, the GATT established a multilateral trade negotiation framework. It further reduced the influence of import-competing domestic interests and simultaneously expanded the participation of exporters. According to Barton et al. (2006), the GATT was also created to help solve a credible commitment problem – small countries were reluctant to negotiate with the United States after WWII, because if the United States failed to commit to the agreements, the small nation would face large potential losses in the form of non-salvageable investments in export industries. With a multilateral framework, large nations have smaller incentives to renege, because agreements are linked for multiple countries and a deviation faced greater economic losses. If a ‘‘small’’ nation is taken as a nation with low economic bargaining power, then this mechanism should have helped the Pacific Rim achieve deepened trade liberalization. But evidence suggests the contrary, at least from 1947 to 1962.


227

Why did the incorporation of a multilateral framework once again neglect the Pacific Rim? The answer is reflected in the creation and design of the GATT. The GATT was not intended at its origin to include developing nations. The United States wanted a small forum of nations to discuss the reduction of trade barriers. In fact, the GATT extended many principles of the previous bilateral agreements, such as unconditional MFN clause. Among the original 23 members of the GATT, only China was from the Pacific Rim. Essentially, the exclusion of the Pacific Rim reflected many of the same problems as under the bilateral agreements. For the United States, incorporating a developing nation in the Pacific Rim to the GATT yielded little benefit in the form of export expansion. Bargaining power once again played an important role. Reducing trade barriers with large nations with greater bargaining power was more urgent than with small nations with little bargaining power. Often, the latter group consisted of nations that were very dependent on foreign trade. In the Pacific Rim, only China could be characterized as a large nation that had a large internal trade and could be self-sufficient without trade. Many other nations, such as Singapore, Indonesia, Thailand, etc., were all small nations with which the early GATT was not concerned. They could not effectively threaten the large nations with closing off their exports, as that would be detrimental to their own economic growth. Instead, in creating the GATT in the post-war era, the United States was more concerned about recovery of major European economies. The structure and rules of the GATT also discouraged participation among Pacific Rim economies. As the primary initiators of the GATT, the United Kingdom and especially the United States held dominant positions. In 1948, the United States alone accounted for 65 percent of GDP of all GATT members and never dropped below 50 percent before 1962 (Barton et al., 2006, pp. 11–13). In fact, until the Doha rounds of this century, every round of trade negotiation was initiated by the Congressional delegation of negotiation power to the President (Barton et al., 2006, p. 44). The United States preferred a multilateral framework that was small in membership and relatively flexible in structure, perhaps believing it was more conducive to achieving deeper trade liberalization. Originally, this framework was intended as a stepping stone to the International Trade Organization (ITO), which would have created a bigger and more centralized international organization, resembling the International Monetary Fund, for example. But this organization never became a reality, partly because the Truman administration did not believe Congress would ratify it and hence did not even submit it for a vote.

228

LEI (SANDY) YE

The GATT continued to operate in a decentralized way, and this dampened the influence that the Pacific Rim could have had in international trade negotiation. For example, if the ITO were implemented, the Pacific Rim would have more representation and total voting power. Evidence from Barton et al. (2006, p. 36) shows that the United States tried to mitigate this potential effect by proposing to allocate votes in the ITO by share of world trade as well as to create a permanent committee of a few nations. Furthermore, while the unconditional MFN principle continued to apply in the GATT, they were often subject to exceptions. For example, existing members of the GATT can choose not to apply MFN to newly admitted members (Barton et al., 2006). In fact, when Japan became a GATT member in 1955, more than 40 percent of existing members refused to grant Japan MFN status in order to protect their domestic industries from Japanese textiles and other labor-intensive imports (Gowa & Kim, 2005). Thus, even though the institutional mechanisms of the GATT were effective in helping the United States reduce trade barriers with its main trading partners, they were not the best design from the perspective of the Pacific Rim. But this outcome might have been avoided if the Pacific Rim had consolidated bargaining power as a regional bloc.

5.3. The Lack of Regional Cooperation As the above discussion has shown, asymmetry in economic bargaining power disadvantaged the Pacific Rim and contributed to the region’s higher trade barriers with the United States. However, this might have changed if the Pacific Rim had worked together as an economic bloc and formulated a coordinated foreign trade policy direction. For example, the Pacific Rim could have attempted collective negotiation with the United States. Such unified regional efforts would more closely coordinate common interests, and thereby increase the bargaining strength of the Pacific Rim. Was such a possibility feasible, and if so, why did it never come to reality? The Pacific Rim possessed some ideal characteristics to forge the kinds of economic cooperation that Europe had already achieved at the time.25 Many countries were main suppliers of raw materials: raw silk in China and Japan, wool in Australia, and rubber and tin in Southeast Asia, to name a few. The region was also relatively less developed in advanced manufacturing industries, partly contributed by high trade barriers. Furthermore, with the exception of China, most other nations in the region were relatively small or sparse in population and were hence particularly dependent on


229

foreign trade. All these reasons suggest a real possibility for these nations to recognize their common interests and create institutions to promote these interests. But historical circumstances during this period prevented such cooperation. In order to formulate long-lasting economic cooperation, a region needs to show some degree of political unity. But the decades leading up to WWII were filled with regional conflicts. Perhaps the most notable was the one between China and Japan during WWII, when Japan extended its economic interests and later military invasion into China and other regions in East Asia starting in the 1920s. Japan’s desire for raw materials to fuel its industrialization resulted in numerous trade disputes and severely disrupted the region’s trade with the United States. Yet China and Japan would be the two biggest potential candidates to take a leading role in forming a coordinated effort in negotiating lower trade barriers with the United States. Regional conflicts excluded this possibility. In addition, there were no natural alliances between nations that could help initiate or expand cooperation. Australia and New Zealand could have been important players, but culturally they did not share a common identity with East Asia, and economically they were members of the British Commonwealth. Hence, they had no particular interest in East Asian trade policy. In addition to minimal regional conflicts, regional cooperation also requires that the participating nations have stable internal governments. The Pacific Rim nations often did not meet this requirement. China had only recently become a republic in 1912, ending a 5,000-year history of imperial rule. The newly formed republic was filled with internal conflicts, as warlords often claimed large powers in different regions of the country. After the war with the Japanese, a civil war ensued until the Communists established the People’s Republic of China in 1949. The civil war severely restrained China’s foreign trade, and the new People’s Republic of China virtually ceased trading with the United States until the economic reforms of the 1970s. In Southeast Asia, domestic conflicts were also widespread. In the early 1920s, many parts of this area were still domains of European colonial powers.26 After WWII, many of these areas became independent, but domestic political conflicts hardly stopped. Japan, of course, also underwent a change in government after its defeat in WWII. Not surprisingly, trade flows in the Pacific Rim plummeted (Frankel & Kahler, 1993). Domestic political instability destabilized the Pacific Rim’s conduct of foreign trade policy. Nations could not even adopt consistent foreign trade policy for themselves, much less forge a broad cooperation among each other.

230

LEI (SANDY) YE

Coincidentally, the most transformative period in American foreign trade policy was also one in which the Pacific Rim faced immense changes. What would have been an opportune time for the Pacific Rim to engage in trade negotiations with the United States and pursue bilateral or multilateral tariff reductions instead was occupied by wars and conflicts. This resulted in a weak bargaining position for the region, which U.S. foreign policy recognized. When considering these broad international trends, the consistency of a higher trade barrier with the Pacific Rim was no surprise after all. The Pacific Rim lacked the features that would have been effective for trade liberalization with the United States, and while its trade barriers with the United States did fall during this period, the pace of decline was much slower than for the rest of the world.27

6. CONCLUSION This paper addresses the central question ‘‘were U.S. trade barriers with the Pacific Rim higher than those with the rest of the world from 1922 to 1962?’’ Estimates of average tariff rates of Pacific Rim imports have shown that this was consistently the case among import-sensitive high-stage manufactured goods. Congressional voting analysis suggests that domestic interests were quite successful in maintaining a certain degree of protection for goods that competed significantly with the Pacific Rim despite the movement to freer trade after 1934. Externally, the incremental nature of U.S. trade liberalization strategies since 1934 plus U.S. foreign policy goals put the Pacific Rim on the low end of the priority list. Concurrently, the Pacific Rim could not strengthen its bargaining power through forms of economic cooperation due to regional conflicts and instability. Several broad observations can be drawn from the results of this paper. First, the effect of trade policy is not one-dimensional. Even with a rulesbased policy rigorously applying the principles of reciprocity and nondiscrimination, which the United States largely achieved by 1962, trade barriers to different regions can vary greatly. Second, political feasibility has been and will remain a top priority in trade liberalization. It is uncertain whether the marginalization of the Pacific Rim was a price to be paid for achieving broader trade liberalization on the part of the United States. Lastly, bargaining power is crucial to a successful trading relationship. U.S. policies in this analysis were not formulated by treating the Pacific Rim as


231

exogenous. Had Pacific Rim economic and political development been different, U.S. trade policy would have adjusted in response. Today, the trading relationship between the two regions differs greatly from that of 1922–1962. Economies like China and Japan are among the largest trading partners of the United States, and their rapid economic growth has diminished many factors that previously restrained freer trade between the two regions. U.S. trade policy is now aligned more closely with global trading frameworks like the World Trade Organization, and the economic bargaining power of Pacific Rim economies has substantially increased. However, LDCs under the multilateral trading framework today face many of the same challenges as those of the Pacific Rim in an earlier period. The highest U.S. tariffs still tend to be on products, such as laborintensive manufactures, in which LDCs have a comparative advantage.28 To overcome fully the implicit geographic bias in an ostensibly neutral trade policy, the United States would have to reverse patterns of political behavior that go well back in American history.

NOTES 1. As early as 1792, Alexander Hamilton made a case for protection in his famous Report on Manufactures. 2. See Taussig (1931) on the history of tariffs of the late nineteenth and early twentieth centuries; Callahan, McDonald, and O’Brien (1994), Eichengreen (1989), and Irwin and Kroszner (1996) for analyses of the Smoot–Hawley Act; Bailey et al. (1997), Lusztig (2004), and Tasca (1938) on the RTAA; and Bagwell and Staiger (1999) and Barton, Goldstein, Josling, and Steinberg (2006) on the GATT/WTO. Other comprehensive studies include Goldstein (1993), Kaplan and Ryley (1994), and Kaplan (1996). 3. The Philippines is, however, largely excluded from the analysis of this paper because of its special trading relationship with the United States during the period considered. 4. Please see Appendix A for a detailed list of goods represented in each category. For the other years, every effort has been made to maintain the same set of goods in each category for consistency. Some discrepancies arise due to slight changes in classification and grouping throughout the different years. However, the tariff rates within a year should not be affected. As for tariff rates across years, the broad trends should not be largely affected by the discrepancies either. 5. When computing the total imports and estimated duties from the Pacific Rim, the breakdown was not by goods, but instead the nine commodity groups as defined in FCNUS (e.g., Table XVI of the 1931 edition). 6. Pacific Rim countries/regions (as named in those years) for which data were collected were China (including Kwantung, which was recorded separately in the source), Japan, Australia, New Zealand, Hong Kong, British Malaya, French

232

LEI (SANDY) YE

Indochina, Netherlands Indies, Thailand, and Korea. Names of countries varied slightly across the years, but the regions covered are the same. 7. Because many of the items considered are actually a broad set of items (e.g., cotton manufactures), many different specific duties may apply. Hence, the percentage of imports of a good from the Pacific Rim is not always equivalent to the percentage of duties collected from that region. However, this assumption is still quite reasonable, because tariff rates on particular commodities were generally the same for all countries of origin. Because data on duties collected broken down by country are simply not available, making this assumption is the only practical option. 8. Similar data for 1922–1938 and 1958 were not available. 9. See Hayford and Pasurka (1991), for example. 10. Even though trade agreements existed before 1934, they played a relatively minor role, and were subject to two-third approval from Congress. 11. The Fordney–McCumber Act was initiated when farmers in the western United States faced severe price declines in their products after World War I. The Smoot–Hawley Act originated from a special session called by President Hoover to aid the farmers, as Midwestern and other regions’ farming interests had been advocating higher duties since Hoover’s election in 1928. In the case of the Smoot– Hawley Act, manufacturing duties were also raised dramatically. President Hoover did not intend to significantly alter the duties on manufactures, calling for ‘‘limited revision’’ when necessary, but the domestic political process in Congress led to a different outcome. 12. Technically, the RTAA was an amendment to the Smoot–Hawley Act. 13. For example, the Republicans now had a group of internationalists in favor of free trade, and the Southern Democrats had begun to turn away from free trade. 14. All Congressional voting records and members’ party affiliations are obtained from data available at the Inter-university Consortium for Political and Social Research (ICPSR, 2004). 15. Data on value-added are used for production of textiles in the 1955 and 1962 votes. Data on total value of farms (land and buildings) are used for votes starting in 1934. 16. See Appendix B for more details on the data source used. 17. Minor deviations to this model occurred in a few instances for reasons that will be explained later. 18. In fact, they have conducted the same regression on party alone for Senate votes on the 1934 RTAA Act and 1945 renewal, and the pseudo-R2 they have estimated are almost identical to my estimates. Irwin and Kroszner did not study votes in the House, but my estimates show that party effects on House votes are even stronger in both cases. 19. To do so, I utilized a Statar command, prtab, written by Long (package spost9_ado from http://www.indiana.edu/Bjslsoc/stata, 2007). prtab calculates the predicted probability of a variable as defined in Long and Freese (2001, p. 120), which for probit models is P^r(y ¼ 1|x) ¼ F(xb^), where F is the c.d.f. of the normal distribution where variance equals 1. 20. In fact, Massachusetts was the only Northeastern state that was still on the top-five list of cotton-manufacturing production for these years. 21. One exception to this is the high significance of cotton in the direction of protectionism for the 1962 House vote. One possible reason is its colinearity with the


233

party variable. In fact, column (4) shows that by removing the party variable and including all other independent variables, the cotton coefficient becomes positive. In any case, this outlier would not detract from cotton states’ general trend. 22. This variable was dropped in the 1930 Senate vote due to colinearity problems. 23. The sugar quota was revised downward in 1948, 1951, and 1956 (Dye & Sicotte, 2004), suggesting that support for protectionism shifted away from tariffs to quotas in this sector. 24. For example, Eichengreen (1989) cites inland agriculture as a coalition against the Smoot–Hawley tariff, and Callahan et al. (1994) show that even border agricultural interests were more likely to vote against the bill once one controls for party. In addition, evidence from Kaplan (1996) has shown that some farming groups resisted the Smoot–Hawley tariff because the higher tariffs on manufacturing would impose higher prices on agricultural implements and inputs. 25. The Western European nations were already quite active in this regard in the post-WWII era. Various forms of economic unions, such as the Benelux Union between Belgium, Luxembourg, and later the Netherlands, were pursued as early as the 1920s (Barton et al., 2006). These efforts eventually fostered the European Economic Community (ECC) and later the European Union (EU). The evolution of the EU was closely related to GATT/WTO. For example, the United States advocated European integration in 1957 by using the customs union exception in Article XXVIIII of the GATT (see Barton et al., 2006, p. 35). 26. These included British Malaya, Netherlands Indies, French Indo-China, etc. 27. The extent to which high trade barriers with the United States negatively impacted the economic development of the Pacific Rim is unfortunately beyond the scope of this paper. The answer to that question would require a much broader study, looking at factors such as the region’s terms of access to markets in other parts of the world. Although the regional discrimination implicit in U.S. policies was to some extent a function of Asia’s economic backwardness at that time, what we can say is that U.S. policies were not helpful in improving the situation. 28. For example, in 2006, the average tariff rate of Bangladeshi and Cambodian exports to the United States is about eight times that of United Kingdom and five times that of France (Elliott, 2009).

ACKNOWLEDGMENTS I thank Gavin Wright for his invaluable advice and insightful suggestions. Comments from Robert Staiger and two anonymous referees as well as data support from Price Fishback are also greatly appreciated.

REFERENCES Bagwell, K., & Staiger, R. (1999). An economic theory of GATT. The American Economic Review, 89(1), 215–248.

234

LEI (SANDY) YE

Bailey, M., Goldstein, J., & Weingast, B. (1997). The institutional roots of American trade policy: Politics, coalitions, and international trade. World Politics, 49(3), 309–338. Barton, J., Goldstein, J., Josling, T., & Steinberg, R. (2006). The evolution of the trade regime: Politics, law, and economics of the GATT and the WTO. Princeton, NJ: Princeton University Press. Brainard, S. L., & Verdier, T. (1997). The political economy of declining industries: Senescent industry collapse revisited. Journal of International Economics, 42, 221–238. Brenner, S. R. (1977). Economic interests and the Trade Agreement Program, 1937–1940: A study of institutions and political influence. Ph.D. thesis, Stanford University. Callahan, C., McDonald, J., & O’Brien, A. (1994). Who voted for Smoot–Hawley? Journal of Economic History, 54(4), 683–690. Carter, S., Gartner, S., Haines, M., Olmstead, A., Sutch, R., & Wright, G. (Eds). (2006). Historical statistics of the United States: Earliest times to the present. New York, NY: Cambridge University Press. Destler, I. M. (2005). American trade politics (4th ed.). Washington, DC: Institute for International Economics. Dye, A., & Sicotte, R. (2004). The U.S. sugar program and the Cuban revolution. Journal of Economic History, 64(3), 673–704. Eckes, A., Jr. (1995). Opening America’s market: U.S. foreign trade policy since 1776. Chapel Hill, NC: University of North Carolina Press. Eichengreen, B. (1989). The political economy of the Smoot–Hawley Tariff. In: Research in economic history (Vol. 12). Greenwich, CT: JAI Press. Elliott, K. (2009). A US Trade Policy for Development: Helping the poorest in a time of crisis. A VoxEu.org publication. Available at http://www.voxeu.org/index.php?q ¼ node/3606 Fishback, P. (2009). Datasets on employment across industries in the 1930 Census matched to congressional districts. (E-mail to author). Frankel, J., & Kahler, M. (Eds). (1993). Regionalism and rivalry: Japan and the United States in Pacific Asia. Chicago, IL: The University of Chicago Press. Goldstein, J. (1993). Ideas, interests, and American trade policy. Ithaca, NY: Cornell University Press. Gowa, J., & Kim, S. Y. (2005). An exclusive country club. The effects of the GATT on trade, 1950–94. World Politics, 57, 453–478. Grossman, G., & Helpman, E. (1994). Protection for sale. American Economic Review, 84, 833–850. Hawke, G. R. (1975). The United States tariff and industrial protection in the late nineteenth century. The Economic History Review, 28(1), 84–99. Hayford, M., & Pasurka, C. (1991). Effective rates of protection and the Fordney–McCumber and Smoot–Hawley Tariff acts. Applied Economics, 23, 1385–1392. Hayford, M., & Pasurka, C. (1992). The political economy of the Fordney–McCumber and Smoot–Hawley Acts. Explorations in Economic History, 29, 30–50. Inter-university Consortium for Political and Social Research and Congressional Quarterly, Inc. (2004). United States Congressional roll call voting records, 1789–1990. Parts 1-202. Computer file. 2nd ICPSR version. Ann Arbor, MI: Inter-university Consortium for Political and Social Research [producer and distributor]. Irwin, D. (1998). The Smoot–Hawley tariff: A quantitative assessment. The Review of Economics and Statistics, 80(2), 326–334.


235

Irwin, D., & Kroszner, R. (1996). Log-rolling and economic interests in the passage of the Smoot–Hawley tariff. Carnegie-Rochester Series on Public Policy, 45, 173–200. Irwin, D., & Kroszner, R. (1997). Interests, institutions, and ideology in the Republican conversion to trade liberalization, 1934–1945. NBER Working Paper No. W6112. Kaplan, E. (1996). American trade policy, 1923–1995. Westport, CT: Greenwood Press. Kaplan, E., & Ryley, T. (1994). Prelude to trade wars: American tariff policy, 1890–1922. Westport, CT: Greenwood Press. Lampe, M. (2009). Effects of bilateralism and the MFN clause on international trade: Evidence for the Cobden–Chevalier network, 1860–1875. Journal of Economic History, 69(4), 1012–1040. Long, S., & Freese, J. (2001). Regression models for categorical dependent variables using Stata. College Station, TX: Stata Press. Lusztig, M. (2004). The limits of protectionism: Building coalitions for free trade. Pittsburgh, PA: University of Pittsburgh Press. Rodrik, D. (1995). Political economy of trade policy. In: G. Grossman & K. Rogoff (Eds), Handbook of international economics (Vol. 3, pp. 1457–1494). Amsterdam: NorthHolland. Sherman, R. (2002). Delegation, ratification, and U.S. trade policy: Why divided government causes lower tariffs. Comparative Political Studies, 35(10), 1171–1197. Stata 9 Commands for Post-Estimation Interpretation of Regression Models (spost9_ado). (2007). Available at http://www.indiana.edu/Bjslsoc/stata. Retrieved on April 28, 2007. Tasca, H. (1938). The reciprocal trade policy of the United States. New York: Russell & Russell. Taussig, F. (1931). The tariff history of the United States. New York: G.P. Putnam’s Sons. U.S. Department of Commerce, Bureau of the Census. (1920). Fourteenth Census of the United States. Washington, DC: U.S. GPO (includes the 1919 Census of Manufactures and 1920 Census of Agriculture). U.S. Department of Commerce, Bureau of the Census. (1921, 1930, 1933, 1935). Biennial Census of Manufactures. Washington, DC: U.S. GPO. U.S. Department of Commerce, Bureau of the Census. (1930). Fifteenth Census of the United States. Washington, DC: U.S. GPO (includes the 1929 Census of Manufactures and 1930 Census of Agriculture). U.S. Department of Commerce, Bureau of the Census. (1934, 1947, 1950, 1956). Statistical abstract of the United States. Washington, DC: U.S. GPO. U.S. Department of Commerce, Bureau of the Census. (1935, 1945, 1950, 1954, 1959, 1964). Census of Agriculture. Washington, DC: U.S. GPO. U.S. Department of Commerce, Bureau of the Census. (1940). Sixteenth Census of the United States. Washington, DC: U.S. GPO (includes the 1939 Census of Manufactures). U.S. Department of Commerce, Bureau of the Census. (1947, 1954, 1963). Census of Manufactures. Washington, DC: U.S. GPO. U.S. Department of Commerce, Bureau of the Census. (1950). Census of Population: 1950. Washington, DC: U.S. GPO. U.S. Department of Commerce, Bureau of the Census. (1960). Census of Population: 1960. Washington, DC: U.S. GPO. U.S. Department of Commerce, Foreign and Domestic Commerce Bureau. (1900–1940, 1946). Foreign commerce and navigation of the United States. Washington, DC: U.S. GPO.

236

LEI (SANDY) YE

U.S. Department of Treasury. (1939, 1941, 1945, 1947, 1950, 1951, 1953, 1955, 1957, 1961). Annual report of the Secretary of the Treasury on the state of the finances. Washington, DC: U.S. GPO. Wright, P. (1935). Trade and trade barriers in the Pacific. Honolulu, HI: Institute of Pacific Relations.

APPENDIX A. CLASSIFICATION OF CATEGORIES USED TO CALCULATE TARIFF RATES Note: This appendix presents the goods and group of goods that are classified as semi-manufactures and finished manufactures in constructing the dataset on ad valorem equivalent tariff rates in Section 3. The classification attempts to closely match the one given by the Foreign Commerce and Navigation of the United States in 1931 (see pp. S68–S69), and modifications were made in other years to best keep the groupings consistent. Parentheses next to certain item groups provide more details regarding items within the group. 1. Semi-manufactures: Leather Furs, semi-manufactures (dressed furs; silk or black fox, dressed or undressed) Cod and cod-liver oil Whale and fish oil Stearic acid Beeswax and other animal wax (bleached beeswax, crude beeswax, manufactures of beeswax, animal wax, n.s.p.f.) Rubber, reclaimed and scrap Bristles, sorted or bunched Shellac Gelatin, inedible, and manufactures of Glue and glue size Casein or lactarene Gums and reins, n.e.s. (dutiable only) Vegetable oils, expressed, inedible, n.e.s. Tar, pitch, and turpentine Extracts for dyeing and tanning Gambier


237

Cotton semi-manufactures (cotton yarn, cotton waste) Jute yarns Yarns of flax, hemp, and ramie Silk waste Spun silk Silk yarns Hat materials Wool, semi-manufactures Rayon waste, yarns, and thread Cork waste Wood, unmanufactured (exclude logs) Sawmill products, except laths and shingles Boards, planks, etc., n.e.s.; and clapboards Veneer and ply woods Marble, onyx, and breccia (in block, rough, square, or tiles) Cement and lime Paper base stock Coke, charcoal, briquettes, etc. Petroleum: topped oils and tops Paraffin and paraffin wax Abrasives, crude and artificial Asbestos (all fibers, suco, asbestos) Precious stones (dutiable only, diamond cut, pearl and sets) Gypsum (exclude crude gypsum and plaster) Mica, cut, split, and manufactures of Talcum, steatite, soapstone, and French chalk (exclude crude) Magnesite (dead, burned, grain, and periclase) Other non-metallic minerals and manufactures of (only dutiable ones listed after salt) Iron and steel semi-manufactures (listings from granular or sponge iron to tin plates, terneplates) Copper (refined in ingots, plates, or bars) and brass (old brass and clippings) Ferro-alloys (dutiable only) Nickel oxide Tin in bars, blocks, etc. Cobalt ore and metal Platinum and platinum metals (ores excluded) Aluminum metal scrap and alloys

238

LEI (SANDY) YE

Lead (pigs, reclaimed scrap, dross, Babbitt, type metal) Nickel alloys, in pigs, n.e.s. Antimony (exclude ore) Zinc in pigs, blocks, etc. Quicksilver or mercury Other metals, alloys, etc. (dutiable only) Coal-tar products Industrial chemicals Pigments Fertilizer materials Perfume materials

2. Finished manufactures: Leather manufactures Fur manufactures Bone and horn manufactures Feathers, artificial, etc., and advanced (dressed feathers, feather, n.s.p.f.) Sponges and manufactures of Other inedible animal products (listing after tankage) Camphor, refined and synthetic Licorice extract Essential and distilled oils Tobacco manufactures (cigars and cigarettes, and manufactures n.s.p.f.) Starch Cotton manufactures (excluding waste and yarn) Jute manufactures (excluding yarns) Manufactures of flax, hemp, and ramie (exclude yarn) Hats, bonnets, and hoods of straw, etc. Hair manufactures Silk manufactures (exclude spun silk and yarn) Rayon manufactures (exclude yarn, waste, and threads) Miscellaneous textiles (all textiles listed after rayon manufactures) Wood manufactures Cork manufactures Refined petroleum oils Marble, breccia, and onyx manufactures Glass and glass products


239

Pottery and other clay products (pottery, tiles, bricks) Chalk manufactures (cubes, blocks, sticks, etc.) Earthy and mineral substances (exclude crude and unmanufactured material) Abrasives (dutiable only) Asbestos manufactures (shingles, other manufactures) Salt Steel-mill products (listing from structural steel to iron to autoclaves, parts, etc.) Iron and steel advanced manufactures Aluminum manufactures (plates, sheets, bars, etc., manufactures n.s.p.f.) Copper manufactures (composition metal, manufactures n.s.p.f.) Brass and bronze manufactures (exclude old brass and clippings) Lead, nickel, and zinc manufactures (bars, rods, plates of nickel, sheets of zinc, dust of zinc, pipes, sheet of lead, and manufactures n.s.p.f.) Other metal manufactures, n.s.p.f. Jewelry and manufactures of precious metals Machinery and vehicle Coal-tar finished products (coal-tar medicinal, other coal-tar) Medicinal products (listings from alkaloids to medicinal preparations containing alcohol) Paint, stains, enamels, and varnishes Fertilizer manufactures Explosives, fireworks, and ammunition Soap and toilet preparations

APPENDIX B. DATA SOURCES FOR REGRESSION VARIABLES Note: This appendix presents the sources of data used on production variables for each vote in the regressions of Appendix C (summarized by Table 4). All data are by value of production (or for the total agriculture variable starting in 1934, value of farms) unless otherwise noted (U.S. Department of Commerce & Bureau of the Census, 1920, 1930, 1940, 1950, 1960, 1935, 1945, 1950, 1954, 1959, 1964, 1921, 1930, 1933, 1935, 1947, 1954, 1963, 1934, 1947, 1950, 1956).

240

LEI (SANDY) YE

1922 Fordney–McCumber Act

Cotton manufactures: 1921 Biennial Census of Manufactures. Silk manufactures: 1921 Biennial Census of Manufactures. All agricultural variables: 1920 Census of Agriculture. Population: 1920 Fourteenth Census of the United States. 1930 Smoot–Hawley Act

Cotton manufactures: 1929 Census of Manufactures. Silk manufactures: 1929 Census of Manufactures. All agriculture variables: 1930 Census of Agriculture. Population: 1930 Fifteenth Census of the United States. 1934 Reciprocal Trade Agreements Act

Cotton manufacturing: 1933 Biennial Census of Manufactures. Silk manufacturing: 1933 Biennial Census of Manufactures. All agriculture variables: 1935 Census of Agriculture (1934 values for cotton, wheat, and wool; 1929 values for sugar; 1935 values for total agriculture). Population: 1930 Fifteenth Census of the United States. 1945 RTAA Renewal Cotton and rayon manufacturing (including silk manufactures): 1947 Census of Manufactures. All agriculture variables: 1945 Census of Agriculture (1945 values for total agriculture and 1944 values for other variables). Population: 1940 Sixteenth Census of the United States. 1949 RTAA Renewal Cotton and rayon manufacturing (including silk manufactures): 1947 Census of Manufactures. All agriculture variables: 1950 Census of Agriculture (1949 values for cotton, sugar, and wheat; 1944 values for sugar; 1950 values for total agriculture). Population: 1950 Census of the Population.

241


1955 RTAA Renewal Textiles: 1954 Census of Manufactures (by value added of textile mill products). All agriculture variables: 1954 Census of Agriculture (1949 values for wheat and 1954 values for other variables). Population: 1950 Census of the Population. 1962 Trade Expansion Act Textiles: 1963 Census of Manufactures (by value-added of textile mill products). All agriculture variables: 1964 Census of Agriculture. Population: 1960 Census of the Population.

APPENDIX C. PROBIT ANALYSIS OF VARIOUS TRADE BILLS Probit Analysis of Vote on 1922 Fordney–McCumber Act, 67th House. Variable

(1)

(2)

(3)

(4)

1.64 0.443 0.420 0.272 (6.96) (5.06) (2.32) (1.06) – – – Party 3.20 (11.76) Cotton Manuf. – 0.008 0.000 0.000 (2.58) (0.04) (0.20) Silk Manuf. – 0.043 0.032 0.035 (4.07) (2.86) (2.92) Wool – – 0.373 0.356 (2.67) (2.52) Cotton – – 0.026 0.026 (6.44) (6.44) Sugar – – 0.010 0.012 (0.34) (0.43)

Constant

(5) 7.20 (2.23) 8.00 (2.54) 0.004 (0.54) 0.047 (1.73) 0.748 (2.19) 0.033 (1.45) 0.316 (2.37)

242

LEI (SANDY) YE

APPENDIX C. (Continued) Variable

(1)

(2)

Wheat

–

–

Agriculture

–

–

Observations Pseudo-R2

298 0.64

(3) 0.005 (1.67) –

301 0.07

301 0.40

(4) 0.003 (0.84) 0.001 (0.82) 301 0.40

(5) 0.002 (0.46) 0.001 (0.36) 298 0.71

Note: Party: binary independent variable, where 0 ¼ Democrats and 1 ¼ Republicans; Vote: binary dependent variable, where 0 ¼ protectionist and 1 ¼ free trade. All production variables are on a per capita basis. Significant at the 10 percent level. Significant at the 5 percent level. Significant at the 1 percent level.

Probit Analysis of Vote on 1922 Fordney–McCumber Act, 67th Senate. Variable

(1)

(2)

Constant

1.405 (3.85) 2.639 (5.99) –

0.298 (1.77) –

0.227 (0.82) –

0.646 (1.50) –

0.008 (1.22) 0.035 (1.65) 0.005 (0.26) 0.076 (1.23) 0.009 (1.57) 0.019 (2.64) –

0.008 (1.31) 0.028 (1.29) 0.009 (0.42) 0.067 (1.10) 0.016 (1.86) 0.019 (2.55) 0.002 (1.28)

56.3 (0.30) 58.1 (0.31) 0.224 (0.48) 0.203 (0.06) 0.004 (0.17) 3.17 (0.32) 0.019 (1.72) 0.014 (1.17) 0.004 (1.79)

71 0.29

71 0.76

Party

Silk Manuf.

–

Wool

–

0.013 (2.24) 0.048 (2.15) –

Sugar

–

–

Wheat

–

–

Cotton

–

–

Agriculture

–

–

Cotton Manuf.


71 0.52

71 0.08

(3)

71 0.27

(4)

(5)

243


Probit Analysis of Vote on 1930 Smoot–Hawley Act, 71st House. Variable

(1)

(2)

(3)

(4)

(5)

1.31 0.200 0.372 0.369 1.11 (9.2) (2.54) (3.51) (2.14) (3.88) – – – 3.13 Party 2.66 (14.36) (11.68) Cotton Manuf. – 0.12 0.005 0.005 0.001 (4.52) (1.56) (1.55) (0.19) 0.011 Silk Manuf. – 0.042 0.030 0.030 (4.77) (3.35) (3.13) (0.71) Wool – – 0.060 0.060 0.088 (1.38) (1.38) (1.22) Sugar – – 0.306 0.306 0.543 (3.17) (3.16) (4.86) Wheat – – 0.005 0.005 0.019 (0.98) (0.82) (1.89) Cotton – – 0.034 0.034 0.009 (6.43) (6.39) (1.18) Agriculture – – – 0.00 0.007 (0.02) (3.87) Constant


374 0.55

375 0.10

375 0.25

375 0.25

374 0.67

Probit Analysis of Vote on 1930 Smoot–Hawley Act, 71st Senate. Variable

(1)

(2)

(3)

(4)

(5)

Constant

1.07 (4.07) 1.95 (5.83) –

0.030 (0.20) –

0.445 (1.86) –

1.25 (2.80) –

0.001 (0.07) 0.032 (1.30) 0.014 (0.57) 2.22 (1.04) 0.012 (1.75)

0.004 (0.41) 0.016 (0.66) 0.025 (1.01) 2.17 (0.86) 0.002 (0.17)

1.12 (1.60) 3.44 (4.53) 0.013 (0.88) 0.023 (0.51) 0.109 (2.58) Dropped

Party

Silk Manuf.

–

Wool

–

0.011 (1.72) 0.060 (2.46) –

Sugar

–

–

Wheat

–

–

Cotton Manuf.

0.012 (0.96)

244

LEI (SANDY) YE

APPENDIX C. (Continued) Variable

(1)

(2)

Cotton

–

–

Agriculture

–

–

83 0.35

86 0.09


(3) 0.123 (2.23) – 86 0.36

(4) 0.122 (2.13) 0.007 (2.13)

(5) 0.008 (0.57) 0.024 (3.15)

86 0.41

83 0.64

(4)

(5)

Note: Sugar dropped due to high number of 0-value cases.

Probit Analysis of Vote on 1934 RTAA, 73rd House. Variable

(1)

1.76 (12.87) Party 3.82 (11.95) Cotton Manuf. – Constant

Silk Manuf.

–

Wool

–

Sugar

–

Wheat

–

Cotton

–

Agriculture

–


381 0.75

(2) 0.657 (8.33) –

(3)

1.82 (7.03) 3.82 (11.37) 0.004 0.004 0.009 0.015 (2.50) (0.54) (0.53) (0.82) 0.073 0.050 0.051 0.010 (4.83) (3.08) (3.02) (0.36) – 0.015 0.014 0.080 (0.52) (0.47) (1.42) – 0.227 0.229 0.473 (1.24) (1.25) (2.04) – 0.008 0.009 0.011 (0.77) (0.79) (0.51) – 0.069 0.069 0.045 (3.94) (3.95) (1.75) – – 0.000 0.000 (0.21) (0.19) 385 0.05

0.464 (4.55) –

385 0.12

0.483 (3.54) –

385 0.12

381 0.77

245


Probit Analysis of Vote on 1934 RTAA, 73rd Senate. Variable

(1)

Constant

1.29 (5.37) 2.38 (6.31) –

Party Cotton Manuf.

(2) 0.337 (2.13) –

(3)

(4)

(5)

0.179 (0.77) –

0.038 (0.11) – 0.008 (0.69) 0.049 (1.38) 0.005 (0.20) 0.646 (2.35) 0.003 (0.17) 0.060 (2.51) 0.000 (0.61)

0.913 (1.93) 2.845 (5.64) 0.010 (0.54) 0.024 (0.39) 0.021 (0.50) 0.724 (2.19) 0.010 (0.42) 0.030 (0.97) 0.002 (1.82)

82 0.14

Silk Manuf.

–

Wool

–

0.014 (1.52) 0.066 (1.97) –

Sugar

–

–

Wheat

–

–

Cotton

–

–

Agriculture

–

–

0.007 (0.61) 0.052 (1.48) 0.009 (0.35) 0.658 (2.39) 0.002 (0.11) 0.059 (2.47) –

80 0.47

82 0.06

82 0.14


80 0.60

Probit Analysis of Vote on 1945 RTAA Renewal, 79th House. Variable

(1)

Constant

1.59 (11.49) 2.45 (13.9) –

Party

(2) 0.138 (1.95) –

–

0.006 (3.62) –

Sugar

–

–

Wheat

–

–

Cotton

–

–

Agriculture

–

–

Cotton Manuf. and Silk Manuf. Wool


391 0.49

392 0.05

(3) 0.025 (0.30) – 0.004 (2.07) 0.032 (0.98) 0.002 (0.04) 0.004 (1.44) 0.034 (4.56) – 392 0.15

(4)

(5)

0.306 (2.40) –

1.77 (8.01) 2.38 (11.64) 0.000 (0.24) 0.077 (1.43) 0.093 (1.64) 0.004 (1.21) 0.012 (1.42) 0.001 (1.73)

0.003 (1.49) 0.011 (0.35) 0.018 (0.32) 0.000 (0.09) 0.036 (4.80) 0.001 (2.83) 392 0.16

391 0.52

246

LEI (SANDY) YE

Probit Analysis of Vote on 1945 RTAA Renewal, 79th Senate. Variable

(1)

(2)

Constant

1.21 (4.86) 1.25 (3.72) –

0.510 (3.09) –

Party Cotton Manuf. and Silk Manuf. Wool

–

0.003 (1.08) –

Sugar

–

–

Wheat

–

–

Cotton

–

–

Agriculture

–

–


75 0.17

75 0.02

(3) 0.712 (3.16) – 0.000 (0.15) 0.054 (1.55) 1.26 (0.52) 0.008 (1.74) 0.014 (1.16) – 75 0.20

(4) 0.787 (2.39) – 0.000 (0.07) 0.050 (1.40) 1.26 (0.52) 0.007 (1.35) 0.014 (1.16) 0.000 (0.31) 75 0.20

(5) 1.47 (3.30) 1.16 (2.71) 0.001 (0.47) 0.078 (1.53) 1.09 (0.40) 0.005 (1.03) 0.004 (0.29) 0.000 (0.08) 75 0.29

Probit Analysis of Vote on 1949 RTAA Renewal, 81st House. Variable

(1)

(2)

1.96 0.833 (11.37) (10.34) – Party 1.78 (8.84) Cotton Manuf. – 0.005 and Silk Manuf. (2.18) Wool – – Constant

Sugar

–

–

Wheat

–

–

Cotton

–

–

Agriculture

–

–


387 0.29

388 0.02

(3) 0.732 (7.62) –

(4)

(5)

1.89 (8.36) 1.69 (7.41) 0.004 0.003 0.002 (1.51) (1.36) (0.70) 0.027 0.010 0.029 (0.63) (0.23) (0.56) 0.760 0.780 0.695 (0.52) (0.55) (0.55) 0.001 0.001 0.001 (0.68) (0.27) (0.37) 0.014 0.015 0.004 (2.81) (2.92) (0.86) – 0.000 0.000 (1.51) (0.36) 388 0.08

0.857 (6.75) –

388 0.09

387 0.30

247


Probit Analysis of Vote on 1949 RTAA Renewal, 81st Senate. Variable

(1)

(2)

Constant

2.04 (4.95) 2.15 (4.62) –

0.51 (3.02) –

Party Cotton Manuf. and Silk Manuf. Wool

0.033 (1.37) –

–

Sugar Wheat

– –

– –

Cotton

–

–

Agriculture

–

–


81 0.37

81 0.11

(3) 0.396 (1.64) – 0.035 (1.41) 0.009 (0.19) – 0.003 (1.27) 0.026 (1.45) – 81 0.23

(4) 0.599 (1.73) – 0.030 (1.26) 0.022 (0.43) – 0.002 (0.52) 0.025 (1.44) 0.000 (0.83) 81 0.24

(5) 2.17 (3.22) 2.23 (3.67) 0.038 (1.27) 0.058 (0.97) – 0.000 (0.13) 0.008 (0.44) 0.000 (0.46) 81 0.48

Note: Sugar dropped in column (5) because it predicts success perfectly.

Probit Analysis of Vote on 1955 RTAA Renewal, 84th House. Variable

(1)

Constant

Textile

1.00 (9.85) 0.767 (5.56) –

Wool

–

0.003 (2.58) –

Sugar

–

–

Wheat

–

–

Cotton

–

–

Agriculture

–

–

Party


405 0.07

(2) 0.719 (8.98) –

405 0.01

(3)

(4)

0.612 (6.34) –

0.519 (4.04) –

0.004 (3.15) 0.022 (0.71) 1.78 (1.59) 0.001 (0.41) 0.010 (2.71) –

0.004 (2.79) 0.032 (1.00) 1.83 (1.64) 0.002 (0.97) 0.009 (2.35) 0.000 (1.08)

405 0.06

405 0.06

(5) 0.994 (6.00) 0.791 (4.95) 0.005 (3.47) 0.036 (1.13) 0.790 (0.71) 0.003 (1.00) 0.004 (1.22) 0.000 (1.61) 405 0.12

248

LEI (SANDY) YE

Probit Analysis of Vote on 1955 RTAA Renewal, 84th Senate. Variable

(1)

(2)

Constant

0.980 (5.38) –

Textile

1.07 (4.46) 0.040 (0.12) –

Wool

–

0.002 (0.79) –

Sugar

–

–

Wheat

–

–

Cotton

–

–

Agriculture

–

–

Party


88 0.0002

88 0.01

(3) 1.75 (5.14) – 0.003 (1.02) 0.163 (2.56) 3.24 (0.51) 0.007 (2.11) 0.003 (0.36) – 88 0.34

(4)

(5)

1.35 (3.13) –

0.947 (1.93) 0.993 (1.79) 0.002 (0.52) 0.263 (2.76) 5.33 (0.75) 0.016 (2.16) 0.004 (0.40) 0.001 (1.05)

0.002 (0.66) 0.240 (2.71) 3.23 (0.55) 0.015 (2.12) 0.001 (0.13) 0.001 (1.29) 88 0.37

88 0.42

Probit Analysis of Vote on 1962 Trade Expansion Act, 87th House. Variable

(1)

(2)

Constant

0.508 (7.09) –

Textile

1.09 (11.06) 1.16 (8.44) –

Wool

–

0.001 (0.78) –

Sugar

–

–

Wheat

–

–

Cotton

–

–

Agriculture

–

–

Party


423 0.15

421 0.001

(3) 0.601 (6.85) –

(4) 0.712 (6.14) –

0.000 (0.30) 0.131 (2.17) 0.017 (0.59) 0.001 (0.50) 0.000 (0.21) –

0.00 (0.09) 0.083 (1.34) 0.0152 (0.53) 0.004 (1.17) 0.001 (0.27) 0.000 (1.48)

421 0.02

421 0.02

(5) 1.62 (9.37) 1.50 (8.93) 0.002 (2.03) 0.188 (2.28) 0.074 (2.46) 0.005 (1.63) 0.008 (2.93) 0.000 (0.79) 421 0.20

249


Probit Analysis of Vote on 1962 Trade Expansion Act, 87th Senate. Variable

(1)

Constant

Textile

2.11 (5.25) 1.41 (2.96) –

Wool

–

Sugar

–

–

Wheat

–

–

Cotton

–

–

Agriculture

–

–

Observation Pseudo-R2

86 0.21

83 0.02

Party

(2) 1.39 (6.43) – 0.002 (0.96) –

(3) 1.58 (5.24) – 0.003 (1.31) 0.029 (0.78) 21.8 (0.79) 0.002 (0.62) 0.009 (1.05) – 83 0.09

(4) 2.00 (4.84) –

(5)

0.004 (1.66) 0.014 (0.31) 19.3 (0.63) 0.005 (0.78) 0.006 (0.63) 0.000 (1.71)

18.3 (2.41) 15.0 (2.26) 0.040 (2.37) 0.513 (2.08) 0.795 (0.26) 0.005 (0.40) 0.055 (0.73) 0.000 (1.20)

83 0.14

83 0.66

APPENDIX D. SUPPLEMENTARY PROBIT USING DISTRICT-LEVEL DATA 1922 Fordney–McCumber Act, 67th House. Variables

Vote (1)

totalCotMill. totalSilMill Party totalWoolWors totalAgr

3.203 (0.272)

(6)

(7)

(8)

1.30e–05 2.48e–05 1.80e–05 (1.43e–05) (1.54e–05) (1.65e–05) 0.000211 0.000208 0.000101 (0.000109) (0.000119) (6.37e–05)

(9)

5.42e–06 (2.38e–05) 0.000300 (0.000266) 3.436 (0.352) 8.26e–05 4.69e–05 1.34e–05 (5.46e–05) (4.12e–05) (4.51e–05) 2.41e–05 1.13e–05 (4.34e–06) (8.13e–06)

250

LEI (SANDY) YE

APPENDIX D. (Continued) Variables

Constant Observations Pseudo-R2

Vote (1)

(6)

1.645 (0.236)

0.470 (0.0812)

298 0.641

301 0.0266

(7)

(8)

0.452 1.124 (0.0819) (0.144) 301 0.0393

(9) 2.156 (0.441)

301 0.124

298 0.653

(8)

(9)

Note: Standard errors in parentheses. po0.1. po0.05. po0.01.

1930 Smoot–Hawley Act, 71st House. Variables

Vote (1)

(6)

(7)

3.15e–05 1.29e–05 2.11e–06 3.27e–05 (1.26e–05) (1.65e–05) (1.65e–05) (2.52e–05) totalSilMill 0.000305 0.000314 0.000299 0.000213 (0.000112) (0.000118) (0.000118) (0.000178) Party 2.660 2.629 (0.185) (0.190) totalWoolWors 0.000551 0.000529 0.000302 (0.000208) (0.000208) (0.000298) totalAgr 1.91e–06 8.15e–07 (1.57e–06) (2.20e–06) Constant 1.305 0.163 0.141 0.200 1.331 (0.143) (0.0701) (0.0707) (0.0858) (0.166) totalCotMill.


374 0.549

375 0.0286


375 0.0612

375 0.0639

374 0.569

1934 RTAA, 73rd House. Variables

Vote (1)

totalCotMill. totalSilMill Party

3.817 (0.319)

(6)

1.89e–05 1.65e–05 (1.33e–05) (1.70e–05) 7.26e–05 6.68e–05 (3.79e–05) (3.80e–05)

totalWoolWors totalAgr Constant Observations Pseudo-R2

1.759 (0.137) 381 0.755

(7)

0.623 (0.0728) 385 0.0143

(8)

(9)

2.46e–05 (3.57e–05) 3.99e–05 (6.95e–05) 3.841 (0.333) 0.000114 0.000102 0.000112 (3.49e–05) (3.42e–05) (5.16e–05) 3.10e–06 2.53e–06 (1.03e–06) (2.50e–06) 0.639 0.498 1.695 (0.0732) (0.0841) (0.167) 385 0.0464

1.20e–05 (1.72e–05) 6.96e–05 (3.94e–05)

385 0.0734

381 0.774


1945 RTAA Renewal, 79th House. Variables

Vote (1)


2.464 (0.181)


1.622 (0.144) 380 0.494

(6)

(7)

(8)

6.31e–05 (2.24e–05) 4.76e–06 (3.01e–05)

6.43e–05 (2.27e–05) 2.85e–06 (3.05e–05)

(9)

5.83e–05 (2.77e–05) 2.89e–05 (3.25e–05) 2.621 (0.202) 0.000126 0.000128 0.000114 (3.90e–05) (3.95e–05) (3.99e–05) 1.13e–06 1.15e–05 (3.02e–06) (4.45e–06) 0.271 0.280 0.311 2.017 (0.0688) (0.0692) (0.107) (0.224)

1.05e–05 (1.37e–05) 2.03e–06 (2.99e–05)

381 0.00116


381 0.0305

381 0.0308

380 0.527

1949 RTAA Renewal, 81st House. Variables

Vote (1)


1.805 (0.213)

(6)

(7)

(8)

(9)

1.35e–05 (2.04e–05) 2.16e–05 (4.40e–05)

0.000236 (0.000118) 1.70e–05 (4.42e–05)

0.000249 (0.000119) 2.75e–05 (4.46e–05)

0.000225 (9.12e–05)

0.000237 (9.26e–05) 4.45e–06 (3.56e–06) 1.063 (0.130)

0.000259 (0.000153) 8.82e–05 (4.81e–05) 1.936 (0.243) 0.000263 (0.000117) 5.53e–06 (4.47e–06) 2.301 (0.258)


2.026 (0.185) 377 0.295

0.950 (0.0813) 378 0.00208

0.940 (0.0830) 378 0.0564

378 0.0609

377 0.353

(8)

(9)


1955 RTAA Renewal, 84th House. Variables

Vote (1)

(6)

(7)

3.48e–05 2.32e–05 2.83e–05 3.32e–05 (1.33e–05) (1.63e–05) (1.67e–05) (1.66e–05) totalSilMill 5.94e–05 5.85e–05 5.04e–05 4.25e–05 (2.89e–05) (2.90e–05) (2.97e–05) (2.97e–05) Party 0.760 0.763 (0.141) (0.145) totalWoolWors 3.15e–05 2.36e–05 1.90e–05 (2.69e–05) (2.72e–05) (2.60e–05) totalAgr 5.20e–06 3.34e–06 (3.54e–06) (3.61e–06) Constant 0.967 0.669 0.672 0.536 0.974 (0.103) (0.0736) (0.0737) (0.117) (0.148) totalCotMill.


380 0.0658

380 0.0240


380 0.0271

380 0.0322

380 0.0954

253


1962 Trade Expansion Act, 87th House. Variables

Vote (1)

(6)

(7)

(8)

(9)

totalCotMill.

4.92e–06 1.13e–05 7.14e–06 1.57e–05 (1.39e–05) (1.72e–05) (1.74e–05) (1.93e–05) totalSilMill 1.23e–05 1.15e–05 3.62e–06 1.55e–05 (3.41e–05) (3.40e–05) (3.38e–05) (3.55e–05) Party 1.133 1.198 (0.142) (0.146) totalWoolWors 1.81e–05 1.15e–05 2.11e–05 (2.91e–05) (2.92e–05) (3.18e–05) totalAgr 4.14e–06 7.34e–06 (3.04e–06) (3.21e–06) Constant 1.086 0.550 0.549 0.663 1.324 (0.101) (0.0701) (0.0701) (0.110) (0.149) Observations Pseudo-R2

398 0.140

398 0.000540


398 0.00137

398 0.00517

398 0.155

A COMPARISON OF FEDERAL FINANCIAL REMEDIATION IN THE GREAT DEPRESSION AND 2008–2009 Barrie A. Wigmore ABSTRACT Studies of Depression-era financial remediation have generally focused on federal deposit insurance and the provision of equity to banks by the Reconstruction Finance Corporation (RFC). This paper broadens the concept of financial remediation to include other programs – RFC lending, federal guarantees of farm and home mortgages, and the elimination of interest on demand deposits – and other intermediaries – savings and loans, mutual savings banks, and life insurance companies. The benefits of remediation or the amounts potentially at risk to the government in these programs are calculated annually and allocated to the various intermediaries. The slow remediation of real estate loans (two-thirds of these intermediaries’ loans) needs further study with respect to the slow economic recovery. The paper compares Depressionera remediation with efforts during the 2008–2009 crisis. Today’s remediation contrasts with the 1930s in its speed, magnitude relative to GDP or private sector nonfinancial debt, the share of remediation going to nonbanks, and emphasis on securities markets.


255

256

BARRIE A. WIGMORE

When President Roosevelt took office in 1933, the country’s economy and financial system had collapsed. The federal government commenced a multiyear series of measures to restore economic growth and employment. These included such measures as devaluation of the dollar, social security, unemployment insurance, the Works Progress Administration, the Civilian Conservation Corps, the Tennessee Valley Authority, and the National Recovery Administration. Initiatives also included a variety of measures to restore the health of the financial system, including the introduction of federal deposit insurance, expansion of the Reconstruction Finance Corporation (RFC), farm and home mortgage guarantees, and the elimination of interest on banks’ demand deposits. This paper examines the programs to remediate the financial system, understood to include the principal intermediaries: banks, savings and loans, mutual savings banks, and insurance companies. This broader view of the financial system contrasts with the frequent focus just on banks. Financial remediation programs differ from outright expenditures to stimulate the economy because there was an expectation that funds advanced would be repaid and that losses on mortgage guarantees would be covered by insurance fees or program earnings. Commitments under remediation programs are therefore not necessarily commensurable because they varied widely in the ex ante expectations for losses and there were no reliable estimates of probable losses. Nonetheless, summing the benefits of remediation or the amounts potentially at risk to the government in these programs yields a worst case government exposure equal to 19 percent of GDP and 10 percent of all private sector debt for the Depression period. Tracing the growth of these commitments over the years 1932–1938 also reveals the extended period over which remediation developed. Not everything was done during Roosevelt’s first 100 days. This paper does not attempt to measure the effectiveness of remediation in restoring financial sector institutional stability or encouraging economic growth. But it does establish that while nonbanks represented the largest share of lending in the 1930s, banks got over 70 percent of the benefit from remediation. And while real estate loans represented two-thirds of all lending, this category of loans continued to experience high levels of distress because remediation programs developed slowly and commercial real estate loans received no remediation at all. These deficiencies in dealing with real estate debt deserve further exploration as a possible explanation for the continuing weakness in the economy prior to the advent of World War II.

Financial Remediation in the Great Depression and 2008–2009

257

In the 2008–2009 crisis, the federal government’s effort to avoid a collapse of the proportions of the 1930s quickly led to numerous remediation programs. Again, there was an expectation that commitments under these programs would be repaid so that the commitments were not the same as normal government expenditures. By the end of 2009, the aggregate benefits of remediation or the amounts potentially at risk to the government in these programs already amounted to 97 percent of GDP and 59 percent of private sector, nonfinancial debt – five times the proportionate level in the 1930s. This calculation is, however, likely to be controversial because of the wide discrepancy between what the government has announced as its commitments for legislative, budgetary, or accounting purposes and what creditors, ratings agencies, and derivatives and other counterparties believe these commitments are actually. This is particularly the case for Fannie Mae (FNMA, the Federal National Mortgage Association), Freddie Mac (FHLMC, the Federal Home Loan Mortgage Corporation), and AIG (the American Insurance Group), institutions which the government now controls. The huge commitments to these three companies alone mean that 62 percent of commitments in the 2008–2009 episode has been to nonbanks. And in contrast with the 1930s, when the Federal Reserve is widely acknowledged to have played no meaningful role in restoring either the financial system or economic growth, the Federal Reserve has complemented Treasury and FDIC remediation with emergency open-market programs that have amounted to a further 18 percent of GDP or 11 percent of private sector, nonfinancial debt. Like the 1930s, commercial real estate is still unattended to. For the purposes of this paper ‘‘financial remediation’’ means federal efforts only, and is concerned with financial instruments such as equity investments, loans, or guarantees which bolstered the credit of either financial intermediaries or borrowers. Financial remediation programs aim to ‘‘cure, correct, or put right’’ (Random House College Dictionary) perceived financial ills. The purpose of financial remediation was and is to repair or sustain the credit process in order to prevent economic decline or to actually increase the granting of credit and thereby stimulate economic growth. The stakes are therefore large, yet the costs of financial remediation programs are not generally estimable at the time because by their very nature they are adopted in crises. The costs, as noted, are generally different from other government expenditures to sustain or stimulate the economy such as social security, public works, unemployment insurance, medicare/ Medicaid, or welfare because the expectation is that the equity investments

258

BARRIE A. WIGMORE

or loans will be paid off or that the costs of guarantees will be recovered through insurance-like fees. Significant financial remediation programs in the 1930s aimed at sustaining home and farm ownership, but with this exception, remediation efforts have not aimed at advancing other social objectives such as income equality, labor rights, equal opportunity, education, or the state of the environment. The structure of this paper is as follows. Section 1 describes each of the elements of financial remediation from 1932 to 1938 and calculates the maximum commitments or liabilities under each program annually. Section 2 describes how financial stress persisted throughout the 1930s in real estate loans which was the dominant (67 percent) area of lending, and also describes the importance of commercial real estate lending which went completely unremediated. Section 3 calculates the disproportionate benefit of remediation efforts gained by banks versus savings and loans, mutual savings banks, and life insurance companies compared to the relative size of their loan portfolios. Section 4 describes each of the financial remediation programs in the current crisis (2008–2009), taking considerable effort to differentiate the practical extent of the liabilities government assumed versus its announced legislative or budgetary costs. The maximum liabilities of each of these programs are aggregated and compared with GDP and total private sector, nonfinancial debt. The Federal Reserve’s emergency open-market programs are described and aggregated in the same fashion. Section 5 compares remediation in the current crisis (2008–2009) and the 1930s.

1. FINANCIAL REMEDIATION EFFORTS, 1932–1938 The Depression saw the first broad effort at the federal level to cure the ill effects on the economy of distress among financial institutions and dysfunctional real estate mortgage structures. The role of the RFC in rescuing the banking system by providing equity to the banks has long been recognized as a critical component of Depression-era financial remediation, and federal deposit insurance has been credited with stabilizing the banking system, most notably by Milton Friedman and Anna Jacobson Schwartz. But there were many other important programs which are aggregated here for the first time.


259

The RFC provided loans to failed or struggling banks and advances to nonbank financial intermediaries that were three times its equity advances to banks; Congress provided an important subsidy to banks by forbidding the payment of interest on demand deposits (mostly business checking accounts); and the farm and home mortgage markets, which had collapsed by 1933, were replaced by markets with government guaranteed mortgages that had extended maturities and much higher loan-to-value standards. The cumulative amounts advanced or guaranteed under these programs from 1932 to 1938 totaled over $14 billion, equal to 19 percent of GDP and 10 percent of all private sector debt in 1931. I have excluded deposit insurance from these calculations because I am unable to quantify the liability. This section provides brief historical details on each of these programs and describes how remediation efforts grew gradually between 1932 and 1938.

1.1. The Reconstruction Finance Corporation The RFC ultimately advanced over $5 billion for financial remediation, but its origins can be traced to faltering federal government steps in 1931 to assist the struggling railroad industry which was the nation’s largest corporate creditor. The Railroad Credit Corporation was set up in 1931, but ultimately lent only $74 million. The National Credit Corporation was created in October 1931 to recycle the benefits of a rate increase that the Federal Commerce Commission allowed the railroads at that time, but it lent only $155 million (Spero, 1939, pp. 17–18). The RFC was set up in January 1932 with a broader mandate to lend to both railroads and commercial banks on an emergency basis. By the end of 1932, it had authorized $337 million in loans to railroads. This was the only program that focused on remediation for nonfinancial corporate borrowers and it came to an abrupt end when there was a public outcry that banks were being bailed out (Wigmore, 1985, p. 311). In the Emergency Banking Act of 1933 that followed the Bank Holiday, the RFC gained powers that made it an all-purpose vehicle for New Deal programs (and eventually war preparation programs), borrowing virtually all (over 93 percent) of its funds from the U.S. Treasury (Statistical Abstract, 1935, p. 201). Our interest here is confined to its financial remediation activities which were mostly complete by the end of 1934. Most importantly, the RFC was empowered to invest in the equity capital of banks, as distinct from making secured loans to them. This has been widely recognized as contributing

260

BARRIE A. WIGMORE

dramatically to stabilization of the banking system after the Bank Holiday. The RFC made equity capital commitments by the end of 1934 exceeding $1.1 billion, equal to 22 percent of the banking industry’s total capital. Initially, this power appeared most important in assisting small banks to qualify for federal deposit insurance, as the RFC authorized only $101 million through October 1933. However, there was renewed concern about the banking system in the fall of 1933 – worries about the Chicago banks’ solvency, concern that too many banks would not qualify for deposit insurance, and impatience with the slow growth in banks’ loan volume – and in response the RFC authorized equity investments between November 1933 and March 1934 totaling $833 million, but under circumstances that mixed political pressure on the banks to accept additional equity and true need (Report of the RFC, 1st Quarter, 1934, p. 32; Jones, 1951, pp. 35–36; Olson, 1988, pp. 77–83). This equity capital program is outlined as part of Table 1 and was 87 percent complete by the end of 1934. The RFC had a host of other financial remediation programs, however, also outlined in Table 1. They have not generally been treated as having great importance, even though they constituted 75 percent of RFC advances for financial remediation and were vital to bank liquidity. The RFC lent over $2.4 billion to 7,345 banks between 1932 and 1938, which was equal to

Table 1.

Advances Authorized by Reconstruction Finance Corporation (Cumulative) ($Millions).

Equity Loans to Savings Loans to Life Mortgage Advances Banks & Insurance Loan to Banks Loans Companies Companies

1932 1933 1934 1935 1936 1937 1938

0 497 1,154 1,237 1,245 1,262 1,321

948 1,794 2,220 2,328 2,394 2,438 2,458

100 122 143 145 146 148 154

83 116 133 133 137 137 139

95 334 418 449 495 538 644

Financial Component of Railroad Advances (See Note)

Total RFC Advances

202 247 275 296 374 386 448

1,428 3,110 4,343 4,588 4,791 4,909 5,164

Note: Only 60 percent of RFC loans to the railroads were for bond, interest, or bank loan payments between 1932 and 1937 (Spero, 1939, pp. 36–41). Source: RFC report for the fourth quarter of 1938 (pp. 47–92).


261

over 10 percent of all commercial banks’ less liquid loans and investments at the end of 1933 (defined to exclude U.S. treasuries) (Banking & Monetary Statistics 1914–1941, p. 19). While $1.1 billion of this was characterized as loans in aid of the reorganization or liquidation of failed banks, in fact most of the banks receiving loans were in parlous condition. Seventy-nine railroads received over $0.4 billion that could be classified as financial remediation – only 60 percent of advances because 40 percent of RFC railroad loans were related to railroad capital spending and ‘‘make-work’’ projects (Spero, 1939, pp. 36–41). Four hundred and twenty-two mortgage loan companies, virtually all of which were in liquidation, received over $0.6 billion to assist in the process. Eighty-four percent of all RFC efforts were complete by the end of 1934, despite the persistence of financial stress for years thereafter.

1.2. Farm Financing By the time of President Roosevelt’s inauguration the mechanisms of farm financing had collapsed under the pressure of a 40 percent decline in farm commodities prices and over-expansion in World War I. Farm finance needed dramatic restructuring to preserve economic activity (approximately 50 percent of all Americans lived in communities of 5,000 people or less), to bail out struggling financial intermediaries, and to preserve the virtues of family farming. Forty-five percent of farm mortgages were delinquent and rural violence was spreading across the country at foreclosure sales and court hearings. Moratoriums had been declared on farm foreclosures in 20 states and other restrictions on creditors had been imposed in 32 states (Jones & Durand, 1954, pp. 4, 12). Life insurance companies – the largest institutional mortgage lenders to farmers – had 24 percent of their farm investments in foreclosure. Insurance company foreclosures would rise to 37 percent by the end of 1935 (Woodruff, 1937, p. 74). Eighty-eight of the 133 joint stock land banks originally set up in the Federal Farm Loan Act of 1916 to assist farm lending were in receivership, and their bonds had dipped below $35 versus $100 par value. The resulting remediation of farm financing extended federal loans of $2.3 billion to farmers on farm mortgages with longer maturities, lower interest rates, and much higher loans relative to value. These loans were financed by the Federal Farm Mortgage Corporation with a U.S. government guarantee.1 This amounted to 25 percent of the $9 billion in

262

BARRIE A. WIGMORE

farm mortgage debt outstanding at the end of 1932 (Federal Reserve, Agricultural Finance Databook, 1976, p. 8). The details of farm financing were complicated by a long political history of support for farm interests, farmers’ suspicion of banks and processors, an ideological desire to establish cooperative business facilities, and pressure from nonfarm interests to make farm financing self-sustaining. This resulted in a mix of private, public, and cooperative financing for farm mortgages and short-term production and marketing credit. There were two important farm acts in 1933 related to farm financing, one in 1934, and others in 1935 and 1937. Table 2 outlines what these acts meant for the remediation of farm loans. In the Emergency Farm Mortgage Act of 1933, Congress authorized over $2 billion in new farm mortgage lending authority for the Federal Land Banks and Land Bank Commissioner. Approximately 90 percent of this mortgage financing was for refinancing existing mortgages, but loans could also be used for working capital or to acquire land lost through foreclosure. Terms for the new loans were extremely lenient – maturities up to 20 years at rates not to exceed 5 percent, and no principal payments for 5 years. While lip service was made to protecting the federal government against losses, in fact appraisals were based on ‘‘normal agricultural values’’ which were defined using crop prices in 1909–1914 which were almost double those in 1933, and loans were up to 75 percent of appraised value, and even included second mortgages. By contrast, the rules for national banks limited farm loans to 50 percent of current appraisals and 5 years maturity – a highly

Table 2.

Federal Farm Credit Remediation ($Millions).

Federal Land Federal Land Bank Bank Mortgage Commissioner Loan Advances Mortgage Loan Advances 1932 1933 1934 1935 1936 1937 1938

28 152 730 249 100 63 51

71 553 196 77 41 29

Cumulative Federal Mortgage Loan Advances

223 1,506 1,951 2,128 2,232 2,312

December 31 Outstanding Farm Credit Administration Short-Term Loans 366 506 532 559 583 653 644

Source: Statistical Abstract (1940, p. 278, which cites Farm Credit Administration Annual Report as the source).


263

dysfunctional structure in a period of declining property values (Farm Credit Administration Annual Report, 1933, pp. 12, 14, 16, 82–83; 1934, pp. 11, 96–99; 1939, p. 18). Many of the changes in farm financing during the Depression related to short-term production financing which need not concern us here. While these changes were important to farmers, they essentially replaced one form of failed federal farm financing with another and kept federal support for short-term farm production financing relatively constant as can be seen in Table 2. This remediation, therefore, had little impact on private financial institutions. The new farm mortgage financing, however, had a large impact on private financial institutions, although it developed slowly. Only $42 million had been advanced under the new mortgage lending powers by September 1933, when a special unit within the Farm Credit Administration was set up to aid farmers facing foreclosure. Loan closings surged to almost $100 million in December 1933, and there was over $1.7 billion in new farm mortgage financing in 1934–1935 (Table 2). The aggregate in new financing reached $2.3 billion in 1938. Large proportions of this refinanced existing borrowers – thereby benefiting their lenders – as will be seen below.

1.3. Housing Urban home mortgage financing was in similar straits when President Roosevelt came into office. Under his administration new federal housing programs and agencies emerged. By 1938, they had lent or guaranteed over $5 billion on home mortgages to cure defaults, alter the structure of home financing, and finance home improvements. This new structure of federally guaranteed home financing has persisted to this day. There is no national data to illustrate the extent of the crisis in home financing when Roosevelt came into office, but in 22 cities home-owner defaults exceeded 21 percent and they were up to 62 percent in Cleveland (Bernanke, 1983, p. 4). In Massachusetts, the default rate for mutual savings bank loans on 1–4 family residences was 22 percent for loans made between 1924 and 1929 (Lintner, 1948, p. 505). Foreclosure rates on 1–4 family homes by the 24 leading life insurance companies reached 23 percent for loans made between 1925 and 1931 (Saulnier, 1950, p. 84). There were approximately 250,000 urban family home foreclosures in each of 1932 and 1933 that probably amounted to $1 billion each year. Urban home mortgage debt declined 15 percent, or almost $3 billion, between 1930 and 1933 because of

264

BARRIE A. WIGMORE

financial pressures (Federal Home Loan Bank Board, Eighth Annual Report, 1940, p. 12). These data understate the home mortgage stress because there was a considerable gap in time before delinquencies turned into foreclosures. Technical financial problems greatly exacerbated the situation as they had for farm financing. As noted, federal banks (and most state banks) were limited to mortgage loans maturing in 5 years at a maximum of 50 percent of current appraisal value (Board of Governors of the Federal Reserve System, Annual Report, 1934, p. 48). Except for savings and loan mortgages, most home mortgages made no provision for amortization of principal. In a market of declining home prices, this invited defaults when it came time to roll over a mortgage. New Deal home mortgage remediation measures revolutionized home financing. In June 1933, President Roosevelt approved the Home Owners Loan Corporation Act, authorizing $2 billion in federal funds to refinance troubled urban home mortgages. The Home Owners Loan Corporation (HOLC) specifically targeted refunding mortgages for the unemployed, those in ‘‘economic misfortune beyond applicant’s control,’’ borrowers in foreclosure, and mortgages held by institutions in liquidation. Authorizations were raised to $4.75 billion in 1935. The HOLC exchanged its own bonds (voluntarily) with lenders at 80 percent of the amount owed by the borrower and then refinanced the borrowers with loans for up to 80 percent of a very generous property appraisal to a maximum of $4,000, no principal payments for 3 years, 5 percent interest, and 15-year maturities. These were unprecedented terms compared to the 3–5 year bullet loans typical anywhere other than savings and loans and the loan limit of 50 percent of appraised value at national banks. There was little effort to protect the government against loss on these loans. As part of giving the borrower a new mortgage, the HOLC paid off any delinquent taxes or assessments, provided funds for necessary repairs, and funded the incidental expenses of the refinancing. Appraisals were well above current market value because replacement cost estimates were used, as well as capitalization of possible rental values over the last 10 years. The program was expected to have large losses. Initially, the federal government only guaranteed interest on the initial HOLC bonds that were exchanged with lenders, which made the bonds trade at a discount. Because the HOLC had to raise funds to finance its operations, the federal government quickly came around to a full guarantee of principal and interest on HOLC bonds. This helped the HOLC to refund most of the initial bonds when interest rates dropped and substantially widen the spread between its costs and the interest it received on its loans. Ultimately, 40 percent of the HOLC loans were foreclosed by the

265


government, but when the final liquidation of all loans was achieved in 1951 the conclusion was that there had been a ‘‘slight profit,’’ basically because the HOLC was subsequently able to refinance its original bonds at very low rates (Harriss, 1951, p. 6). The government also moved dramatically to restructure home mortgages that were not in default because of the problem with five-year terms and the constraints on renewing a mortgage at only 50 percent of appraised value when prices had fallen substantially. The National Housing Act of June 1934 created the Federal Housing Agency (the ‘‘FHA’’) to provide a federal guarantee (insurance) for virtually any middle class home mortgage. Nationwide standards were established under which borrowers and lenders could apply for a federal guarantee on a loan in return for a small fee that was expected to cover future losses. Losses were expected to be small and the guarantees never appeared even in the government’s accounting for its ‘‘contingent liabilities,’’ which were limited to guaranteed bonds. The loans were for 20 years at up to 80 percent of appraised value to a maximum of $16,000. ‘‘Character loans’’ without specific collateral requirements for home improvement could also be guaranteed. Annual commitments under the HOLC home mortgage relief program and the FHA guarantee programs are outlined in Table 3. Large volumes under these programs occurred later than both RFC remediation and farm mortgage relief. HOLC financing accelerated from a modest start in 1933 of only $103 million to over $3 billion by the end of 1936 when the program effectively ended. At this point, it amounted to 29 percent of all institutional

Table 3. HOLC Mortgage Loans 1933 1934 1935 1936 1937 1938

103 2,203 2,903 3,103 3,153 3,228

Federal Home Mortgage Remediation (Cumulative) ($Millions).

FHA 1–4 Family Mortgage Guarantees

FHA Home Improvement Guarantees

FHA Total

HOLCþFHA Total

96 407 842 1,363

30 253 499 559 732

30 349 906 1,401 2,095

103 2,233 3,252 4,009 4,554 5,323

Note: HOLC, Home Owners’ Loan Corporation; FHA, Federal Housing Administration. Sources: Federal Housing Administration, Annual Report (1940, p. 7); Report of the Federal Home Loan Bank Board (1939, Chart XVII on p. 46).

266

BARRIE A. WIGMORE

nonfarm 1–4 family mortgage financing.2 FHA mortgage guarantees did not become significant until 1936, but reached over $2 billion by 1938. The HOLC and FHA combined reached $5.3 billion in 1938, equal to over 50 percent of all 1–4 family home mortgage financing. Refundings accounted for 63 percent of FHA guarantees in 1935–1936, according to the FHA (Federal Housing Administration, Annual Report, 1935, p. 18; 1936, p. 28). This appears to correspond to refinancing on existing homes; however, refinancing was surely greater than this. New homes receiving FHA guarantees probably repaid builders’ loans, and home improvement loans improved the collateral of existing lenders and facilitated sales.

1.4. Eliminating Interest on Demand Deposits The Emergency Banking Act of 1933 provided a dramatic subsidy to banks by eliminating interest on ‘‘noninterest bearing, demand deposits’’ (checking accounts, mostly maintained by businesses) which were 59 percent of all deposits at the time.3 This probably increased bank earnings by almost $1.5 billion (cumulatively) by the end of 1938 and was a different form of remediation because there was no sense in which it was going to be paid back. It was more akin to a business tax remitted to the banks with the aim of increasing system strength. The Emergency Banking Act also required the Federal Reserve henceforth to regulate banks’ time deposit rates. This was a further, but lesser, subsidy whose value I have not tried to estimate. The subsidy of eliminating interest on demand deposits was immediate. As soon as banks reopened after the Bank Holiday, they no longer paid interest on demand deposits. A rough estimate of the subsidy to banks is outlined in Table 4. Banks’ net current earnings in the ensuing years were 175 percent higher than they would have been if they had had to pay 1 percent interest on demand deposits. Over the 5.5 years from mid-1933 to 1938 this amounted to almost $1.5 billion in incremental earnings (and capital), putting it on a scale greater than $1.3 billion of RFC capital advances to the banks, although the benefits of these earnings were more drawn out than the RFC’s quick direct action. One might expect that the low level of money market rates during the 1930s would have minimized the benefit of free demand deposits, but it did not. While rates on prime 4–6-month commercial paper dropped from 1.375 percent at the end of 1932 to only 0.875 percent at the end of 1934, bank time deposit rates averaged 2.8 percent in 1933, 2.6 percent in 1934, and 2.4 percent in 1935 (Federal Reserve, Banking & Monetary Statistics 1914–1941, p. 265).

267


The Effect on Operating Earnings of Interest-Free Deposits (All Member Banks – $Millions).

Table 4.

Demand Deposits

1933 1934 1935 1936 1937 1938

Operating Earnings Before Write-Offs

17,700 21,100 25,500 29,500 30,300 30,000

Totals

1% Interest Cost on Demand Deposits

Operating Earnings if 1% Demand Deposit Cost

Operating Earnings Increase from Interest-Free Demand Deposits 45% 115% 214% 284% 261% 357%

378 394 374 399 419 384

118 211 255 295 303 300

260 183 119 104 116 84

2,348

1,482

866

Note: Assumes no interest on demand deposits for two-thirds of 1933. Source: Banking & Monetary Statistics 1914–1941 (p. 263).

Table 5.

1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938

Federal Reserve Member Banks’ Interest Rate Margins.

Interest Costs (%)

Rates on Business Loans (%)

Interest Costs versus Loan Rates (%)

2.20 2.10 2.10 1.70 1.50 1.20 0.90 0.70 0.46 0.44 0.41

5.24 5.88 5.12 4.58 5.01 4.75 4.02 3.58 3.22 3.06 3.00

42 36 41 37 30 25 22 20 14 14 14

Note: Business loans in 7 northern and eastern cities and 11 southern and western cities. NYC banks’ loans are excluded because they made so few business loans. Loan rates are annual averages of monthly rates. Interest costs are annual interest divided by year-end deposits. Sources: Federal Reserve Annual Report (1932, p. 160, Ibid. 1933, p. 226, Ibid. 1934, p. 178, Ibid. 1935, p. 158, Ibid. 1937, pp. 101, 140–141) and Survey of Current Business (1940, Supplement, p. 51).

Nor was the advantage of free demand deposits competed away, despite Friedman and Schwartz’s claims to the contrary (Friedman & Schwartz, 1963, p. 444, Footnote 22). Table 5 outlines the annual interest cost divided by all year-end deposits for Federal Reserve member banks for the years

268

BARRIE A. WIGMORE

1928–1938. Despite relatively stable time deposit rates, interest costs dropped to only 0.4 percent because over half of all deposits were free. The annual averages for monthly business loan rates for the same years for banks in 18 northern, eastern, southern, and western cities remained above 3 percent, which meant that deposits cost only 14 percent of loan rates in 1936–1938 versus 30 percent in 1932.4 Earnings of savings & loans and mutual savings banks presumably also benefited from the elimination of interest on demand deposits. But the absence of profit motivation in these institutions and their lower vulnerability to suspension may have led to more of this benefit being passed through to customers. In any case, the benefit was small because they had smaller demand deposit bases.

1.5. Deposit Insurance Deposit insurance, introduced in the Emergency Banking Act of 1933, provided a federal guarantee of commercial bank deposits up to $2,500, which was subsequently increased to $5,000 in 1934. Neither savings and loan nor mutual savings bank deposits were eligible. The Federal Housing Act of 1934 provided similar insurance to savings and loan corporations, but they were very slow to sign up for it (Federal Home Loan Bank Board, Statistical Summary, 1948, p. 5). Mutual savings banks basically refused to sign up for federal deposit insurance. Academic opinion on the merits of deposit insurance has evolved since Friedman and Schwartz gave it credit for stabilizing the commercial banking system after the Bank Holiday (Friedman & Schwartz, 1963, p. 440). O’Grada and White (2002), Calomiris and Mason (1997), and Wigmore (1985, 1987) have shown that it was the largest depositors that provoked bank suspensions rather than those benefiting from deposit insurance. Calomiris and Mason (2003b) have shown that adding bank panics as a variable to regressions that explain bank failures adds no explanatory power beyond the underlying fundamental data, and Calomiris (1990) has questioned the necessity of having deposit insurance at all. Calomiris and White (1994) and White (1998) have reduced deposit insurance to a political victory for small banks and a system fraught with distortion of risk-taking incentives. It has been roundly blamed for fomenting the savings & loan crisis of the early 1990s (White, 1991). Calomiris (1990) turned the whole question of deposit insurance on its ear,


269

suggesting its best role is to discipline risk-taking and that the best structure would be based on peer sponsorship and enforcement. Despite this shift in opinion, both noneconomic historians and journalists amidst the crisis today give deposit insurance repeated credit for stabilizing the banking system.5 Deposit insurance did have a benefit to small depositors since they suddenly had a federal guarantee, but it is difficult to calculate what benefit deposit insurance provided to banks rather than depositors. The benefit occurred insofar as deposit insurance was able to forestall runs on otherwise solvent banks or to give insolvent banks the chance to recover. It did nothing for bank capital or the liquidity of bank assets. As of mid-May 1933, over 75 percent of all member bank deposits were greater than the initial $2,500 insurance cap. This only dropped to 62 percent when the higher $5,000 cap became effective in October 1934.6 Looked at another way, the larger the bank, the fewer of its deposits were insured. The 96 largest banks, with $50 million or more in deposits, had only 26 percent of their deposits insured as of October 1934 when the cap was $5,000. There were 631 banks with deposits between $5 and $50 million that had only 50 percent of deposits insured. It is difficult to see how these banks gained any benefit since larger depositors with their superior knowledge of and insight into the banks’ affairs could draw out more than enough deposits to force a suspension of payments. But there were over 13,000 small banks with deposits of less than $5 million each that had 69 percent or more of deposits insured. Deposit insurance presumably had a benefit to these small banks. These small banks held approximately $9 billion in deposits versus $36 billion in total system deposits. There were off-setting costs however. The banks had to pay for deposit insurance, and the stronger banks were hurt by a system that kept the weaker banks in business (FDIC, Annual Report, 1934, p. 20; 1938, pp. 80–81). I simply see no way to quantify the benefit that banks collectively received from deposit insurance. It is straightforward to quantify the various other financial remediation efforts. The benefit assumption for small banks cannot have been very large based on the savings & loan experience at the time. Savings and loans were generally small, with an average asset size of less than $1 million in 1934, but they were slow to sign up for deposit insurance. Savings & loans with deposit insurance held only 5 percent of industry assets at the end of 1934, a figure that did not reach 50 percent until 1938 (Federal Home Loan Bank Board, Statistical Summary, 1948, p. 5). In the same vein,

270

BARRIE A. WIGMORE

the great majority of the mutual savings banks also refused to participate in federal deposit insurance.

1.6. The Cumulative Impact of Remediation The cumulative impact of Depression-era financial remediation was much greater and took much longer to accumulate than one would think focusing on the climactic period of the Bank Holiday of 1933 and the many new programs introduced by President Roosevelt upon gaining office. The expansion of the RFC powers was vitally important and quick, with most of its impact felt by the end of 1934, but the subsidy to bank earnings of eliminating interest on bank deposits resulted in an accumulated increase in bank equity over the next five years even greater than that resulting from RFC equity advances. Federal home mortgage loans from the HOLC and FHA guarantees of home mortgages built up to $5.3 billion by 1938 – equivalent to the funds advanced by the RFC – but occurring later. Federal Farm Land Bank mortgage loans ramped up from zero to over $2 billion in 1934–1935. Two factors underlay the growth in these programs – the administrative momentum necessary for large scale and the tendency of Congress to expand the programs as the economy failed to recover. The amounts cited for these programs did not all, on a dollar for dollar basis, represent the same prospective or retrospective burden on taxpayers. There were different loss assumptions and different risks with different programs. Some involved equity advances as opposed to loans, and some involved outright loans as opposed to guarantees of loans made by others. What these programs had in common was that there was an amount of capital that was advanced, lent, or guaranteed, and in the worst case this represented a federal loss or liability. The amounts advanced, lent, or guaranteed are cumulated annually for each of these programs in Table 6. The most significant anomaly in adding up these various programs is the bank equity gained from the elimination of interest on banks’ demand deposits. As noted, this was effectively a tax on business depositors remitted to the banks; thus taxpayers did not face a potential or actual liability. These subsidies are, nevertheless, included in Table 6 aggregation. The resulting picture of financial remediation during the Depression shows a modest impact in 1933 of only $3.7 billion, growing to $14.7 billion by 1938. This was equal to 19 percent of 1931 GDP, the year before

271


Table 6. RFC Equity Plus Loans

1932 1933 1934 1935 1936 1937 1938

1,428 3,110 4,343 4,588 4,791 4,909 5,164

Total Financial Remediation Effort (Cumulative) ($Millions). Housing: Federal Cumulative Total Reme- Total as a HOLC Farm Land 1% Deposit diation % of GDP Loans plus Bank Subsidy to in 1931 of FHA Loans Banks $76 Billion Guarantees

103 2,233 3,252 4,009 4,554 5,323

70 292 1,576 2,021 2,207 2,310 2,391

118 329 584 879 1,182 1,482

1,498 3,685 8,613 10,647 12,158 13,297 14,773

2% 5% 11% 14% 16% 17% 19%

Total as a % of Private Sector Debt in 1931 of $148 Billion 1% 2% 6% 7% 8% 9% 10%

Sources: See Tables 1–4. Source for GDP and Private Debt – Bureau of the Census, ‘‘Historical Statistics of the United States’’ (1975, pp. 224, 989).

remediation began with the creation of the RFC, and 10 percent of all outstanding private sector debt at that time.

2. REAL ESTATE LENDING If we are to understand the Depression era’s remediation programs we need to know what financial institutions were doing the lending. Table 7 may present an unfamiliar image to most readers since commercial banks accounted for only 38 percent of all institutional lending in 1931. Table 7 illustrates the distribution of loans of all types (business, consumer, personal, mortgage) among both member and nonmember commercial banks, savings and loans, mutual savings banks, and life insurance companies from 1931 to 1938. These intermediaries dominated institutional lending. Others categories, such as finance companies, mortgage companies, fidelity and guaranty companies, credit unions, and charitable institutions were insignificant. Table 7 excludes loans on securities, loans to brokers and other banks, and money market investments which the Federal Reserve traditionally included in its reporting as bank loans. Table 7 includes as loans ‘‘real estate owned’’ (REO) which is real estate taken over by the lender as a result of foreclosure. Federal Reserve data did not treat REO as loans, but it became increasingly significant relative to loans for all types of intermediaries as the 1930s progressed because, as we shall see below, intermediaries’ holdings of

39,011 36,138 32,884 31,202 30,832 31,729 33,009 34,643

11,375 9,106 7,608 7,294 7,657 8,812 9,886 11,363

All Member Banks

29% 25% 23% 23% 25% 28% 30% 33%

% of Total

3,523 2,919 2,023 1,795 1,852 1,978 2,225 2,250

Nonmember Banks

9% 8% 6% 6% 6% 6% 7% 6%

% of Total

6,260 5,790 5,275 4,792 4,455 4,387 4,414 4,445

Savings & Loan Associations

16% 16% 16% 15% 14% 14% 13% 13%

% of Total

6,138 6,257 6,248 6,127 5,992 5,878 5,814 5,733

Mutual Savings Banks 16% 17% 19% 20% 19% 19% 18% 17%

% of Total

Loans by Financial Intermediaries ($Millions).

11,715 12,066 11,730 11,194 10,876 10,674 10,670 10,852

Life Insurance Companies

30% 33% 36% 36% 35% 34% 32% 31%

% of Total

Note: Loans exclude open-market paper, loans on securities, and loans to brokers and other banks. Loans include REO (real estate owned) and life insurance policy loans. Nonmember banks’ loans on this basis are estimated using the same ratio to ‘‘total loans’’ as member banks. Sources: Banking & Monetary Statistics 1914–1941, Federal Home Loan Bank Board Yearbook (1946, p. 7), Welfling (p. 87), Life Insurance Fact Book (1946, p. 42).

1931 1932 1933 1934 1935 1936 1937 1938

Total Loans

Table 7.

272 BARRIE A. WIGMORE


273

foreclosed real estate continued to increase throughout the decade. Based on this definition of loans, combined member and nonmember banks held 38 percent of all loans in 1931, the combination of savings and loans and mutual savings banks (whose savings and lending functions and nonprofit organization were similar) had 32 percent, and life insurance companies had 30 percent. The other important requirement for understanding remediation is clarifying that real estate loans rather than commercial or industrial loans were the critical variable in lending activity. Real estate loans (on residential, farm, and commercial properties) were $25 billion or 65 percent of these institutions’ total loans in 1931 and vitally more important to nonbanks than to banks.7 Table 8 outlines real estate loans by these intermediaries from 1931 to 1938. Real estate loans were 29 percent of member bank loans in 1931, 34 percent of nonmember bank loans, virtually 100 percent of savings and loan and mutual savings bank loans, and 71 percent of life insurance company loans (they also made significant policy loans). Within the markets for real estate loans there was considerable specialization. Savings and loans and mutual savings banks held 45 percent of 1–4 family mortgage loans. Life insurance companies held 22 percent of farm mortgage debt. Life insurance companies and mutual savings banks each held 33 percent of the mortgages on commercial property. Banks were not dominant lenders in any category. In both home and farm mortgage lending, ‘‘others’’ were important lenders, accounting for 36 percent and 49 percent, respectively, of loans in 1931. The ‘‘other’’ category included owner-financed sales and local money-lending as well as loans by trust departments of banks, other intermediaries such as mortgage companies and finance companies, and by endowments and charities (Federal Home Loan Bank Board, Statistical Summary, 1948, p. 14; Banking & Monetary Statistics 1914–1941, p. 8; Morton, 1956, p. 171). ‘‘Others’’ undoubtedly held a large share of commercial real estate loans as well. Commercial or business lending by banks in 1931 probably only accounted for 12 percent of all loans. The first opportunity to isolate commercial lending in bank loan data is in the fourth quarter of 1938 when all member banks began to report their loans in much greater detail (Banking & Monetary Statistics 1914–1941, p. 76, Table 1). Previously, commercial loans were lumped under ‘‘other’’ loans along with loans to private individuals, governments, institutions, and other borrowers. ‘‘Commercial’’ loans at the end of 1938 were 58 percent of ‘‘other’’ loans. Applying this ratio to member bank ‘‘other’’ loans in 1931 suggests that member bank commercial loans were 41 percent of all member bank loans,

39,011 36,138 32,884 31,202 30,832 31,729 33,009 34,643

25,209 24,596 22,976 21,797 21,212 21,126 21,250 21,788

Total Real Estate Loans

65 68 70 70 69 67 64 63

% of Total

3,250 3,131 2,634 2,587 2,651 2,772 2,890 3,186

29 34 35 35 35 31 29 28

Real Member Estate Banks Real % of Total Member Estate Loans Loans

1,215 1,158 857 755 778 826 861 961

NonMember Banks Real Estate Loans

34 40 42 42 42 42 39 43

Real Estate % of Total Nonmember Loans

6,260 5,790 5,275 4,792 4,455 4,387 4,414 4,445

Savings & Loan Real Estate Loans

100 100 100 100 100 100 100 100

Real Estate % of Total S&L Loans

6,138 6,257 6,248 6,127 5,992 5,878 5,814 5,733

100 100 100 100 100 100 100 100

8,346 8,260 7,961 7,536 7,336 7,263 7,271 7,463

71 68 68 67 67 68 68 69

Real Estate Life Mutual Real Estate Insurance % of Total % of Savings Life Banks Real Mutual Companies Insurance Real Savings Estate Loans Estate Loans Loans Loans

Real Estate Loans of Financial Intermediaries (Including REO) ($Millions).

Note: Loans exclude open-market paper, loans on securities, loans to brokers, loans to other banks, and all investments. Real estate loans include REO (real estate owned; see text). Nonmember banks’ real estate loans are estimated using the same ratio of real estate loans to total loans as country member banks. Sources: Banking & Monetary Statistics 1914–1941 (pp. 23, 101), Federal Home Loan Bank Board Yearbook (1946, p. 7), Welfling (p. 87), Life Insurance Fact Book (1946, p. 42).

1931 1932 1933 1934 1935 1936 1937 1938

Total Loans

Table 8.

274 BARRIE A. WIGMORE

275


or $4.5 billion, and only 12 percent of total loans by the four types of financial intermediaries. Applying the same 58 percent ratio to loans of nonmember banks would have increased this $2 billion, although given the rural nature of nonmember banks this would probably result in an overstatement. There appears to have been a misplaced emphasis on business credit as opposed to real estate lending in many studies analyzing the Great Depression.8 Other than railroads, utilities, and oil companies, businesses were not big borrowers until the latter third of the twentieth century. Although bank failures virtually disappeared after the Bank Holiday, the high level of distress in real estate lending for the remainder of the 1930s suggests that there were continuing problems for nonbanks. A good measure of financial stress is REO which is essentially real estate taken over in foreclosure and held by the lender. The difficulty in selling any property so acquired during the Depression resulted in it accumulating on lenders’ books throughout the 1930s. Table 9 outlines the growth of foreclosed real estate (REO) for the four types of intermediary. It was less of a problem for banks than other intermediaries. Member commercial banks’ REO peaked at 14 percent of total real estate loans including REO in 1935 and nonmember banks were probably about the same (based on country member banks). But the share for savings and loan quadrupled from 6 percent of loans in 1931 to 26 percent in 1936, which was especially important because real estate loans constituted over 85 percent of savings and loan assets. Mutual savings banks’ REO rates Table 9.

1931 1932 1933 1934 1935 1936 1937 1938

Foreclosed Property (REO) as Share of Total Real Estate Loans.

All Member Banks (%)

Country Member Banks (%)

Savings & Loans (%)

Mutual Savings Banks (%)

Life Insurance Companies (%)

7 9 10 12 14 13 12 11

8 10 11 13 15 14 13 13

6 11 16 23 26 26 23 20

4 6 8 11 13 16 16 16

8 11 16 22 27 29 28 27

Note: Ratios are REO divided by REOþmortgage loans. Sources: Life Insurance Fact Book (1946, pp. 37, 42), Federal Home Loan Bank Board Yearbook (1946, p. 7), Welfling (p. 87). Banking & Monetary Statistics 1914–1941 (pp. 74, 98–101).

276

BARRIE A. WIGMORE

also quadrupled from 4 percent to 16 percent of loans between 1931 and 1936. Their real estate loans amounted to 60 percent of assets. For the life insurance industry, REO rates rose from 8 percent in 1931 to 29 percent in 1936. Real estate loans were over 50 percent of life insurance companies’ assets net of policy-holder loans. Commercial real estate mortgages suffered particular distress and there was no remediation for them, as opposed to farm and home mortgages. Government assistance to commercial real estate remains a complicated, politically thorny issue to this day.9 There are varying estimates of the amount of commercial real estate debt in the 1930s, but in all cases these estimates are large. Goldsmith’s data indicate that commercial mortgages amounted to $24 billion in 1929.10 The Department of Commerce estimated they were $15 billion in 1930, but its subsequent data suggested $21 billion.11 Morton estimated that mortgages on ‘‘Income-Producing Properties’’ held by the same four intermediary groups studied in this paper were $10.3 billion in 1929 (Morton, 1956, p. 171). Morton’s number did not include approximately $6.2 billion in public real estate bonds that were included in Goldsmith’s estimate. Nor did Morton include ‘‘other’’ lenders which were surely significant since they accounted for over one-third of home and farm mortgage lending. In any case, commercial mortgages amounted to somewhat less than straight corporate debt of $30 billion, but at least as much as 1–4 family residential mortgage debt of $19.7 billion, and much more than farm debt of $9.6 billion. The extreme distress in commercial mortgages is illustrated in Chart 1 by year-end prices per $1,000 bond for an index, compiled by E. T. Watson Co., of 250 commercial mortgage bonds that were publicly traded.12 During the 1920s, there was a public market for commercial mortgage debt on rental apartment buildings, offices, industrial buildings, clubs, hotels, and movie theaters. The index covers numerous cities and types of properties and also retains defaulted bonds throughout the period. The average price for these bonds dropped from $970 in 1929 to $186 at the end of 1932, and barely recovered above $400 at any time in the rest of the 1930s. Foreclosure rates for commercial mortgages created between 1926 and 1929 ultimately ranged between 30 and 47 percent (Morton, 1956, p. 171; Saulnier, 1950, p. 84). Further analysis is needed to explore whether the continuing stress in commercial real estate had the same debilitating economic effects on economic recovery that Anari, Kolari, and Mason (2002) found from the slow resolution of failed banks and that Field (1992) emphasized with respect to failed residential real estate development. Commercial real estate construction was over 25 percent of private plant and equipment

277

Financial Remediation in the Great Depression and 2008–2009 $1,200 $1,000 $800 $600 $400 $200 $0 1929

Chart 1.

1930

1931

1932

1933

1934

1935

1936

1937

1938

Commercial Real Estate Bond Prices (per $1000 Bond).

expenditures in 1929. F. W. Dodge’s ‘‘Contracts Awarded for the Construction of Office and Loft Buildings in Thirty-Seven Eastern States’’ averaged approximately 54 million square feet in 1925–1929 but had recovered to only 8 million square feet in 1936–1938 (Survey of Current Business, February 1940, p. 9; Fisher, 1951, p. 151). Commercial real estate is an awkward fit for traditional theories of investment growth and macro public policy-making. Field (1992) has delineated the ‘‘hangover’’ from premature subdivision in the 1920s that inhibited the revival of residential real estate development in the Depression. Commercial real estate development is also subject to intense local regulation, has an unusually long economic life, tends to be very highly leveraged (due to our laws, customs, taxes, and institutional structures), is very sensitive to income taxes, and has exaggerated cycles.

3. REMEDIATION BENEFITS BY FINANCIAL SECTOR The benefits of Depression-era financial remediation were disproportionately received by banks relative to their share of loans. Table 10 illustrates the annual, cumulative remediation received by each type of intermediary from the various programs between 1932 and 1938. The banking industry received almost $7 billion, or 71 percent, of these benefits versus $1.2 billion received by savings and loans, $0.8 billion by mutual savings banks, and $0.9 billion by

278

BARRIE A. WIGMORE

Table 10.

1932 1933 1934 1935 1936 1937 1938

Remediation Funds Received by Intermediary (Cumulative) ($Millions).

Banks

% of Total

Life Insurance

Savings & Loans

Mutual Savings

Total

1,024 2,593 4,526 5,337 5,955 6,462 6,999

83 83 72 71 70 71 71

182 264 569 688 787 829 887

0 166 737 971 1,092 1,161 1,231

27 90 412 555 649 711 780

1,233 3,114 6,244 7,551 8,483 9,163 9,897

Note: See Tables in appendix for each industry. Sources: Annual Reports for RFC, FHA, Farm Credit Administration, and author’s calculations; Welfling (p. 91); Banking & Monetary Statistics 1914–1941; Spero (1939, pp. 36–41); Life Insurance Factbook (1946, p. 49).

life insurance companies. The remediation received by each type of intermediary from each of the programs is outlined in Appendix, Tables A1–A10. The difference between total remediation of almost $15 billion in Table 6 and the almost $10 billion in Table 10 was received by ‘‘others.’’ In both home and farm mortgage lending, ‘‘others’’ were important lenders, accounting for 36 percent and 48 percent, respectively, of loans in 1932. This ‘‘other’’ category included owner-financed sales and local moneylending, but it also included loans by trust departments of banks, mortgage companies, finance companies, endowments, and charities. The question arises why banks received 71 percent of remediation benefits when they held only 38 percent of the loans. Banks’ total assets were much greater than those of the other intermediaries, but this simply reflected a great deal of lending and investment within the financial sector that can be netted out. Their regulatory structure may have made banks more amenable to federal remediation efforts. The life insurance, savings and loan, and mutual savings bank industries had no federal regulators comparable to the Federal Reserve, the Comptroller of the Currency, and ultimately the FDIC. These other industries prized their state-level regulation jealously and focused their political efforts at that level. When FDR declared the Bank Holiday of 1933, it was the New York State Superintendent of Insurance who effectively created a holiday for the life insurance companies, requiring all New York registered life insurance companies (accounting for 80 percent of the life insurance written in the United States) to suspend


279

indefinitely all policy-holder loans, policy redemptions, cash dividends, and withdrawals of cash on deposit. Were bankers more aggressive because they were profit-seeking? Most savings and loans were not-for-profit organizations, mutual savings banks were just what their name implied, and many life insurance companies were mutual organizations. Were banks more effective politically? The American Bankers Association was a powerful lobby. Jesse Jones, the head of the RFC, was himself a Texas banker. Perhaps the fact that all banks were profit-seeking entities also had an impact on their political power. Did the fact that all corporations needed daily bank relationships, at a minimum for transferring funds, also give banks powerful political allies that the other financial industries lacked?

4. COMPARING 2008–2009 AND THE 1930S With the benefit of better understanding of the policy mistakes made during the Depression, the federal government acted quickly, massively, and more comprehensively in 2008–2009 to avert a similar result. There have been 13 remediation programs under the aegis of the Treasury, the FDIC, and the Federal Reserve, not counting Federal Reserve emergency openmarket policies. The various programs are outlined below. Several important programs have had explicit Congressional approval, but many have relied upon existing authority. The aggregate commitments or potential liabilities (worst case) exceeded $14 trillion, equal to 97 percent of GDP and 59 percent of private nonfinancial debt in 2008 – five times the proportionate scale of remediation in the Depression. The Federal Reserve provided additional, unprecedented, emergency open-market purchases equal to 18 percent of GDP and 11 percent of private, nonfinancial debt in 2008. Again, as in the Depression, the government commitments in the various programs are not all commensurable, but it is the essence of such a crisis that the risks are unknown, the losses unpredictable, and the ultimate downside is exactly what is trying to be avoided. Accordingly, the earlier procedure of aggregating the commitments will be followed here. This time, remediation commenced quickly in conjunction with the declines in financial markets and appears to have peaked within just 18 months of the initial program (July 2008) to allow federal trusteeship of the two government sponsored entities for housing financing – the Federal National Mortgage Association and the Federal Home Loan Mortgage Corporation.

280

BARRIE A. WIGMORE

4.1. Treasury Programs The Treasury initiated four remediation programs. The first and largest occurred on September 6, 2008, when both the Federal National Mortgage Association (FNMA) and the Federal Home Loan Mortgage Corporation (FHLMC) were put into conservatorship (bankruptcy) based on the authority given the Treasury under the Housing and Economic Recovery Act of July 30, 2008, to buy debt and equity securities of these Government Sponsored Entities. Treasury announced, among other things, that it was prepared to inject $100 billion of preferred equity (later raised to $200 billion) into each of these companies. The remedial obligation incurred with respect to these companies will be dealt with below, as well as for AIG. The Troubled Asset Rehabilitation Program (TARP) for up to $700 billion was authorized by the Emergency Economic Stabilization Act of October 2008. Commitments as of June 30, 2009, to financial intermediaries totaled $418 billion, to revitalizing securities markets $155 billion, to auto companies (excluding their finance subsidiaries) $61 billion, and to troubled home-owners and small businesses $65 billion (Special Inspector General for the Temporary Asset Recovery Program, Quarterly Report to the Congress, July 21, 2009, pp. 37–38). When it was authorized, the expressed intent was to support the markets for the complex securities created around real estate mortgages, but the program quickly morphed into providing most of its money to increase the equity of distressed banks and other financial institutions, similar to the RFC’s preferred stock program. All of the large banks receiving TARP preferred stock have either repaid or arranged to repay these advances as of the end of 2009. The Treasury has made substantial profits in every case because of the preferred dividends and common stock warrants that permitted the taxpayer to share in the gains of financial recovery. The ex ante risks of losses in the TARP were, however, at the time fearsome and incalculable. In August 2009, after considerable financial recovery, the Treasury estimated the possible losses at $340 billion. In December 2009, it reduced this estimate to $140 billion, with most of the financial losses attributable to AIG (a nonbank) (The New York Times, 12/9/09). Problems continue to escalate among regional banks, however, which could require additional TARP advances, and in 2009 the Treasury treated TARP as a revolving fund so commitments could aggregate more than $700 billion even though the Treasury said that it was prepared to reduce the fund to $550 billion.


281

The Treasury also guaranteed $3.2 trillion of money market mutual fund accounts in September 2008 following the bankruptcy of Lehman Brothers and the announcement that Reserve Funds, a prominent money market mutual fund organization, would not be able to repay par ($1) on its accounts because of its investments in short-term Lehman Brothers obligations. The Treasury’s guarantee was advanced without Congressional approval, utilizing $50 billion in assets in the Exchange Equalization Fund (not the first time this Fund has been used oddly). Treasury press releases made clear that all money market mutual fund accounts were fully guaranteed so it is unclear how the meager $50 billion in the Exchange Equalization Fund covered the $3.2 trillion in liabilities, except insofar as Treasury’s ex ante estimate of possible losses through this program was small. The program was terminated September 18, 2009, with no losses and a profit of $1.2 billion (U.S. Department of the Treasury, Press Room, Release TG-293, September 18, 2009). The Treasury also provided a direct line of equity credit of $30 billion to AIG in March 2009 at the worst point in the financial crisis as part of the Federal Reserve’s plan to pay off AIG’s derivative liabilities (AIG is dealt with more comprehensively below).

4.2. FDIC Programs The FDIC was involved in five financial remediation programs, none of which required Congressional approval because of existing FDIC authority. In September 2008, it temporarily increased the ceiling on deposit insurance to $250,000 (ultimately through 2013). No effort has been made here to quantify the remediation effects of this program for the same reasons as those mentioned above with respect to the initiation of deposit insurance in 1933. However, quantifying three other programs is quite straightforward. Shortly after the FDIC temporarily expanded its deposit guarantee to $250,000, it guaranteed 100 percent of the balances above $250,000 in what it defined as ‘‘noninterest bearing transactions accounts’’ – essentially business checking accounts. This amounted to $761 billion as of September 30, 2009. This program has been extended until June 30, 2010 (FDIC website). The initial guarantee was provided to banks for free, but the FDIC subsequently developed insurance fees (modified several times) in an attempt to make the program self-funding. No estimate of losses under this program is available. In the third quarter of 2009, the FDIC announced that its

282

BARRIE A. WIGMORE

Deposit Reserve Fund to cover losses was negative $8.2 billion (FDIC, ‘‘Quarterly Bank Profile, Deposit Insurance Fund Trends, Third Quarter 2009’’). The FDIC also offered to guarantee medium-term unsecured debt to facilitate refunding of existing unsecured debt (including commercial paper) by banks and bank-holding companies. The limit was 125 percent of their already outstanding unsecured debt as of September 30, 2008, and maturing in the next 9 months. Eighty-nine issuers (57 banks and 32 holding companies or nonbank affiliates of depository institutions) participated in this program with $307 billion outstanding as of the end of September 2009 which was only 50 percent of the possible amount. The program for further issues expired October 31, 2009 (FDIC, ‘‘Quarterly Bank Profile, Temporary Liquidity Guarantee Program, Third Quarter 2009’’). The fee structure for this guarantee program was modified several times as well. At a much smaller level, the FDIC participated with the Federal Reserve to the extent of $21 billion in loss-sharing agreements for securities portfolios owned by IndyMac, Citigroup, and Bank of America Corporation. In March 2009, the FDIC agreed to participate in the Treasury’s Legacy Loan Program to set up highly leveraged limited partnerships that would take troubled real estate loans off banks’ balance sheets with the government and private investors providing equity and the FDIC guaranteeing the debt. This program has not, by the end of 2009, advanced to a significant level (FDIC (75 Years) 2008 Annual Report, pp. 89, 100–101). Clearly, the FDIC’s guarantee of all business checking accounts ($761 billion) and its guarantees of unsecured debt refinancing for banks and bank-holding companies ($307 billion) were its critical programs.

4.3. Federal Reserve Remediation Programs The Federal Reserve moved very aggressively during the present crisis. It dramatically broadened both the financial assets and the intermediaries eligible for Federal Reserve loans. This was more akin to open-market operations (which will be dealt with below) than remediation, although the line between the two became blurred because of the risks the Federal Reserve assumed. The Federal Reserve’s remediation acts occurred when it accepted unconditional risk beyond open-market operations in specific rescue missions that threatened systemic disruption – Bear Stearns, AIG, Citigroup, and Bank of America Corporation. The Federal Reserve’s initial


283

assumption of a loss-sharing risk of $29 billion to facilitate J. P. MorganChase & Co.’s acquisition of Bear Stearns in May 2008 appears small in retrospect. It committed $85 billion to AIG in September 2008 following the Lehman Brothers bankruptcy, $245 billion to a loss-sharing pool with Citigroup in January 2009, and $97 billion to a similar pool with Bank of America Corporation at the same time (never finalized) (Federal Reserve, Annual Report 2008, pp. 8–10, 40). Each of these (but not Bear Stearns because of its modest size) will be dealt with separately below.

4.4. Remediation of FNMA, FHLMC, and AIG The largest risks were assumed in the federal remediation of FNMA, FHLMC, and AIG which justifies individual consideration of each. Under Generally Accepted Accounting Principles (GAAP) these three companies are now completely controlled by the U.S. Government. Government statements and actions have convinced the public that these entities will not be allowed to fail. FNMA declared in its 2008 10K that the federal conservator ‘‘y succeeded to all of the power and authority of the Board of Directors, Management, and the shareholders’’ and that the conservator ‘‘y indicated that our obligations will be paid in the normal course of business during the conservatorship’’ (FNMA, Annual Report on Form 10K for 2008, pp. 21–22). Similar language applied to FHLMC. This alone has allowed the rating agencies to maintain their AAA ratings of FNMA and FHLMC, a critical under-pinning for their continued financing and that of the $4.3 trillion of mortgages that they have guaranteed. Without similar support AIG would not be able to continue its insurance business or unwind its financial obligations in the ordinary course of business. Interpretation of the liabilities assumed by the government for these three entities is both complicated and controversial. The liabilities asserted for political consumption (Table 11) are only a modest fraction of the reality. The Treasury only committed to buy $100 billion of preferred stock from each of FNMA and FHLMC (subsequently raised to $200 billion each and then made unlimited on December 24, 2009), established lines of credit for them estimated at $265 billion by FNMA and at $250 billion for FHLMC,13 and undertook to buy an unannounced amount of agency-guaranteed mortgage-backed securities ($176 billion as of September 30, 2009). The Federal Reserve also committed to buy $100 billion of each of the agencies’ direct debt and $500 billion of outstanding mortgage-backed securities

284

Table 11.

BARRIE A. WIGMORE

Treasury Commitments to FNMA and FHLMC ($Billions) (as of September 30, 2009). Commitments

FNMA pfd FNMA line of credit FHLMC pfd FHLMC line of credit Treasury MBS holdings Totals

Drawdowns

200 265 200 250 176

61 0 52 0 176

1,091

289

Note: FNMA, Federal National Mortgage Association; FHLMC, Federal Home Loan Mortgage Corporation; MBS, Mortgage-backed securities. Sources: FNMA 10Q (September 30, 2009, pp. 75, 77). FHLMC 10Q (September 30, 2009, pp. 55–56, 111). Line of credit is estimated.

(subsequently raised to $1.25 trillion) which is discussed under the Federal Reserve’s open-market operations below. In reality, however, the government assumed enormous liabilities for FHMA and FHLMC which plausibly aggregate $7.2 trillion for debts, guarantees, and derivatives liabilities! The financial liabilities of these two institutions and AIG are outlined in Table 12. Their outstanding short-term and long-term debt of $1.6 trillion is easily understood. Less well understood is that the government became responsible for these two institutions’ guarantees of over $4.3 trillion of home mortgages which were the underlying assets for the troubled and hugely complicated mortgage-backed securities markets. No one would expect 100 percent losses on these guarantees, but how far home prices will fall and how homeowners will behave when their mortgages are ‘‘under water’’ was still unknown at the end of 2008. FNMA and FHLMC only provided $109 billion (2.5 percent) in reserves against losses from these guarantees. Under new accounting rules effective in 2010 they will have to consolidate these liabilities (FNMA 10Q 3rd Quarter, 2009, pp. 138–139; FHLMC 10Q 3rd Quarter, 2009, p. 116). Even less well understood are the derivatives liabilities of FNMA and FHLMC which are also outlined in Table 12. The AIG debacle made clear that even the architects of a derivatives portfolio might not understand its risks and that derivatives liabilities could be enormous. The accounting disclosures for derivatives do little to facilitate understanding. A simplified way of thinking of derivatives is this. Two parties make a contractual bet for an agreed period of time (months or years) on some publicly observable

285


Table 12. Liabilities of Government-Controlled Companies ($Billions). FNMA (9/30/09) Short-term debt Long-term debt Guarantees Notional derivative liabilities Total

241 562 2,441 589 3,833

Source: FNMA 10Q (September 30, 2009, pp. 73, 75, 83, 172). FHLMC (9/30/09) Short-term debt Long-term debt Guarantees Notional derivative liabilities Total

365 438 1,874 692 3,369

Source: FHLMC 10Q (September 30, 2009, pp. 56, 107, 116, 140, 144). AIG (12/31/08) Short-term debt Long-term debt Notional derivative liabilities Insurance liabilities Total

56 137 785 471 1,449

Source: AIG 10K (2008, pp. 193, 258, 262–263).

financial value (stocks, interest rates, foreign exchange rates, bond prices, financial indices, or even risk premiums) where the losing party posts acceptable collateral (defined) for its losses daily. FNMA and FHLMC have huge contracts along these lines that they basically reveal in two ways – the net amount of their losses which was $5 billion at the end of September 2009, and the ‘‘notional amount’’ or underlying principal value underlying their ‘‘bets’’ where they were winners (assets) or losers (liabilities) which was $2.5 trillion for both assets and liabilities combined. Since we have no way of calculating the potential losses under these derivatives contracts, how quickly disruptive demands for collateral might arise, or how seriously derivative assets would be dissipated in bankruptcy, Table 12 includes the notional value of derivative liabilities of $1.3 trillion as part of the liabilities assumed in federal financial remediation. This is just the liability side, because in bankruptcy the asset side can dissipate rapidly as was discovered with Lehman Brothers since the party owing money to the bankrupt entity can say the contract – which

286

BARRIE A. WIGMORE

should run for a further period – was cancelled by bankruptcy. There is also the problem of the bankrupt party losing the personnel to track down who owes it money. At the end of 2009, there was no way to tell what losses remediation of FNMA and FHLMC might ultimately entail, but their extreme leverage means that the losses have the potential to be dramatic. For example, at the end of 2007 FNMA had debt and mortgage guarantees in excess of $3 trillion backed by $47 billion in equity and reserves – only 1.5 percent of these liabilities. Subsequently, FNMA’s stock market value declined from a peak of $69 billion in August 2007 to virtually zero at the end of 2009. FNMA lost $59 billion in the first 9 months of 2009, its nonperforming loans quadrupled to 4.7 percent from the first quarter of 2008, and it expressed ‘‘y significant uncertainty regarding the full extent of future credit lossesy’’ (FNMA 10Q 3rd Quarter, 2009, pp. 4–5). The government kept FNMA and FHLMC in business so that their huge obligations could devolve in an orderly fashion, but it was possible from the vantage point of the end of 2009 that their losses could reach $1 trillion if home prices did not recover. The remediation of AIG represents a similar contrast between an apparent government commitment of $230 billion and a plausible aggregate liability of almost $1.5 trillion. Remediation of AIG began in September 2008 when it appeared that AIG might fail under the pressure of cascading investment losses and huge requirements to post collateral for derivatives transactions. The Federal Reserve bore the brunt of the original commitment to sustain AIG when it lent AIG $85 billion. This was reduced to $60 billion in November when the Treasury committed $40 billion to AIG for preferred stock under TARP and $30 billion under an equity line of credit and gained control of almost 80 percent of AIG’s common stock. The Federal Reserve further committed $50 billion to Maiden Lane II and III which were special-purpose corporations set up to dispose of AIG derivatives liabilities to counterparties. Counterparties were further allowed to keep $35 billion of collateral posted by AIG under these agreements. AIG was allowed to use $15 billion of Federal Reserve Commercial Paper Funding Facilities secured by mortgage-backed securities. These measures are outlined in Table 13. Subsequent restructurings of these measures were simply to improve the Federal Reserve’s creditor position. Thus, an aggregate government commitment of $230 billion with draw downs of $152 billion permitted AIG to continue as a going concern. As with FNMA and FHLMC, however, the true potential liabilities assumed by the government were much greater. The Treasury issued the

287


Table 13.

Government Commitments to AIG ($Billions) (as of September 30, 2009).

TARP Treasury equity line of credit Federal Reserve line of credit Maiden Lane II Maiden Lane III Maiden Lane III collateral ceded Federal Reserve CPFF Double-counting Totals

Commitments

Drawdowns

40 30 85 20 30 35 15 25

40 3 39 20 30 35 10 25

280

202

Note: CPFF, Commercial Paper Funding Facility. Sources: AIG 10Q (September 30, 2009, p. 100), AIG 10K (December 31, 2008, p. 251). AIG website accessed on 12/2/09.

following statement related to its help to AIG on March 2, 2009, at the depth of the financial crisis: The steps announced today provide tangible evidence of the U.S. government’s commitment to the orderly restructuring of AIG over time in the face of continuing market dislocations and economic deterioration. Orderly restructuring is essential to AIG’s repayment of the support it has received from U.S. taxpayers and to preserving financial stability. The U.S. government is committed to continuing to work with AIG to maintain its ability to meet its obligations as they come due (AIG, Annual Report on Form 10K for 2008, p. 202)

All of AIG’s financial obligations are outlined in Table 12. AIG’s shortand long-term debt was only $192 billion, but the notional value of its derivatives liabilities was approximately $800 billion. Its insurance liabilities were an additional $471 billion. While these insurance liabilities are not debts, they are contractual financial obligations to customers and if it appeared that AIG could not honor them its business would collapse. These insurance liabilities are as fundamental as deposit liabilities for a bank. The aggregate of all these liabilities was almost $1.5 trillion. Potential losses in the bailout of AIG are unusually uncertain because its problems may seriously affect the profits and values of its subsidiary insurance operations, irrespective of whether they are retained or sold.

288

BARRIE A. WIGMORE

4.5. Remediation of Citigroup and Bank of America Remediation of Citigroup and Bank of America received extensive media attention because the seriousness of their problems required repeated assistance, but the magnitude of the government’s commitment did not approach that to FHMA, FHLMC, or AIG, and the government never gained control of Citigroup or Bank of America. Table 14 outlines the remediation assistance to Citigroup. The government provided $45 billion in preferred stock in two stages to Citigroup during 2008, converted $25 billion of that to common equity in March 2009, and structured a loss-sharing arrangement on $301 billion of securities in February 2009. Citigroup made use of the liquidity assistance available from the FDIC to issue $32 billion of guaranteed unsecured notes to refund commercial paper and maturing debt. The Federal Reserve Bank of New York stated that it did not expect to have to advance funds under its liability of $245 billion in the Citigroup loss-sharing agreement (Federal Reserve, Annual Report 2008, p. 40), and Citigroup announced that it would redeem $20 billion of TARP preferred stock, raise new equity, facilitate the government’s sale of its $25 billion of common stock during 2010, and terminate the loss-sharing agreement on the $301 billion portfolio of troubled real estate and credit card assets (The New York Times, 12/14/09). Bank of America Corporation received similarly structured commitments from the government, although the loss-sharing agreement for an asset pool of $118 billion was never consummated, and it paid off the government in November 2009. Discussion of Citigroup and Bank of America should not omit mention of the remediation that did not happen – Lehman Brothers. There is no knowing whether Lehman Brothers would have passed through a relatively brief period of government assistance as the other banks did or ended up in government control like FNMA, FHLMC, and AIG. Table 14. Citigroup Remediation ($Billions) (as of February 27, 2009). Original pfd stock Second stage pfd stock Conversion to common equity Loss-sharing agreement FDIC guaranteed unsecured debt

20 25 25 301 32

Totals

403

Source: Citigroup 10K (2008, p. 2).

289


4.6. The Cumulative Impact of Contemporary Remediation The cumulative impact of the various current financial remediation programs outlined above was over $14 trillion and happened so quickly that it can be effectively aggregated at one time in Table 15. The amounts cited are the peak amounts for each program, some of which by the end of 2009 had already been reduced to zero. Others, such as commitments to FHMA and FHLMC, may continue to grow. As noted, the amounts cited do not all, on a dollar for dollar basis, represent the same liability for

Table 15. Government Agency or Recipient US Treasury US Treasury FDIC FDIC FDIC Federal Reserve FNMA FHLMC AIG Citigroup BofA Total Double–counting True total

Contemporary Financial Remediation. Program(s)

TARP Money Market Mutual Fund guarantees FDIC guarantees of bank debt FDIC transactions accounts guarantees FDIC loss sharing agreements Bear Stearns Acquisition Federal conservatorship Federal conservatorship (TARP, Treasury, Federal Reserve) (TARP, Treasury, Federal Reserve, FDIC) (TARP, Federal Reserve, FDIC)

Peak Amounts ($Billions)

Peak Date

769 3,200

6/30/2009 9/18/2008

350 684

May 2009 12/31/2008

21 29 3,833 3,369 1,467

12/31/2008 May 2008 9/30/2009 9/30/2009 March 2009

378

6/30/2009

164

March 2009

14,264 219 14,045

Note: Includes BofA loss-sharing agreement; does not include federal deposit insurance. Sources: TARP, Department of the Treasury, The Next Phase of Government Financial Stabilization and Rehabilitation Policies (September 2009); FDIC: FDIC Annual Report (2008) and FDIC website; Federal Reserve/Bear Stearns Acquisition, J. P. Morgan-Chase, Annual Report on Form 10K (2008); FNMA: FNMA Annual Report on Form 10K (2008) and Form 10Q for Third Quarter (2009); FHLMC: FHLMC Annual Report on Form 10K (2008) and Form 10Q for Third Quarter (2009); AIG: AIG Annual Report on Form 10K (2008) and Form 10Q for Third Quarter (2009); Citigroup: Citigroup Annual Report on Form 10K (2008) and Form 10Q for Third Quarter (2009); Bank of America: Annual Report on Form 10K (2008) and Form 10Q for Third Quarter (2009).

290

BARRIE A. WIGMORE

taxpayers. There were different risks, different loss assumptions, equity advances versus loans, outright loans versus guarantees of loans made by others, and undertakings that whole entities such as FNMA, FHLMC, and AIG would pay their bills versus formal guarantees. As during the 1930s, all of the programs had in common that there was a principal amount of capital that was advanced, lent, guaranteed, or at risk that in the worst case represented a federal loss or liability. All of this remediation was done in an atmosphere of tremendous uncertainty. There is some repetition in the numbers because FNMA, FHLMC, AIG, Citigroup, and Bank of America received funds under the broader generic programs. This repetition has been eliminated as ‘‘double-counting.’’ The resulting aggregate remediation of $14 trillion was equal to 97 percent of GDP and 54 percent of total private nonfinancial debt in 2008. The remediation of FNMA and FHLMC represents the greatest risk of loss to the taxpayer. Not only are their liabilities the largest among all of the financial intermediaries, but these two entities are also the most highly leveraged, and the housing assets underlying their liabilities are probably the least liquid. AIG also represents a great risk of loss, both because of the very complicated nature of its liabilities, but also because of the risk that its business will deteriorate seriously. By the end of 2009, all of the large banks but Citigroup had paid back their TARP funds with profits to the government, and Citigroup was in the process of paying back its preferred stock and facilitating disposition of the federal government’s common stock holdings. The Treasury’s guarantee of money market mutual fund assets had ended with a slight gain. There are no estimates of FDIC losses on its guarantees of business checking accounts and banks’ unsecured mediumterm debt, but it is to be expected that losses will arise in these programs as foreclosures of medium-sized banks increase.

4.7. Federal Reserve Open-Market Activities The Federal Reserve engaged in the most extensive open-market activities in history in response to the current crisis. Open-market policy was expanded from the traditional transactions in U.S. treasuries, the purpose of which is to control bank reserves, money supply, and short-term interest rates, to over $2.6 trillion in transactions designed to revive various securities markets that provide credit – particularly mortgage-backed securities, assetbacked securities (credit card receivables, insurance receivables, student loans, small business loans), and commercial paper.


291

The innovations began in March 2008 after several months of disruption in the mortgage securities markets. The Federal Reserve included $100 billion of mortgage-backed securities among open-market purchases and at the same time, discount facilities for the Primary Dealers in United States Treasuries were expanded whereby the Federal Reserve lent its Treasuries against mortgage-backed securities collateral (Term Securities Lending Facility or ‘‘TSLF’’). It also changed its rules to permit discounting of any securities eligible for general market collateral (Primary Dealer Credit Facility or ‘‘PDF’’). In September 2008, the Federal Reserve authorized purchases of $500 billion in mortgage-backed securities as part of sustaining FNMA and FHLMC after they were placed under federal conservatorship. This was increased to $1.25 trillion in March 2009 at the worst point in the crisis. Direct purchases of mortgage-backed securities ultimately amounted to almost $900 billion, but the TSLF and PDF represented another $650 billion of market support for mortgage-backed and asset-backed securities where the Federal Reserve accepted them as collateral. Loans under these facilities were still almost $1 trillion at the end of 2009 (http://federalreserve.gov, ‘‘Statistics & Historical Data,’’ ‘‘Factors Affecting Reserve Balances-H.4.1,’’ December 10, 2009). What losses the Federal Reserve might sustain in these securities remained to be seen. When Reserve Funds failed to sustain the $1 par value of its money market funds following the Lehman Brothers bankruptcy in September 2008, over 90 percent of all commercial paper outstanding ($1.7 trillion) was for financial or asset-backed, special-purpose issuers, demand for which evaporated. Asset-backed, special-purpose issuers accounted for over $700 billion and 44 percent of all commercial paper outstanding and had hundreds of pages of underlying legal documentation which few investors had read and which therefore required a high level of general confidence for acceptance. The reception for financial issuers was not much better (http:// www.federalreserve.gov, ‘‘Statistics & Historical Data,’’ ‘‘Commercial Paper’’). There was a coordinated federal effort to support commercial paper among the FDIC, which guaranteed refunding for banks and bankholding companies, the Treasury, which supported FNMA and FHLMC, and discount facilities set up by the Federal Reserve for Commercial Paper Funding (CPFF) and for Asset-Backed Commercial Paper for Money Market Mutual Funds (AMLF). These Federal Reserve facilities amounted to over $500 billion at year-end 2008, but were virtually zero at the end of 2009 (http://www.federalreserve.gov, ‘‘Statistics & Historical Data,’’ ‘‘Factors Affecting Reserve Balances-H.4.1,’’ December 10, 2009).

292

BARRIE A. WIGMORE

Table 16.

Federal Reserve Open-Market Initiatives 2007–2009 ($Billions).

Program Purchase of MBS CP funding facility (CPFF) Asset-backed CP for MMMF (AMLF) Central Bank swap facilities Term asset-backed lending (TALF) Term securities lending (TSLF) Primary dealer facility (PDCF) Total % 2008 GDP of 14,441 % 2008 private nonfinancial debt of 24,000

Date 9/30/2009 1/21/2009 10/08 12/08 9/30/2009 3/4/2009 10/1/2008

Peak Amount 883 351 150 583 46 493 147 2,653 18% 11%

Sources: Federal Reserve, Annual Report 2008, Treasury, The Next Phase of Government Financial Stabilization and Rehabilitation Policies, Federal Reserve, Statistical Release H.4.1, SIGTARP, Quarterly Report 2nd Q 2009.

The peak outstanding under each of the Federal Reserve’s emergency open-market programs are summarized in Table 16. When they were initiated, these specially authorized facilities blurred the line between open-market activities and remediation. The Federal Reserve did not have the time or experienced people to study the documents and evaluate the credits of these securities. They involved incredibly complex structures, underlying collateral that could not be accurately evaluated, and risks that were difficult to model and were often weakly understood even by those who originated them. The Federal Reserve bought on faith as part of its responsibility to sustain the banking system and the economy. These programs aggregated over $2.6 trillion, equivalent to 18 percent of 2008 GDP and 11 percent of 2008 private, nonfinancial debt. It must again be emphasized that the potential losses in these various programs are different and uncertain. Some of them will clearly have no losses. Others are a question mark. At the time they were initiated there was an atmosphere of acute risk and uncertainty and sorting out what programs would be cost-free and what would bear high costs was impossible. This aggregation of the Federal Reserve’s various innovations, unadjusted for risk, nevertheless provides a starting point for estimating the magnitude of the remediation effort.


293

5. COMPARING THE 1930S AND 2008–2009 The differences between financial remediation in the Depression and 2008–2009 are striking, in part because the Depression-era experience has been etched in the American consciousness as the ultimate economic disaster to be avoided at all costs and a cauldron of unexpected new social and economic policies. Presumably economic policy has also advanced in understanding. The Depression, as Bernanke is reputed to have declared, is the holy grail of economic theory. 1. The current remediation developed with remarkable speed in 2008–2009 compared with the Depression when three years of economic decline occurred before the RFC was even created and the peak in remediation was not reached until 1938. 2. The range and focus of programs differed markedly in the two periods. In the Depression, over 50 percent of remediation was for home and farm mortgage borrowers, and virtually none for the securities markets after the RFC was chastised in 1932 for helping refinance several railroads (the largest bond issuers). In the present crisis, 98 percent of remediation has gone to financial intermediaries and Federal Reserve support for the securities markets in its emergency open-market operations exceeded $2.6 trillion. 3. The one common feature has been the exclusion of any significant remediation for commercial real estate even though it accounted for approximately 25 percent of private capital spending in both periods. 4. The current financial remediation also extended far beyond the banking system which in the 1930s received the benefits of 71 percent of total remediation. Direct remediation for FNMA, FHLMC, and AIG alone (nonbanks) amounted to 62 percent of the total in 2009. The comparison is tendentious, of course, because securitization, derivatives, investment bankers, federal mortgage agencies, and securities markets generally play such a large role in current credit markets. Further research is necessary to establish the distribution of remediation among other insurance companies, investment bankers, and finance companies, as well as to determine the ultimate beneficiaries of remediation – foreigners and domestic institutions holding FHMA and FHLMC debt or mortgagebacked securities, counterparties to FNMA’s, FHLMC’s, and AIG’s derivative liabilities, holders of financial- and asset-backed commercial paper, investors in money market mutual funds, etc.

294

BARRIE A. WIGMORE

Table 17.

Remediation in 1938 and 2009 Compared ($Billions).

1938 aggregate remediation

2009 aggregate remediation 2009 Federal Reserve emergency open-market programs

% of 1931 GDP of $76 19% % of 2008 GDP of $14,441 97% 18%

% of 1931 Private Sector Debt of $148 10% % of 2008 private non-final Debt of $24,000 59% 11%

Sources: Bureau of the Census, ‘‘Historical Statistics of the United States’’ (pp. 224, 989); Federal Reserve, Flow of Funds Accounts.

5. The current remediation has been approximately five times as large proportionately compared with the 1930s, equal to 97 percent of GDP or 54 percent of private nonfinancial debt (six times if Federal Reserve emergency open-market programs are included). Table 17 summarizes data already cited in the text that compare peak remediation in 1938 with 1931 GDP and total private sector debt, and both peak remediation in the current crisis and Federal Reserve emergency open-market programs with 2008 GDP and private sector, nonfinancial debt.14 Other comparisons are available, of course, but the basic purpose of remediation is to sustain or grow private sector nonfinancial credit in order to sustain or grow GDP. 6. The potential for loss in the current remediation is greater, both because of its greater magnitude and because of the very high leverage in FNMA and FHLMC.

NOTES 1. The Federal Farm Mortgage Corporation was the financing arm of the Farm Credit Administration, while the Federal Land Banks were the lending arms. The Federal Farm Mortgage Corporation sold public debt with a federal government guarantee that was treated as a contingent liability by the government. 2. Federal Home Loan Bank Board, Fourth Annual Report, 1937, pp. 29–32, 41. This excludes individuals who provided $6.9 billion of noninstitutional 1–4 family urban home financing, equal to 39 percent of all 1–4 family urban home mortgage financing (Federal Home Loan Bank Board, Statistical Summary 1948, Table 10, p. 14). 3. Including interbank and local government demand deposits, but excluding federal government deposits (Board of Governors of the Federal Reserve System, Annual Report 1934, p. 148). 4. I have excluded New York City banks from this data series because they made few business loans (only 7 percent of total assets at the end of 1932) (Board of


295

Governors of the Federal Reserve System, Banking & Monetary Statistics 1914–1941, p. 83). Data are for Central Reserve City member banks in New York City. Business loans are defined as 77 percent of ‘‘Other loans’’ (1938 ratio). 5. Earlier historians such as Schlesinger (1958) and Friedel (1973, pp. 229–236) emphasized deposit insurance, but even financially sophisticated recent historians such as Meltzer (2003, pp. 434–435) and Black (2003, pp. 276–279, 349) have given it credit. 6. Board of Governors of the Federal Reserve System, Federal Reserve Annual Report (1934, p. 20). Nonmember banks (not included in this calculation) were approximately 18.5 percent of all deposits. 7. Some economists may not consider farm land ‘‘real estate,’’ but it was so treated by the Federal Reserve and life insurance companies. Farmland was treated as real estate from the beginning of the Republic when westward expansion schemes bankrupted Robert Morris to homestead financing after the Civil War to incentives for railroad development to agricultural expansion in World War I to land tracts for western home-building in our time. 8. Although Bernanke (1983) and Calomiris and Mason (2003a) appear to have resolved much of the debate about financial factors in the Great Depression, the only credit data in their regressions was for businesses. 9. It is unclear what remediation is suitable for commercial real estate. The federal government is unlikely to guarantee commercial mortgages, subsidize commercial tenants, or provide equity to developers. One possibility might be to selectively provide pooled long-term refinancing in cases that appear likely, in the opinion of some outside authority, to be economically viable given long-term recovery. 10. Goldsmith (1954, pp. 117, 119). Bonds ¼ $6.2 billion, nonresidential mortgages ¼ $9.5 billion, and residential mortgages for greater than 4 units ¼ $8.1 billion. The total of $24 billion may double count a portion of the $6.2 billion of bonds. 11. US Department of Commerce, Statistical Abstract (1936, p. 275). Later issues suggest that this excludes $6 billion in public real estate bonds (which could include some double-counting). 12. E.T. Watson Co. is sometimes referred to as Burr & Co. which took it over in 1934. Their index covers 1932–1938 and was reported monthly in The New York Times. I have used my own estimates for the end of 1929–1931 based on irregularly reported prices in The New York Times and The Annalist for 44 bonds in 1929, 135 bonds in 1930, and 55 bonds in 1931. 13. Federal National Mortgage Association, 10Q (September 2009, p. 77). I estimated for FHLMC based on relative size. 14. Some might think 1929 was a better base year, as the peak of the prior expansion, but remediation did not begin until 1932.

ACKNOWLEDGMENTS The author is indebted to Peter Temin (MIT) for repeated assistance and encouragement over many years, and to Alexander Field (editor) for suggestions, questions, assistance, time, and patience beyond the call of duty in the preparation of this paper.

296

BARRIE A. WIGMORE

REFERENCES 1930s: Anari, A., Kolari, J., & Mason, J. R. (2002). Bank asset liquidation and the propagation of the great depression. Wharton Financial Institutions Center Working Paper no. 02–35, August. Bernanke, B. (1983). Non-monetary effects of the financial crisis in the propagation of the great depression. American Economic Review, 73, 257–276. Black, C. (2003). Franklin Delano Roosevelt, champion of freedom. New York: Perseus Books Group. Board of Governors of the Federal Reserve System. Annual Reports (1932–1938). Washington, DC: Government Printing Office. Board of Governors of the Federal Reserve System. (1942). Banking & monetary statistics 1914–1941. Washington, DC: Government Printing Office. Board of Governors of the Federal Reserve System. (1977). Agricultural finance databook. Washington, DC: Government Printing Office. Calomiris, C. W. (1990). Is deposit insurance necessary? A historical perspective. Journal of Economic History, 50(June), 283–295. Calomiris, C. W., & White, E. (1994). The origins of federal deposit insurance. In: C. Goldin & G. Libecap (Eds), The regulated economy: A historical approach to political economy (pp. 145–188). Chicago, IL: University of Chicago Press. Calomiris, C. W., & Mason, J. R. (1997). Contagion and bank failures during the great depression: The June 1932 Chicago banking panic. American Economic Review, 87(December), 863–883. Calomiris, C. W., & Mason, J. R. (2003a). Consequences of bank distress during the great depression. The American Economic Review, 93, 937–947. Calomiris, C. W., & Mason, J. R. (2003b). Fundamentals, panics, and bank distress during the depression. The American Economic Review, 93, 1615–1647. Farm Credit Administration. Annual Reports (1933–1939). Washington, DC: Government Printing Office. Federal Deposit Insurance Corporation. Annual Reports (1933–1938). Washington, DC: Government Printing Office. Federal Home Loan Bank Board. Annual Reports (1934–1940). Washington, DC: Government Printing Office. Federal Home Loan Bank Board. (1949). Statistical summary, 1948. Washington, DC: Government Printing Office. Federal Housing Administration. Annual Reports (1934–1940). Washington, DC: Government Printing Office. Field, A. J. (1992). Uncontrolled land development and the duration of the depression in the United States. Journal of Economic History, 52, 785–805. Fisher, E. M. (1951). Urban real estate markets: Characteristics and financing. New York: NBER. Friedel, F. (1973). FDR – Launching the new deal. Boston, MA: Little, Brown and Company. Friedman, M., & Schwartz, A. J. (1963). A monetary history of the United States, 1867–1960 (Paperback edn). Princeton, NJ: Princeton University Press. Goldsmith, R. W. (1954). The share of financial intermediaries in national wealth and national assets, 1900–1949. NBER Occasional Paper no. 42, New York: NBER.


297

Harriss, C. L. (1951). History and policies of the home owners loan corporation. New York: NBER Financial Research Program Studies in Urban Mortgage Financing. Institute of Life Insurance. (1947). Life insurance fact book, ‘‘1946’’. New York: Institute of Life Insurance. Jones, J. H., with Edward Angly. (1951). Fifty billion dollars my thirteen years with the RFC (1932–1945). New York: The Macmillan Company. Jones, L. A., & Durand, D. (1954). Mortgage lending experience in agriculture. Princeton, NJ: Princeton University Press. Lintner, J. (1948). Mutual savings banks in the savings and mortgage markets. Cambridge: Harvard University Press. Meltzer, A. H. (2003). A history of the federal reserve, volume 1: 1913–1951. Chicago, IL: University of Chicago Press. Morton, J. E. (1956). Urban mortgage lending: Comparative markets and experience. Princeton, NJ: Princeton University Press. O’Grada, C., & White, E. N. (2002). Who panics during panics? Evidence from a nineteenth century savings bank. NBER Working Paper no. 8856. NBER, Cambridge, MA. Olson, J. S. (1988). Saving capitalism: The reconstruction finance corporation and the new deal 1933–1940. Princeton, NJ: Princeton University Press. Reconstruction Finance Corporation (various dates 1933–1938). Report of the Reconstruction Finance Corporation (title varies). Washington, DC: Government Printing Office. Saulnier, R. J. (1950). Urban mortgage lending by life insurance companies. New York: NBER Research Program Studies in Urban Mortgage Financing. Schlesinger, A. M., Jr. (1958). The coming of the new deal 1933–1935. Boston, MA: Houghton Miflin Company. Spero, H. (1939). Reconstruction finance corporation loans to the railroads 1932–1937. New York: Bankers Publishing Company. The New York Times, various issues, New York: The New York Times Company. US Department of Commerce, Bureau of Foreign and Domestic Commerce. (various years, 1930–1941). Statistical Abstract of the United States.Washington, DC: Government Printing Office. US Department of Commerce, Bureau of the Census. (1940). Survey of current business (February). Washington, DC: Government Printing Office. US Department of Commerce, Bureau of the Census. (1975). Historical statistics of the United States colonial times to 1970. Washington, DC: Government Printing Office. Welfling, W. (1968). Mutual savings banks, the third century. Cleveland, OH: The Press of Case Western Reserve University. White, L. J. (1991). The S&L debacle. New York: Oxford University Press. White, E. (1998). The legacy of deposit insurance. In: M. Bordo, et al. (Eds), The defining moment, the great depression and the American economy in the twentieth century (pp. 87–124). Chicago, IL: University of Chicago Press. Wigmore, B. A. (1985). The crash and its aftermath, a history of securities markets in the United States, 1929–1933. Westport, CT: Greenwood Press. Wigmore, B. A. (1987). Was the bank holiday of 1933 caused by a run on the dollar? Journal of Economic History, 47, 739–755. Woodruff, A. M., Jr. (1937). Farm mortgage loans of life insurance companies. New Haven, CT: Yale University Press.

298

BARRIE A. WIGMORE

2008–2009: AIG. AIG. Bank Bank

Annual Report on Form 10K for 2008, New York, NY. Quarterly Report on Form 10Q. (September 30, 2009), New York, NY. of America Corporation. Annual Report on Form 10K for 2008, Charlotte, NC. of America Corporation. Quarterly Report on Form 10Q. (September 30, 2009), Charlotte, NC. Board of Governors of the Federal Reserve System. (2009). Annual Report 2008. Washington, DC: Board of Governors of the Federal Reserve System. Board of Governors of the Federal Reserve System. (September 2009). Flow of Funds Accounts of the United States. Washington, DC. Citigroup Inc. Annual Report on Form 10K for 2008, New York, NY. Federal Deposit Insurance Corporation. (2009). FDIC (75 Years) 2008 Annual Report. Washington, DC: Federal Deposit Insurance Corporation. Federal Home Loan Mortgage Corporation. Annual Report on Form 10K for 2008. Washington, DC. Federal Home Loan Mortgage Corporation. Quarterly Report on Form 10Q. (September 30, 2009). Washington, DC. Federal National Mortgage Association. Annual Report on Form 10K for 2008. Washington, DC. Federal National Mortgage Association. Quarterly Report on Form 10Q. (September 30, 2009). Washington, DC. http://www.fdic.gov, ‘‘Industry Analyses’’, ‘‘Quarterly Banking Profile’’ (third quarter 2009). http://www.federalreserve.gov, ‘‘Statistics & Historical Data,’’ ‘‘Commercial Paper’’ (December 16, 2009). http://www.federalreserve.gov, ‘‘Statistics & Historical Data’’, ‘‘Factors Affecting Reserve Balances-H.4.1’’ (various dates). J. P. Morgan-Chase & Co. Annual Report on Form 10K for 2008, New York, NY. Special Inspector General for the Temporary Asset Recovery Program. (2009). Report to Congress (July 2009). Washington, DC. The New York Times. (2009). New York Times, December 9. U.S. Department of the Treasury. (2009). The next phase of government financial stabilization and rehabilitation policies (September 2009). Washington, DC. U.S. Department of the Treasury, Press Room (various dates). Washington, DC.

APPENDIX. DERIVATION OF TABLE 10 DATA The published information available on how remediation funds were spread among the various financial intermediaries ranges from detailed to unavailable and has therefore required various assumptions to assemble


299

Table 10. Tables A1–A4 detail the annual benefits from each of the remediation programs for banks, savings and loans, mutual savings banks, and insurance companies, respectively. The data for beneficiaries of RFC loans or equity advances are available in RFC published reports, but there are no data with respect to beneficiaries of its aid to refund railroad bonds and loans. Spero (1939) documented that only 60 percent of RFC loans to railroads was for financial purposes so I have applied that ratio to RFC railroad loans to establish the amount each year for financial remediation in Table A5 (Spero, 1939, pp. 36–41). I have been forced to estimate the share of that 60 percent allocation that went to each intermediary on the basis of their relative holdings of railroad bonds and loans in 1930, as outlined in Table A6. This gave banks the benefit of 38 percent of RFC loans to railroads each year, savings and loans zero, mutual savings banks 13 percent, and life insurance companies 49 percent. The Farm Credit Administration published detailed data in its annual reports on the intermediaries benefiting from farm mortgage refinancing by the Federal Land Banks and the Land Banks Commissioner. These data are outlined in Table A7. The HOLC published various intermediaries’ aggregate (rather than annual) shares of its refinancing activities over the 1933–1938 period – banks, 17 percent; savings and loans, 25 percent; mutual savings banks, 13 percent; and life insurance companies, 5 percent (Welfling, 1968, p. 92 and Report of the Federal Home Loan Bank Board, 1939, Chart XVII on p. 46). I have applied these shares each year throughout the period to annual HOLC refinancing in Table A8. Allocating the benefits of FHA mortgage refunding activity among financial intermediaries is more difficult. The FHA only published these data for two years, apparently based on existing versus new home financing. I have accepted this interpretation (which I think is too narrow) to calculate FHA refinancing each year from 1935 to 1938 in Table A9 and allocated benefits among intermediaries based on their shares of the total FHA program each year from 1934 to 1938. The annual benefit to savings and loans and mutual savings banks of eliminating interest on demand deposits, outlined in Table A10, has been calculated on the same 1 percent rate assumption as for banks and the assumption that these intermediaries’ demand deposits bore the same ratio to total deposits as did country member banks.

Table A1.

1932 1933 1934 1935 1936 1937 1938

RFC Equity

RFC Loans to Banks

0 497 1,154 1,237 1,245 1,262 1,321

948 1,794 2,220 2,328 2,394 2,438 2,458

Remediation Funds Received by Banks (Cumulative) ($Millions).

FHA HOLC Farm Refinancing Refinancing Mortgage 17% Share Refinancing

18 375 494 528 536 549

144 307 427 539

RFC Rail Loans 38% Share 76 93 103 111 141 145 168

74 345 439 461 472 481

1% Interest Subsidy

Bank Total

1,024 2,593 4,526 5,337 5,955 6,462 6,999

118 329 584 879 1,182 1,482

Note: Percentages indicate share of program over period to 1938. Sources: Tables 1, 4, A5, A7–A9.

Table A2. RFC

1932 1933 1934 1935 1936 1937 1938

122 143 145 146 148 154

Remediation Funds Received by Savings & Loans (Cumulative) ($Millions).

FHA Refinancing

33 79 110 131

HOLC Farm Refinancing Mortgage 25% Share Refinancing

26 551 726 776 788 807

0 0 0 0 0 0

RFC Rail Loans 0% Share 0 0 0 0 0 0 0

1% Savings & Interest Loan Subsidy Total

18 43 67 91 115 139

0 166 737 971 1,092 1,161 1,231

Note: Percentages indicate share of program over period to 1938. Sources: Tables 1, A6–A10.

Table A3. RFC

1932 1933 1934 1935 1936 1937 1938

0 0 0 0 0 0 0

Remediation Funds Received by Mutual Savings Banks (Cumulative) ($Millions). FHA Refinancing

4 16 22 26

HOLC Farm Rail RFC Refinancing Mortgage Loans 13% 13% Share Refinancing Share

1% Interest Subsidy

Mutual Savings Total

0 0 0 0 0 0 0

44 88 134 180 227 274

27 90 412 555 649 711 780

13 286 377 403 410 420

27 33 37 40 50 52 60


301


Table A4. RFC

1932 1933 1934 1935 1936 1937 1938

Remediation Funds Received by Life Insurance Companies (Cumulative) ($Millions). FHA HOLC Farm Mortgage Refinancing Refinancing 5% Refinancing Share

83 116 133 133 137 137 139

5 110 145 155 158 161

14 34 59 76

RFC Rail Loans 49% Share

Life Insurance Total

99 121 135 145 183 189 219

182 264 569 688 787 829 887

22 191 250 277 286 291


Table A5.

RFC Railroad Financing Beneficiaries (Annual) ($Millions).

RFC Railroad Loans

Banks Savings & Loans Mutual Savings 38% Share 0% Share Banks 13% Share

Life Insurance Companies 49% Share

1932 1933 1934 1935 1936 1937 1938

202 45 28 22 78 12 61

76 17 10 8 29 5 23

0 0 0 0 0 0 0

27 6 4 3 10 2 8

99 22 14 11 38 6 30

Totals

448

168

0

60

220

Sources: Tables 1 and A6 for percentage shares.

Table A6. All Member Banks 983 17%

Railroad Bonds and Loans in 1930 ($Millions).

Nonmember Banks

Bank Loans

749 13%

500 8%

Savings & Mutual Savings Loans Banks 0 0%

797 13%

Life Insurance Companies

Total

2,908 49%

5,937 100%

Note: Nonmember banks and mutual savings banks are assumed to hold railroad bonds in the same proportions as country banks – 25% of all investments. Sources: Life Insurance Fact Book (1946, p. 49), ‘‘Banking & Monetary Statistics 1914–1941’’ (pp. 23, 77, 102), author’s estimate for bank loans.

302

BARRIE A. WIGMORE

Table A7.

1931 1932 1933 1934 1935 1936 1937 1938

Federal Farm Mortgage Loan Beneficiaries ($Millions).

Farm Mortgage Loans Advanced (Cumulative)

Refinancing (Increment) (%)

Refinancing (Increment) $Millions

Banks Share

Life Insurance Share

Others Share

42 70 292 1,576 2,021 2,207 2,310 2,391

90 88 88 78 69 70

200 1,130 392 145 71 57

74 271 94 22 11 10

22 169 59 27 9 5

86 429 149 54 26 25

1,995

482

291

769

Note: refinancing in 1935 was not reported. I have used the 1934 percentage. Sources: Farm Credit Administration Annual Report (1933, pp. 82–83; Ibid. 1934, pp. 96, 99; 1936, pp. 118, 122; 1937, p. 120; 1938, pp. 123–124, 128).

Table A8.

HOLC Refinancing Beneficiaries (Annual) ($Millions).

HOLC Savings & Loans Banks 17% Mutual Savings Refinancing 25% Share Share Banks 13% Share

Life Insurance Companies 5% Share

1933 1934 1935 1936 1937 1938

103 2,100 700 200 50 75

26 525 175 50 13 19

18 357 119 34 9 13

13 273 91 26 7 10

5 105 35 10 3 4

Totals

3,228

808

550

420

162

Sources: Welfling (1968, p. 92) and Annual Report of the FHLBB (1939, Chart XVII on p. 46).

Table A9.

FHA Refinancing Beneficiaries (Annual) ($Millions).

FHA Existing Homes Financing 1935 1936 1937 1938 Totals

64% 52% 45% 31%

FHA Savings & Refinancing Loans Share

Banks Share

Mutual Life Insurance Savings Companies Share Banks Share

205 287 221 212

33 46 31 21

144 164 119 113

4 11 7 4

14 20 24 17

925

131

540

26

75

Sources: Federal Housing Administration, Annual Report (1935, p. 18; 1936, p. 28; 1937, p. 37; 1938, pp. 66, 77).

303


Table A10. Savings Institutions’ Benefits from Interest-Free Deposits ($Millions). Savings & Assumed S&L Loan Demand Deposits Deposits 46%

1933 1934 1935 1936 1937 1938

5,896 5,523 5,220 5,165 5,178 5,190

2,712 2,541 2,401 2,376 2,382 2,387

1% Interest Cost on Demand Deposits 18 25 24 24 24 24 139

Mutual Assumed Mutual 1% Interest Savings Demand Cost on Bank Deposits 46% Demand Deposits Deposits 9,488 9,738 9,871 10,056 10,170 10,278

4,364 4,479 4,541 4,626 4,678 4,728

44 45 45 46 47 47 274

Note: Assumes the same proportion of demand deposits as country member banks and no interest on demand deposits for two-thirds of 1933. Sources: FHLBB Statistical Summary, 1948 (p. 7), Banking & Monetary Statistics 1914–1941 (p. 23).

TRENDS IN FOOD CONSUMPTION IN THE UNITED STATES, 1840–1910 AN EXPERIMENT IN ECONOMETRICAL HISTORY William N. Parkerw This paper is a progress report on some historical research into food consumption trends in the United States and their implication for American economic growth. The time period covered runs at a maximum from the first Census of Agriculture in 1940 to the beginning of what an historian at least must consider the reasonably reliable national estimates by the Department of Agriculture in 1909 (USDA, 1953). The period is thus determined not primarily by any historical or theoretical considerations, but simply by the fact that it lies in the pleasant twilight between archeological darkness and the modern statistical glare. For this reason the research must be called ‘‘historical’’ and does not perhaps deserve (if I may be permitted a Veblenian adjective) the more ‘‘honorific’’ title: ‘‘econometrical.’’1 Statistical records and economic theories, however, like certain antibodies in the blood, appear to grow in society with the progress of the diseases they combat. Where men are not consciously interested in ‘‘economizing,’’ they are not often interested in counting, so that in the study of preindustrial societies where our statistical records are most deficient, our theoretical tools are also the most blunt. As we glide forward in the study of the


305

306

WILLIAM N. PARKER

industrialization process, economic theory fits us with a pair of skates that fit increasingly well and grow – with the growing dominance of economic motivation in society – increasingly sharp. But the absence of quantitative data – an absence which did so much to produce those heroic feats of speculative theory in the nineteenth century – leaves us gliding on ice which has only very slightly thickened. Only as a congealed twentieth century capitalism coats economic theory with incrustations of quantitative data frozen solid by government regulatory bureaux can the economist skates begin to feel safe. But by this time he often appears to have skated off the ice entirely and is left wobbling on his ankles on the dry land bargaining with a sociologist for a pair of roller skates. If then we are to examine the transition from an agricultural to an industrial economy in the United States and elsewhere, we must necessarily be ready to skate on some very thin ice. But the theorist and the statistician must stand beside the historian – as the three, performing a splendid Schumpeterian waltz, keep one another from falling through the ice into the pond. I wish to begin with a brief summary of the main trends in per capita food consumption in the United States between 1840 and 1910, as the present research has uncovered them. Then I should like to sketch the main lines of historical investigation into the implications and explanations of these trends that are being pursued. I will close with a few methodological remarks.

THE MAIN TRENDS IN PER CAPITA CONSUMPTION The main trends in food consumption per capita unearthed so far from the Census and other material may be briefly summarized. Only for the meats – pork, beef, and veal – have the estimates been carried back to 1840. It seems not impossible that wheat flour may be extended before 1880, and possibly even the series for dairy products before 1870. In these cases, however, prudence has to date prevailed. To one familiar with the evidence of the European experience, the most striking fact uncovered by the investigation is that per capita consumption of meat in the United States failed to rise as per capita income grew during the nineteenth century. This fact is attributable largely to the movement of per capita pork consumption, which – as best we have been able to measure

Trends in Food Consumption in the United States

307

it – appears to have declined by at least 30 percent between 1840 and 1910. The trend of per capita beef and veal consumption is distorted by the production boom of the 1880s, but a rise of 20 percent over the whole period is, if anything, perhaps too high. These trends, however, are based on the carcass weights; taking into account the difference in the retail equivalent of the carcass weights of the two meats, we may estimate a decline in the per capita consumption of about 15 percent for pork, beef, and veal taken together between the 1840s and 1900s.2 Some rise in poultry consumption occurred at least after 1880, and a rise in lamb and mutton consumption is possible. These could at most balance out the decline in the major meats to produce a roughly stable per capita trend for the period for all meats and poultry taken together. Against this conclusion, which has some statistical basis, one incommensurable element must be set – the indeterminate, but certainly declining, amounts of wild game consumed. A second – only slightly less surprising – fact is the absolute decline after 1880 (when milling data first became available) in the per capita consumption of wheat flour and cornmeal. The decline in wheat flour consumption was moderate – no more than 10 percent, but the fall in the two types of flour taken together was nearly 40 percent, in the 30-year period – nearly all of this being due to the decline in cornmeal consumption per capita. Consumption of the other grains taken together – rye, buckwheat, rice – was not large enough in 1910 that any increase in their consumption, even from zero, could have balanced the decline in cornmeal. The trend of per capita consumption of potatoes, both Irish and sweet, was very stable. Against these declining trends – meat and meal – must be set several sharply rising ones. Per capita consumption of dairy products probably rose between 1870 and 1900 – with a rise in butter and cheese consumption preceding a rise in the consumption of fluid milk. In fluid milk equivalent, the rise may have amounted to 15–20 percent although estimates of human consumption of milk on farms are wholly lacking. Per capita consumption of sweetening in the form of sugar and syrups at least doubled during the period from 1870 to 1910. A fall of 50 percent or over in consumption of molasses and syrups of all kinds – cane, sorghum, and maple – was far overbalanced by the steady and rapid rise in the consumption of refined sugar. Beginning about 1890, consumption of citrus fruits shows an extremely rapid increase; good data are not available for the other fruits, but for them a fairly stable consumption per capita appears not unlikely. No adequate data are available on vegetables other than potatoes.

308

WILLIAM N. PARKER

The per capita trends examined so far may be summarized as follows: Declining

Steady

Rising

Pork Cornmeal Molasses

Beef Wheat flour Potatoes

Sugar Dairy products Rice Poultry Eggs Citrus fruit

It is reassuring to observe that, at the points where the two studies overlap, these conclusions match closely those arrived at by Holbrook Working in a study done in 1925 (Working, 1926). The estimates for flour and cornmeal were derived by Working’s methods from the same data. Working, however, was primarily interested in flour consumption; only for flour, cornmeal, and sugar did he make estimates going back more than a few years before 1910. His estimate of meat consumption, which shows a per capita decline to 1917 begins with 1907, and for this and the other commodities he was not able to draw on the various revised production series published since 1924 by the Department of Agriculture. The study done in 1930 in the Department of Commerce by Montgomery and Kardell (1930) does not carry meats back before 1899, and several of its estimates – notably those for potatoes and eggs – appear questionable, either because of recent revisions of the production data or because of inadequate allowances for disappearance before consumption. The important study of the production data by Strauss and Bean (1940) is directed to deriving estimates of farm income and many of its methods of estimation or interpolation were found unacceptable in deriving a consumption series.

METHODS OF ESTIMATION OF PORK CONSUMPTION The trend in pork consumption which produced the most striking results in the present study also offered the most interesting problems of estimation. It may be of some interest, therefore, to indicate how the conclusion of a rather sharply declining trend was arrived at. (The essential data and sources are given in Tables 1–4.) The problem had two parts: (1) to estimate the trend of inventories from the Census count of swine on farms,

309


Table 1.

1840 1850 1860 1870 1880 1890 1900

Estimates of June 1 Hog Inventories, Cycle Averages, 1840–1900a.

January 1 Inventory (BAE Est.)

Cycle Average of January 1 Inventories

June 1 Inventory (Census)

Estimated June 1 Inventory-Cycle Average

(1)

(2)

(3)

(4)

26,301 30,354 33,513

26,301 30,354 33,513 37,832 48,087 53,895 62,661

33,781 44,327 48,130 51,055

36,037 42,841 45,171 20,887

49,773 57,427 62,868

Source: Column 1: United States Department of Agriculture (1938). Column 2: Ibid. Years chosen for cycle or cycle phase around the Census year as follows: 1867–1874, 1877–1883, 1888–1893, 1898–1902. Column 3: Decennial Censuses of Agriculture, figure for 1870 omitted because of known undercounting. Column 4: 1840, 1850, 1860, same as Column 3 (see text), 1870, sum of 1870 figures in Table 2, line 3 and Table 3, line 1; 1880, 1890, 1900, Column 3 times Column 2 divided by Column 1. a Thousand head.

(2) to estimate from a variety of sources the ratio of the weight of slaughter to the inventory as the period progressed. With respect to the estimation of inventory trend, the principal problems were, first, to locate the Census year in the commercial hog cycle (not discussed here) and second, to decide whether to use the June 1 inventory as reported by the Census, or to use the January 1 annual series estimated by the Department of Agriculture from 1867 (USDA, 1938), and projected back for the Census years 1840, 1850, and 1860 by Robert Gallman in a recent careful study (Gallman, 1956) of national income in the 1840–1880 period. The January 1 inventory series, which runs annually from 1867 – interpolations between the Census years being made mainly on the basis of the contemporary estimates of the crop reporters3 – made it possible to locate the Census years within the cycle, which shows up with a certain regularity from the start. The January inventories, therefore, could be taken as averages of a cycle, or one-half of a cycle (upswing or downswing) centered on the Census year. For the three Census years before 1870, no such adjustment was possible on the basis of the USDA series. The problem

310

Table 2.

WILLIAM N. PARKER

Estimate of Commercial Pork Production, Cycle Averagesa, 1840–1900. 1840

1850

1860

1870

1880

1890

1900

1. Estimated total commercial 1,077 pork (thousand head) 2. Estimated commercial 0.66 slaughter ratio; applicable to June 1 inventory 3. Estimated June inventory 1,616 destined for commercial slaughter (thousand head) 4. Percentage share of inventory 6 equivalent of commercial slaughter in total inventory (percent) 5. Percentage share of inventory 94 equivalent of farm slaughter in total inventory (%) 6. Average live weight per head 200 of commercial slaughter (lbs) 7. Total live weight of 215 commercial slaughter (million pounds)

2,635

3,749

6,544

16,840

21,382

37,133

0.66

0.66

0.70

0.82

0.87

0.89

3,953

5,623

9,348

18,921

24,577

41,525

13

17

25

39

46

66

87

83

75

61

54

34

230

260

274

252

239

227

606

974

1,793

4,244

5,110

8,429

Line 1: 1900: Average of 1899–1902, Agricultural Statistics, 1942 (p. 404). Cincinnati Price Current figure is 36,730. Other years: Estimated on basis of series for Western pack and Easter pack from Cincinnati Price Current, Statistical Annual, 1881–1913. Line 2: 1840, 1850, 1860: Estimated equal to farm slaughter ratio (see Table 3, Line 2, source note). 1870, 1880, 1890: Interpolated between 1860 and 1900 on basis of increase in ‘‘summer’’ pork pack, shown in Cincinnati Price Current, op. cit., assuming a relation between extension of ‘‘summer’’ (March–October) slaughter and declining age at slaughter. 1900: Obtained by subtracting the June inventory equivalent farm slaughter from total inventory (Table 1, Column 4), and dividing the remainder into the estimated nonfarm slaughter. Estimates of farm and nonfarm slaughter from Agricultural Statistics, 1942 (p. 404), derived by USDA largely from statistics of federally inspected slaughter. June inventory equivalent of farm slaughter derived by dividing farm slaughter by assumed farm slaughter ratio (see Table 3, Line 2, source note). USDA series begins with 1899. Average of years 1899–1902 for farm and nonfarm slaughter, were used in the computation. This yields figures for farm and nonfarm slaughter of 13,950 and 37,133, respectively. Dividing 13,950 by 0.66 yields 21,136, which subtracted from 62,661 leaves 41,525 as inventory for commercial slaughter of 37,133. Line 3: Line 1 divided by Line 2. Line 4: Line 3 divided by June inventory (Table 1, Column 4). Line 5: 100% minus Line 4. Line 6: Cincinnati Price Current, op. cit. (Parker indicated in his manuscript that he needed to add source references for the 1840 and 1850 numbers – Ed.). Line 7: Line 1 times Line 6. a See Table 1, Column 2, source note.

311


Table 3.

Estimate of Pork Produced from Hogs Slaughtered on Farms, Cycle Averages, 1840–1900a. 1840

1. Estimated June inventory for 24,685 farm slaughter (thousand head) 2. Estimated farm slaughter 16,292 (thousand head) 3. Estimated live weight of farm 3,258 slaughter (million pounds)

1850

1860

1870

1880

1890

1900

26,401

27,890

28,484

29,166

29,318

21,136

17,425

18,407

18,800

19,250

19,350

13,950

6,921

4,602

4,700

4,812

4,837

3,488

Source: Line 1: 1900 and 1870: Line 2 divided by assumed farm slaughter ratio (0.66). Other years: June 1 inventories (Table 1, Column 4) minus commercial inventories (Table 2, Column 3). Line 2: 1900: USDA, Agricultural Statistics, 1942 (p. 404). 1899–1902 average. 1870: Line 3 divided by estimated live weight (250 pounds). Other years: Line 1 times assumed farm slaughter ratio of 0.66 (see text). Line 3: 1870: Interpolated between 1860 and 1880. Other years: Line 2 times assumed live weights of 200 pounds for 1840, 225 for 1850, and 250 thereafter (see text). a See Table 1, Column 2, source note.

then was to estimate both the amplitude of the cycle and the point at which the Census count fell in these decades. This problem, especially the second part of it, is probably not insoluble. The maximum amplitude in the period 1870–1910, however is 715 percent from the trend, and there is a very strong reason to suppose that the cycle was concentrated mainly in the commercial sector of the industry. This sector becomes smaller very rapidly as we move back from 1870, so that the chances of any Census figure being more than 10 percent from the trend are extremely small in this period. Consequently an adjustment on this account was not attempted. The choice between June Census inventories and the estimated January inventories offered more difficulty. Both the USDA series and Gallman’s projection to 1814 appear to have been derived mainly to furnish a basis for income estimates. The advantage of January over June inventories for this purpose is not clear, but for deriving consumption estimates, the use of January inventories offers almost insuperable difficulties. The difference in the inventory between the two dates is of course due to a difference in the seasonal pattern of births relative to slaughter. Although the seasonal pattern of slaughter changed markedly as between summer and winter during the period, it remained fairly well balanced as between the two halves

312

WILLIAM N. PARKER

Table 4.

Total and Per Capita Pork Production and Consumption, Cycle Averages, 1840–1900a. 1840

1. Live weight of slaughter 3,473 (million pounds) 2. Carcass weight of slaughter, 2,004 excluding lard (million pounds) 20 3. Exports of pork and pork products, excluding lard, in carcass equivalent (million pounds) 4. Domestic consumption 1,984 (million pounds) 5. Population (thousand head) 17,120 6. Consumption per capita 116 (pounds)

1850

1860

1870

1880

1890

1900

4,527

5,576

6,493

9,056

9,947

11,917

2,612

3,217

3,746

5,225

5,739

6,876

90

75

221

757

650

984

2,522

3,142

3,525

4,468

5,089

5,892

23,261 108

31,513 100

40,575 87

50,464 88

63,721 80

76,226 77

Source: Line 1: Table 2, Line 7 plus Table 3, Line 3. Line 2: Line 1 times 3,577, factor derived from Holmes (1907, pp. 55, 61). Holmes gives 220 pounds as average live weight, 160 pounds as corresponding carcass weight including lard, and 33 pounds as average weight of lard. Line 3: Export data from Gries (1929). Carcass equivalents of exports given as follows:

Bacon Ham Canned pork Fresh, salted, and pickled pork

1.16 1.04 1.18 1.00

Exported weight

Line 4: Line 2 minus Line 3. Since exports were preponderantly bacon and ham, domestic consumption included a greater than average proportion of leaner cuts. Line 5: United States Department of Commerce (1960), Series B, pp. 31–39. Line 6: Line 4 divided by Line 5. a See Table 1, Column 2, source note.

of the calendar year. The winter slaughter appears to have been distributed over the four months from November to February. The introduction of the so-called summer pack, extending commercial slaughtering from March to October, was made possible by the increasing use of ice after 1870. It had the effect simply of distributing the slaughter more evenly throughout the year. The fly in the ointment, however (to use a somewhat inappropriate metaphor) appears to be the spring pigs. The concentration of pig births in the spring appears to have been a development beginning especially in the


313

North Central states after 1870 as Iowa and the Dakotas come increasingly into production. Both the regional differences in the June and the estimated January inventories in these years and the regional differences in the pig count made annually after 1920 by the Department of Agriculture suggest this.4 Moreover, the difference between the June and the estimated January inventories grows between 1880 and 1890 and between 1890 and 1900. The principal objection to the use of the annual January inventory series is that January inventories do not include as large a portion of all the animals slaughtered during the year as do the June inventories. The number of spring pigs slaughtered before the end of the year shows up in the June inventories, but must be taken into account by a rise in the slaughter ratio applicable to the January 1 inventories. Where this practice (i.e., spring pigs slaughtered in nine months) is on the increase – as it appears to have been after 1870 – very queer results are produced in transforming slaughter ratios based on June inventories into ratios applicable to the January inventories, from which the slaughtered pigs are missing. Consequently, for the years except 1870 when the Census is clearly in error, the Census count was accepted and adjusted after 1870 for its place in the hog cycle by a factor derived from the relation of the January inventory for the Census year to the cycle average computed from the January inventory at that point. Since it is almost certain that the reporting by the Census became, if anything, more complete as the century progressed, the use of the Census inventory estimates may understate the fall in inventories per capita. To this extent our estimate of the decline in pork consumption is conservative. Given the estimated cycle-adjusted trend of June inventories, the problem then was to transform these into estimates of the movement of the slaughtered weight produced. For this purpose it was necessary to consider essentially a single technological relationship: the weight–age curve for hogs. (The estimates of slaughter ratio and average weight were thus interrelated, since too low an assumed slaughter ratio could be at least partly compensated for by too high an estimate of the average weight at slaughter.) Where the cut is made in this curve to produce the meat supply in any one year depends in commercial farming, of course, on the meat–feed price relationship, and certain relative price conditions can cause hogs to be fattened at younger ages. The relationship set by technology is less a line than a border zone; nevertheless, over a period of half a century, our interest must lie not in the year-to-year changes but in shifts of the curve permitted by improvements in breeding and feeding practices. Now on the basis of what is known about regional differences after 1900, it appears extremely probable that most of the improvements in hog raising

314

WILLIAM N. PARKER

came in the commercial sector of the industry. It was necessary, therefore, to estimate the weight–age relationship for the farm and the commercial sectors separately. No attempt was made, of course, to draw an exact curve; instead slaughter ratios and average live weights at slaughter were computed from the available evidence. This evidence consists of two sorts: (1) the USDA estimates available from 1899 of numbers slaughtered on farms and commercially and of the carcass weight of the total slaughter, and (2) the various contemporary nineteenth-century sources. The method involved first an assumption of a weight–age relationship for farm-slaughtered hogs which yielded an average age at slaughter of 18 months with an average weight of 250 pounds in 1900. On this basis, the USDA estimate of the numbers slaughtered on farms in 1900 was transformed into an inventory equivalent. The remainder of June inventory – shown by the Census – was that available for commercial slaughter, and this relative to the estimated commercial slaughter in the USDA series in 1900 gave a commercial slaughter ratio (i.e., slaughter/inventory in June) just under 0.90, implying an average of just over one year. Average weights of the commercial slaughter in the West and the numbers slaughtered were available back to 1863 in the remarkable series of the Cincinnati Price Current (1881–1913). It was possible to estimate from these the total commercial slaughter. The commercial slaughter ratio was assumed equal to the farm slaughter ratio at the beginning of the period and was raised beginning in 1870 to its 1900 level, in proportion to the increase in the summer pack in the total commercial slaughter. Since this development was accompanied by a less than proportionate decline in the average weight at slaughter, the argument for a rising slaughter ratio which – combined with improvements in breeding – might have induced a relatively higher weight at earlier ages over the period, appeared quite strong. From the commercial slaughter and the commercial slaughter ratio, the estimated inventory equivalent of the commercial slaughter was derived, and this subtracted from the total inventory left the numbers available to be slaughtered on farms. So summary a statement of the method used does not do full justice to the variety of contemporary sources on which the estimates of the relationships have been based. The peg upon which the whole structure hangs is the estimate of the age–weight relationship for farm-slaughtered hogs; the assumption of a 200-pound hog in 18 months at the outset of the period, however, does not appear unduly generous, and this weight is allowed to rise to 250 pounds by 1860. A slightly lower weight might be assumed for 1840,


315

but a less rise is also possible, so that the net result in the movement of the final per capita figure might not be greatly changed. Nevertheless it would be desirable, first to experiment with other assumptions to see what per capita consumption would be yielded consistent with the inventory and commercial slaughter data, and, second to direct further research toward determining a weight–age curve for hogs on American farms in the nineteenth century. Two further shadows that fall across the meat estimates must be mentioned: first, changes in the relationship between live weight at slaughter and retail food equivalent, and second, changes in the numbers of animals kept in villages and cities out of reach both of the agricultural Census taker and the B.A.E. estimator. However, it would be tedious to pursue these questions further at this time.

MAIN LINES OF FURTHER INVESTIGATION I would like now to sketch the main lines of investigation into certain implications of these trends. In doing so, the discussion will go beyond what has been clearly established by the data sifted to date and will suggest certain conclusions that may perhaps never be fully established. The epistemological problems involved in this sort of venture I shall comment on in the concluding paragraph of this paper. First, it appears desirable to consider whether these trends in output relative to population were accompanied by rising production costs in the farm products taken relative to one another and to other products in the economy. Here we have mainly the series on farm prices in about a dozen states to go on, as exploited by Bean and Strauss in their estimates of farm income. I am indebted to my colleagues, Messrs. Potter, and Christy, at Resources for the Future for the general conclusion that between 1870 and 1910, except perhaps for hogs, no striking rises in farm prices relative to the wholesale price level occurred. Their work is showing a number of qualifications to this general conclusion. However, the representativeness of the Bean and Strauss data on several commodities is open to serious question. Accepting this conclusion, however, for purposes of argument, one has the impression from farm price data of a relative elastic supply situation in agriculture in the period. The major means by which this elasticity was produced – opening of new lands, farm machinery, regional specialization

316

WILLIAM N. PARKER

permitted by the railroad – are familiar stock in trade of the economic historian. I should like to add to this lore two new suggestions: (1) that perhaps the major factor may have been something we can call the ‘‘flexibility of the livestock production function.’’ It may well have been the shift in livestock feeding from grazing to hay and grain feeds that permitted the growth of the livestock population without sharp rises in the relative prices of livestock products. But such conversion by the techniques of 1850 would have implied a large labor input just at the time when the demand for an industrial labor force was growing most rapidly. Then unless there had been a speeding up of immigration, or an even more rapid rate of labor-saving technological change, there would have been a drag on the rate of industrial growth. It is here then that the labor-saving techniques in the field crops are seen to have had their main effect. (2) that apart from the drain of labor from agriculture, a great deal of the continuous improvement in productivity, and the sensitivity to commercial demand, may possibly be attributable to the pressures induced by the general price deflation between 1873 and 1896. American economic history has always had a strong agrarian tinge, stemming perhaps from the school of Turner, and it has been popular to look on the deflation and all the other squeezes which the farmer suffered in that period as an unmitigated evil. It must be remembered that real costs were falling and markets expanding very rapidly in these same years. The American farmer might well have responded to an improvement in his earning capacity in the manner traditional to peasant agriculture: by withdrawing from commercial markets to increase his consumption of leisure time or to improve the conditions of his living by work on and around the farmhouse. What then would have happened to surpluses of farm products and to their prices, to the real wages of city dwellers, and to the inclination of the young farm population to move to the cities? Fortunately, if we speak in cold terms made familiar by modern experiments in the industrialization of peasant economies – the American farmer was crushed by a burden of debt and by excessive fixed charges and money demands under a falling price level, and this pressure kept him tied to the commercial market. The sources of supply elasticity, however, are not quite to the point of the present investigation. A more novel, and in a sense more important, task is to examine the problem posed for agricultural supply by the demand movement indicated by the per capita consumption trends (assuming that


317

these did occur at relatively stable farm prices). This involves measuring, under the existing production functions in agriculture, the demand for productive factors in agriculture derived from the observed movement in consumption. Such a measurement is proceeding under great difficulties, but one fact appears quite certain: at constant yields in the feed crops (which was in fact about the case), the amount of land required to feed an American did not rise. The stable or declining trends in the major land-using foods – the meats and cereals – are sufficient to insure this. Instead, the composition of output moves toward the more heavily labor-using products. If vegetables, which probably rose sharply, and cotton are included, this movement would be even more pronounced. In short, the level and composition of output appears to have altered in such a way as to reduce somewhat the impact of the growth of population and of per capita incomes on the demand for agricultural land. And the check appears not to have occurred directly through rising farm prices. An examination of why demand moved in such a way as to produce these per capita trends is complex, and certain of its elements are quite impossible to quantify. We have, for example, no good index of retail food prices; however, apart from the movement of distribution margins, it is certain that the growing proportion of nonfarm consumers meant a rise in the average margin borne by the food consumer and altered relative retail food prices in striking ways. It is possible, for example, that in the cities and villages, the ratio between the price of flour and the price of sugar was relatively more favorable to sugar than on farms. Similarly the ratio between the price of beef and the price of pork was more favorable to pork on the farms – where home processing was easier for pork than for beef – than in the cities, where both had passed through much the same channels of distribution. A great deal of these relative-price calculations depends of course upon the valuation put at the farm on family labor, but before the automobile and with a high value put on money income, this value was undoubtedly low. Essentially, the attempt is being made in the present research to establish through the evidence of budget studies and other contemporary sources some rough picture of the diet typical to western, southern, and eastern agriculture, to urban workers, and – where possible – to immigrants. If from the sketchy data can be produced a set of diets which, when weighted by the changing relative shares of these components in the population, produces a movement in food consumption close to that actually observed, we shall come as close to proving the assumed set of diets as one can get in historical work. The declining per capita consumption of meat – a phenomenon which

318

WILLIAM N. PARKER

appears to be peculiar to the young agricultural counties in the process of industrialization – can thus be explained by the high consumption of pork in farm diets, and possibly by the lower meat consumption characteristic of immigrant dietary habits. Since both the composition of the diet and its level are controlled ultimately by physiological requirements in relation to occupation and manner of living, it would be useful to follow Raymond Pearl’s pioneering work (Pearl, 1920) and compute caloric and protein intake, and the data appear to be complete enough to permit that. As Working pointed out, the declining flour and cornmeal consumption after 1890, and in our data the two series combined yield a declining caloric intake. The decline in cornmeal consumption is the strongest factor reducing the average caloric intake. Similar calculations are being made for the protein intake, where some rise will probably be shown. Thus the elements of the dietary shift, involving – at least in terms of calories an absolutely lower level of consumption, which has been studied by nutritionists in the Bureau of Human Nutrition and elsewhere for the years since 1909 – appear in the preceding 30- or 40-year period. To changes in relative retail food prices, in the composition of the population among different diets, in the physiological requirements and other physical elements affecting ‘‘tastes’’ must doubtless be added a decline in the prestige value of a ‘‘good table’’ – an element which colors even urban tastes in a predominantly agricultural society but which disappears even from agriculture in a predominantly industrial society. However, I realize that the discussion of social elements of this sort, which wholly defy quantification, is as distressing to econometricians as statistics are to historians, and I will not follow this line of influence further in this company. I would be less than fair to my own profession, if I were to imply that the explanation of trends of this nature can rest wholly on statistical investigation. Yet the value of statistics and theory in describing and explaining the movement of economic history is very great indeed. The study of food consumption demonstrates two principal uses. First, contemporary statistics are used to define a limited problem – one of the many trends exhibited by the economy over a portion of its history – and in the explanation of why that trend occurred we are able to uncover a portion of the network of economic and social relationships through which the impulse of economic growth was transmitted. The study, I think, is showing that apart from the well-known improvements in agriculture productivity which kept farm prices relatively stable, numerous more deeply hidden elements on the side of demand served to relieve the pressure of our growth on our agricultural resources. Without at lest a crude statistical measure of per capita


319

consumption, this chain of investigation would hardly suggest itself, and its relative importance could not be weighed. Second, both in making statistical estimates of consumption and in explaining the trend, we are able to draw on statistical and econometric work of the period that we may call the Era of (relatively) Good Statistics. The estimate of pork consumption, for example, involves the projection backward of relationships between inventory and carcass weight of the output that are evident in the data since 1899. Recent trends in some of these technological relationships must be carefully set against contemporary evidence of agricultural technology, since it is as dangerous to project backward as it is to project forward. Yet – as the Red Queen remarked to Alice – it is one of the advantages of living backward that one’s memory works both ways. It is a poor sort of projection that looks only forward; indeed I often wonder why econometricians are so concerned with looking forward, since by living a few years we will know what will have happened, whereas when we project backward, we illuminate a portion of human experience that must otherwise be forever lost. In addition, as we all know, prophets are far more often unmasked by the future than are historians embarrassed by the results of more diligent research. In addition to the backward projection of certain technological relationships, the examination of the recent studies of income elasticities of farm and urban consumers – as well as some crude attempts of our own to make statistical sense out of some of the very scanty contemporary budget study material – is proving suggestive in the explanation of the declining trend of consumption of the typical frontier farm products – notably pork and cornmeal. In this way, on a larger scale, with the help of recent statistical measurements and the relationships suggested by economic theory, it may ultimately be possible to weave together from the contemporary statistics and the documentary sources a portion of the picture of the economic mechanism by which American growth was produced in the nineteenth century. If such a picture does no violence either to the historical evidence or to our half-intuitive preconceptions about economic interrelationships, it will impress us as being true, since the degrees of freedom are sufficiently numerous in historical work that we can never rule out the possibility of several alternative, equally plausible interpretations. But it will have the merit of being self-consistent and systematic. The existence even of several alternative explanations of the growth of output and productivity in the United States would have the advantage of directing historical research toward problems in which the validity of one explanation over another could be tested.

320

WILLIAM N. PARKER

In that case the phrase ‘‘further research’’ would lose the dreary suggestion of endless unplanned fact-finding expeditions – a legacy of the nineteenth century concept of the scientific method in history. Further research is after all the historian’s substitute for experiment, and it must be applied at crucial points to permit us to choose between theories. Perhaps the greatest advantage of employing theory and statistics in economic history – even in a crude and halting way – is that one need no longer give inquiring students that most discouraging of all advice: that they ‘‘follow their own interests.’’ The existence of a stock of material is of course a prerequisite of research, but amid the abundance of unexploited historical materials, real interest is given only to questions that are of some crucial importance for a more general understanding of our historical experience. The conscience of modern scientific method has cast an enchantment over traditional historiography, and the American past sits chained up and inarticulate as the lady in Comus. My fellow economic historians, on hearing that I have met the econometricians face to face, may indeed address me with that excess of humanistic learning with which historians continue to be burdened, with Milton’s own exclamation: ‘‘What! Have ye let the false Enchanter scape? Oh ye mistook; ye should have snatched his wand And bound him fast. Without his rod reversed And backward mutters of disserving power We cannot free the Lady that sits here In stony fetters fixed and motionless.’’

NOTES 1. The term ‘‘econometrical history’’ can of course be produced by a few simple linguistic transformations from the term ‘‘quantitative economic history’’ which A. P. Usher used to characterize his and Clapham’s work. 2. Because of the greater proportion of fat cuts in pork, the caloric equivalent of this consumption may have fallen more rapidly, and the protein content less rapidly than the total weight. On the other hand, exports were preponderately fat cuts (bacon), leaving a disproportionate share of leaner cuts for domestic consumption. 3. For Livestock, tax assessors’ listings were used in revising the interpolations. For hogs, however, listings frequently excluded young animals. 4. For this reason, Gallman’s adjustment of the pre-1870 Censuses to a January 1 basis by the use of the 1920–1924 relationship offers an instance of the dangers of projecting backward the technological relationships of a later date.


321

ACKNOWLEDGMENTS This research has been generously supported by a grant from Resources for the Future, and thanks are due especially to Harold J. Barnett and his colleagues of that organization, to Richard Ruggles (Yale), and to R. W. Pfouts (University of North Carolina) for useful comments and criticisms.

REFERENCES Cincinatti Price Current. (1881–1913). Statistical Annual. Gallman, R. (1956). Value added by agriculture, mining, and manufacturing in the United States, 1840–1880 (Chapter III, pp. 104–129). Doctoral dissertation, University of Pennsylvania. Gries, C. (1929). Foreign trade of the United States, hogs pork and pork products. Reports FS 44. United States Department of Agriculture, Bureau of Agricultural Economics, November 27. Holmes, G. K. (1907). Meat supply and surplus. Bulletin 55. United States Department of Agriculture, Bureau of Statistics, Washington, DC. Montgomery, E. G., & Kardell, C. H. (1930). Apparent per capita consumption of principal foodstuffs in the United States. US Department of Commerce, Domestic Commerce Series, no. 38. Pearl, R. (1920). The nation’s food: A statistical study of a physiological and social problem. Philadelphia, PA: W. B. Saunders. Strauss, F., & Bean, L. H. (1940). Gross farm income and indices of farm production and prices in the United States, 1869–1937. Technical Bulletin no. 703. United States Department of Agriculture (December). United States Department of Agriculture. (1938). Livestock on farms, January 1, 1867–1919. Revised estimates, Mimeo (January). Washington, DC: Department of Agriculture. United States Department of Agriculture. (1942). Agricultural Statistics, 1942. Washington, DC: Government Printing Office. United States Department of Agriculture. (1953). Consumption of food in the United States, 1909–52. In Agriculture Handbook No. 62 (September) and Supplement (October, 1955). United States Department of Commerce. (1960). Historical statistics of the United States, colonial times to 1957. Washington, DC: Government Printing Office. Working, H. (1926). The decline in the per capita consumption of flour in the United States. Wheat Studies of the Food Research Institute (Vol. II, No. 8). Stanford University.

Research in Economic History, Volume 27

Research in Economic History, Volume 25 (Research in Economic History)

Research In Economic History

Research in Economic History, Volume 24

Research in Economic History. Volume 21

Research in Economic History Volume 26

Research in Economic History, Volume 22, Volume 22

Advances in Ecological Research, Volume 27

Advances in Food Research, Volume 27

Advances in Cancer Research, Volume 27

Advances in Food Research, Volume 27

Aboriginal History Journal; Volume 27

Research in Economic History (Vol 20)

Advances in Catalysis, Volume 27

Advances in Parasitology Volume 27

Progress in Optics, Volume 27

Advances in Immunology Volume 27

Choice in Economic Contexts, Volume 25: Ethnographic and Theo Enquiries (Research in Economic Anthropology) (Research in Economic Anthropology)

Research in the History of Economic Thought and Methodology, Volume 24 A: A Research Manual

Research in Personnel and Human Resources Management, Volume 27

Research in Social Movements, Conflicts and Change, Volume 27

Research in the History of Economic Thought and Methodology: Documents From the History of Economic Thought

An Economic History of Film (Routledge Explorations in Economic History)

Economic History of Sweden (Routledge Explorations in Economic History, 16)

Grantville Gazette Volume 27

Geographers, Volume 27

Organometallic chemistry, Volume 27

Horticultural Reviews, Volume 27

Research in the History of Economic Thought and Methodology

Torrid Teasers Volume 27

Research in the History of Economic Thought and Methodology, Volume 25B: Documents From the History of Economic Thought (Research in the History of Economic ... History of Economic Thought and Methodology)

Research in Economic History, Volume 27

Research in Economic History, Volume 25 (Research in Economic History)