The Japanese Tax System, Third Edition

THE JAPANESE TAX SYSTEM This page intentionally left blank The Japanese Tax System Third Edition HIROMITSU ISHI O...

Author: Hiromitsu Ishi

33 downloads 1138 Views 22MB 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

THE JAPANESE TAX SYSTEM

This page intentionally left blank

The Japanese Tax System Third Edition

HIROMITSU ISHI

OXFORD UNIVERSITY PRESS

This book has been printed digitally and produced in a standard specification in order to ensure its continuing availability

OXFORD UNIVERSITY PRESS

Great Clarendon Street, Oxford 0X2 6DP Oxford University Press is a department of the University of Oxford. It furthers the University's objective of excellence in research, scholarship, and education by publishing worldwide in Oxford New York Auckland Bangkok Buenos Aires Cape Town Chennai Dar es Salaam Delhi Hong Kong Istanbul Karachi Kolkata Kuala Lumpur Madrid Melbourne Mexico City Mumbai Nairobi Sao Paulo Shanghai Taipei Tokyo Toronto Oxford is a registered trade mark of Oxford University Press in the UK and in certain other countries Published in the United States by Oxford University Press Inc., New York © Hiromitsu Ishi 1989, 1993, 2001 The moral rights of the author have been asserted Database right Oxford University Press (maker) Reprinted 2004 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, or under terms agreed with the appropriate reprographics rights organization. Enquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above You must not circulate this book in any other binding or cover And you must impose this same condition on any acquirer ISBN 0-19-924256-9

For the late Dr Carl S. Shoup


Foreword

One of the most difficult but rewarding undertakings in the field of tax theory and practice is that of tracing the development of a given country's tax system over a period of decades. Japan's tax system is particularly interesting for such a study, since it has changed notably through periods of war, post-war reconstruction, high-level economic development, and moderated economic growth. Those of us, including myself, whose inability to read Japanese shuts us off from a considerable literature that might inform us on this subject will be especially appreciative of this treatise written in English by Professor Hiromitsu Ishi of Hitotsubashi University. Professor Ishi's thorough and perceptive analysis of the development of Japan's tax system over time, coupled with his detailed description and critique of that system as it stands today provides the reader with an unusual opportunity to view a tax system as a virtually living, organic whole, in enough detail to allow comparisons of generally accepted tax theory and tax principles with the real world of an existing tax system. Professor Ishi brings special qualifications to the writing of this book. He has composed a volume on taxation that is part of a Postwar Fiscal History of Japan undertaken by the Ministry of Finance. In the course of this study Professor Ishi has had access to first-hand information, including confidential documents. For many years he has served as a member of the Tax Advisory Commission, and is now involved in drawing up plans for current tax reform. Through his activities in scholarly international tax organizations, including the International Institute of Public Finance, and through several research and teaching posts abroad, notably in Australia and the United States, Professor Ishi has developed a comprehensive outlook with which to appraise the tax system of his own country, as shown, for example, by the numerous comparative tables in this volume. This outlook assures that many of his appraisals and conclusions will be widely applicable. Among the significant developments in the Japanese tax system since the adoption in 1949-50 of the Tax Mission recommendations, none has been more important than the movement, first, away from equity, towards tax preferences designed to achieve certain economic goals and facilitate certain political compromises, followed by a turnaround, now gathering momentum in Japan as elsewhere, toward fairness among taxpayers similarly circumstanced, the economic goals to be achieved by other means. Professor Ishi's detailed account of these phases is among the most important of his contributions. January 1989 C A R L s. S H O U P

Preface to the First Edition Japan's image in the eyes of the world has changed greatly in recent years; Japan is now seen as a great power, comparable with the USA in terms of economic and competitive strength. The story of the economic miracle of the Japanese economy has attracted the world-wide attention of academics and non-academics alike. Reflecting this long-standing interest in Japan, a large literature on the Japanese economy already exists, much of which is now available in English. It seems to me, however, that very little material is available in English regarding the Japanese tax system and tax policy issues. In fact, no book has yet been written to explain the actual performance of tax policies from an academic point of view specifically for English readers. Europeans, Americans, and others interested in understanding the Japanese tax system must continue to have difficulty gaining access to reliable literature. The main aim of this book is to help fill this gap. There is a great need for a systematic description of how the Japanese tax system has developed during or since the postwar economic growth period, and how it has contributed to the superlative performance of the Japanese economy. In addition, it is important to assess Japan's present and future tax reforms because the most sweeping reforms of the Japanese tax system have been debated in the past several years. Many see a need to reform the tax system in order to adapt to ongoing changes in the economy. The tax debate will be with us for years to come. In this sense, it is an appropriate time to publish this book. Generally speaking, this book is non-technical and is written not only for academics but also for the broader tax community (bureaucrats, businessmen, tax accountants and lawyers, etc.). It should also be of great help to general readers who are not tax specialists but are interested in Japanese economic policy in general. Instead of using sophisticated theoretical and econometric models, the discussion in this book will be developed in terms of what general insights are to be derived or what general lessons are to be learned from this course of study, and will include simple statistical procedures. Many people and circumstances have helped me in this undertaking. I am fortunate to be a member of the government Tax Advisory Commission and the Postwar Financial History Project at the Ministry of Finance. These valuable experiences have had a major influence on my ideas about tax issues and have enabled me to have easier access to the resources needed to pursue my research. Members of staff of the Research Division, Tax Bureau at the Ministry of Finance read parts of earlier drafts and saved me from some major errors. My deep thanks should be expressed to Dr Carl S. Shoup for writing the foreword in this book. In addition, I have accumulated many debts during my career to overseas friends who have contributed to my intellectual development and research

Preface to the First Edition

ix

through discussions and the exchange of papers: the late Joseph A. Pechman, Henry Aaron, Oliver Oldman, Richard M. Bird, Mervyn King, and Sijbren Cnossen, among others, have greatly assisted me in the writing of such a book in English. Final publication of the book would not have been possible without the warmhearted support from Andrew Schuller at Oxford University Press, who first invited me to write it. I also owe a special debt to Gillian Bromley and Sue Hughes for their kind help in preparing the manuscript. The Nissan Institute of Japanese Studies at Oxford University very kindly provided me with office space a couple of times for my writing; I wish to express my gratitude to Director Arthur Stockwin, and Jenny Corbett. Finally, I am greatly indebted to Hiroko Sugimoto at Hitotsubashi University, without whose dedicated typing and other assistance the manuscript could not have been completed, to David Gross, who assisted tremendously in editing the final drafts in English, and to Sinji Yamashige, who provided assistance in computing the empirical analyses. I would like to thank all of them very much.

H.I. Tokyo, December 1988

Preface to the Second Edition Four years have passed since I wrote the first edition of this book. At that time, the Takeshita Cabinet was making a strenuous effort to get the Tax Bill approved, including the introduction of the consumption tax (Japan's VAT), in the Diet. It was during the height of the movement among the opposition parties against the consumption tax. In preparing the first edition, I was unfortunately unable to witness and review the whole process of the Takeshita tax reforms. Since that time, the Japanese economy has changed considerably, generating the 'bubble' phenomenon, and there were two important tax reforms at the beginning of the 1990s. I think we need to study these changes in taxation and in the Japanese economy. Two tax reforms in particular are important: (1) amendments to the consumption tax, and (2) land tax reform, both of which I think must be considered in detail to understand recent developments in the tax system. This is the main reason why I am motivated to revise the earlier edition of the book. Another reason is that almost all the data cited in the first edition are now out of date; a new edition has made it possible to make use of the latest data available to us. All the equations have been recalculated to include the more recent period and the data in figures and tables have been completely revised. In the present edition, three chapters—Chapters 4, 16, and 17—have been added, while one chapter has been dropped. At the same time, Part IV has been virtually rewritten to cover recent tax developments, while changes have been made in each chapter. Fortunately, I have been deeply involved in the deliberations on recent tax reforms as a member of the Tax Advisory Commission (in particular, as chairman of the Subcommittee on Land Taxation in 1990). I am greatly obligated to those persons who have helped me in revising this book. (A number of book reviews have come out in relevant journals and literature.) I should like to express my gratitude in particular to Richard M. Bird, Peter Jackson, Simon James, Sylvain Plasschaert, William Vickrey, and Richard Woodworth for their constructive comments and suggestions. I owe much to both Andrew Schuller for the publication of this book and Tracy Mawson for her kind help with the manuscript. For the opportunity to work on the revision, I wish to express my indebtedness to the Nissan Institute of Japanese Studies at Oxford University, and in particular to Director Arthur Stockwin and Jenny Corbett who provided me with pleasant surroundings for my work. I also wish specially to thank the Daiwa Anglo-Japan Fund for financial assistance. As I was at the time of writing the first edition, I am again greatly indebted to Hiroko Sugimoto for her endless aid in the production of the manuscript and

Preface to the Second Edition

xi

to Yoshihiro Kaneko for computational calculations. Others who have assisted in one way or another include the staffs at the Ministries of Finance and Home Affairs, and at the National Tax Administration in collecting data or computing empirical results. I would like to express my sincere gratitude to all of them. H.I. Oxford, October 1992

Preface to the Third Edition Since the second edition of this book was published in 1993, the Japanese economy has completely changed from a success story of good performance to the recent agonies of prolonged recession after the collapse of the bubble phenomenon in 1991. After nearly zero real growth rates continued for the three years 1992-94, the stagnant state of the economy continued for a long period with little sign of recovery, although there was a slight upward movement during the two years 1995—96. What was worse was that once again the economy experienced minus growth rates during the years of 1997-98, as a result of increasing the consumption tax rate, the Asian economic crises, and successive bankruptcies of big banks in 1997. A great deal of pessimism has spread all over the country. Many people, in and out, of Japan even believe that Japan's successful era is over and that it will never return to the success of previous days. To combat the uncertain state of the economy, the Japanese government has repeatedly acted to stimulate aggregate demand by using a Keynesian type of fiscal expansionary policy with increased public investment and tax reductions. But the government has also occasionally tried to increase tax revenues to make up for revenue losses due to successive tax cuts. As a consequence, the Japanese tax system has been affected by a round of tax changes in its macroeconomic policies. In particular, the individual and corporate income taxes, and the consumption tax (Japan's value-added tax) experienced big changes in the 1990s. The main aim of this revised edition is to clarify the process and the effect of another tax reform following the two previous tax reforms under the Nakasone and Takeshita Cabinets in the late 1980s which were treated in previous editions. For this purpose, it has been necessary to trace back the tax reform process institutionally, to update all the data used in the second edition, and to revise all figures and tables. In this edition, the contents have been substantially altered, particularly in the latter part. One chapter—Chapter 16—is added to discuss environmental taxes which have become important and is included in Part IV, an independent Part, different from the second edition. At the same time, Part III has been rewritten to include land tax reform (Chapter 13) and public finance moved to Chapter 18 in Part V. In contrast, Parts I and II basically remain unchanged apart from the updating of the relevant figures and tables. As was in the case with previous editions, I am greatly obliged to those who have assisted me in revising this book. My special thanks to Andrew Schuller at Oxford University Press for his kind support, to Hiroko Sugimoto at Hitotsubashi University for her warmhearted assistance, and Toshiya Hatano at Chiba University for computational calculations. Equally my deep thanks to the staffs of the Tax Bureaux at the Ministry of Finance and the Ministry of Home Affairs who always

Preface to the Third Edition

xiii

helped me in collecting necessary data and computing empirical results. To all of them I wish to express my warm thanks. While I was revising my work, I was informed that Dr Carl S. Shoup had passed away at the age of 97.1 have been indebted to his scholarship for a long time, and he was kind enough to write the Foreword for the first edition of this book in 1989. To Dr Shoup I dedicate this book as a token of gratitude.

H.I. Tokyo, April 2000


Contents

List of Figures

xvii

List of Tables

xix

List of Abbreviations

xxiii

Introduction

hxx I. Overview

1

1. General Characteristics of the Tax System 2. Background of the Japanese Tax System

3 16

3. Two Strategies of Postwar Tax Policy 4. Tax Administration and Tax Equity

36 53

II. Individual Income Tax 5. Basic Structures 6. The Erosion of Individual Income Tax

73 75 94

7. Inflation Adjustments 8. The Taxation of Investment Income and Savings

110 130

9. Effects of Taxation on the Distribution of Income

147

III. Corporate Taxation and Taxes on Capital

165

10. Principles of Corporate Taxation

167

11. Corporate Tax Levels and Tax Incentives 12. Inheritance and Gift Taxes

187 204

13. Land Tax Reform

219 IV. Indirect Tax System

14. The Traditional Framework of Indirect Taxes 15. The Value Added Tax 16. Design of Environmental Taxes

249 251 268 300

xvi

Contents V. Recent Tax Developments

317

17. Rebuilding the Tax System

319

18. Local Taxation and Intergovernmental Fiscal Relations

349

19. Appraisal and Further Reform

383

References

397

Name Index

409

Subject Index

411

List of Figures 1.1 1.2 1.3 2.1 3.1 3.2 4.1 4.2 4.3 4.4 4.5 5.1

5.2 5.3 5.4 5.5 6.1 6.2 A6.1 7.1 7.2 7.3 8.1 9.1 9.2 9.3 9.4 10.1 11.1 11.2 11.3 11.4 12.1

Trends in the relative shares of direct and indirect taxes, 1934-1998 Trends in tax collection costs, 1965-1998 The taxation process under Japanese and US-Western systems Changes in the level and composition of Japanese tax revenues, 1885-1995 Actual and anticipated rates of economic growth, 1955-2000 The gap between government expenditures and tax revenues, 1960-1998 Administrative cost as a percentage of tax revenue, selected countries, 1960-1995 Comparison between per-staff tax revenue (T/N) and administrative cost (ON) Relative share of withheld income tax, 1960-1998 Tax delinquency as a percentage of revenue, 1960-1996 Tax gap among three income sources, 1970-1990 Marginal rates of income tax (excluding local taxes) up to an income of ¥10 million for a family of four with a single wage-earner; selected countries, 1998 An exposition of Japan's income tax schedule Variations of progressivity of income tax, 1951-1997 Tax progressivity by income class: self-assessed income tax, 1970 Tax progressivity by income class: taxes on wage and salary incomes, 1970 Income tax erosion by income class, 1996 The ratio of tax erosion, selected years Steps in creating a comprehensive tax base, 1996 Changes in effective tax rates, actual and after inflation adjustment, 1960-1975 and 1975-1990 Tax thresholds for a working family of four, 1970-1992 Tax thresholds and inflation, 1970-1992 Movements of personal savings rates in six major countries, 1965-1995 The Lorenz curve, before and after taxes The movement of income distribution in terms of the Gini coefficient, 1951-1997 Redistributive effects of income tax, 1951-1997 Redistributive effects of self-assessed income taxes, by types of income, 1951-1997 Trends in self-financing by corporate firms, 1968-1998 Corporate tax revenue losses from special tax measures as a percentage of corporate tax revenue, 1965-1997 Effective corporate tax rates before and after adjustment for special tax measures, by size of corporation, 1973-1996 Nominal tax rates of corporate taxes, Japan and USA, 1970-1993 Average effective tax rates of corporate taxes, Japan and USA, 1970-1993 Computation of the Japanese tax base

10 11 13 19 49 50 55 59 61 64 68

83 85 87 89 90 102 104 108 123 128 129 137 150 151 152 155 183 189 196 199 203 207

xviii 12.2 13.1 13.2 13.3 13.4 13.5 14.1 14.2 15.1 16.1 16.2 16.3 17.1 17.2 17.3 17.4 17.5 17.6 18.1 18.2 18.3

List of Figures Computation of taxes due Long-term trends of land prices (nationwide, residential land), 1956-1998 Movement of land tax revenue and land price Long-term gains taxes of individuals for land sales after 1989 Corporate capital gains taxes for the sales of land after 1982 A schematic chart of property tax assessment Alcohol tax as a percentage of alcohol consumption and national tax revenues, selected fiscal years Total tax burdens of car-related taxes on the standard passenger car: international comparisons, 1997 Mechanism of the local consumption tax Alternatives of the environmental tax Tax rates, revenues, and earmarking of energy taxes, 1998 Patterns of household environment-related expenditures as a percentage of annual income, 1992 Percentage changes of non-entitlement expenditures over the preceding year, 1970-2000 Trends in the bond dependency ratio, 1970-2000 Package scheme of tax reductions and increases, 1994-1998 Trends of major national taxes, 1983-1999 Individual income tax structure among major countries, 1999 Effective corporate tax rates among major countries, 1999 Tax shares between central and local governments, fiscal 2000 Trends in the ratio of local taxes to total annual revenues at local level, 1955-1998 Relative shares of four major taxes, 1955-1998

208 220 242 244 245 246 256 263 298 306 307 312 322 323 340 342 343 346 353 355 358

List of Tables 1.1 Source of tax revenues in Japan and seven major countries, 1995 1.2 Tax levels in OECD countries: tax revenues as a percentage of GDP, selected years 1.3 Sources of tax revenues, by source, selected years 1.4 Tax revenue as a percentage of national income, by level of government, selected years 2.1 Correlation of per capita real income (yIN), openness (M/Y, (M+X)/Y), and agricultural products' share (Ag/Y) with tax shares (T/Y), 1885-1997 2.2 Regression equations for tax sources as a function of economic development variables, 1885-1944 and 1953-1997 A2.1 Shoup Mission recommendations and their modifications 3.1 Comparison of estimated revenue loss from special tax measures and total individual and corporate income tax revenue, 1958-1998 3.2 Percentage distribution of the estimated revenue loss from special tax measures, by type of incentive, 1958-1997 3.3 Estimated annual tax changes, fiscal years 1950-1998 3.4 Year-to-year income elasticity of the national tax system, 1965-1998 4.1 Tax collection methods by taxable income in 1996 4.2 Conceptual adjustment between TS-base and NIS-base incomes 5.1 Minimum taxable level of the individual income tax, selected countries, 1998 5.2 Different income tax liabilities among one-earner and two-earner couples with a combined income of ¥10 million, 1998 5.3 Income tax liability and income-splitting, L998 5.4 .Statutory rates of income taxes at all levels of government, 1999 5.5 Major changes of tax rates, national income tax, 1949-1999 5.6 Taxes on personal income as a percentage of GDP and at the income level of an average production worker (APW), selected countries 6.1 Comparison of actual and comprehensive income tax base and yield, 1996 6.2 Comparison between actual and comprehensive income tax rates, 1996 7.1 Effect of 10 per cent inflation on the tax liability of a family of four with an annual income of ¥5 million, 1975 7.2 Ten per cent inflation, statutory tax rate, and the deduction for employment income, 1975 7.3 Ten per cent inflationary effects on income tax liability for a family of four, selected income levels, 1975 7.4 Inflation and the self-assessed income tax liability, 1960-1990 7.5 Inflation and withheld income tax on wages and salaries, 1960-1990 7.6 Income taxes and fiscal dividend, 1960-1975 8.1 Personal saving with non-taxable interest income, 1987 8.2 Total principal outstanding of personal saving, by source, as of March 1986

5 7 9 10 21 23 33 39 41 43 47 60 67 76

78 79 81 82 93 100 101 112 113 114 118 120 126 131 132

xx 8.3 8.4 8.5 8.6 9.1 9.2 9.3 A9.1 A9.2 10.1 10.2 10.3 10.4 10.5 10.6 10.7 11.1 11.2 11.3 11.4 12.1 12.2 12.3 12.4 12.5 13.1 13.2 13.3 13.4 13.5 13.6 14.1 14.2 14.3

List of Tables Special tax measures for interest income: reduced tax rates in the case of separate taxation at source Proportion of families using the privileged saving system, by income class, 1986 Proportions of portfolio investments in the household sector, selected years Equations estimating personal savings, 1965-1998 Percentage distribution of taxpayer, taxable income, and tax yield under the self-assessed income tax, 1955-1995 Redistributive effects of taxation, by household, 1972 Redistributive effects of taxation, by household, 1984 Redistributive effects of the self-assessed income tax, 1951-1997 Redistributive effects of the withheld income tax on wage-salary incomes, 1951-1997 Numbers of corporations, by the scale of paid-in capital, 1996 Nominal tax rates among major advanced countries, 1998 Historical patterns of the corporate tax rate, 1950-1999 The ratio of tax credits Types of company taxation in OECD member countries, 1987 Adjustment of corporate and individual income taxes, 1974 Comparison of the total tax burden of corporation and non-corporation in 1961(before tax revision) Revenue losses of special tax measures of corporations, by type, 1977, 1987, and 1998 Effective corporate tax rates before and after adjusting for special tax measures, by size of corporation, 1973-1996 An international comparison of corporate tax rates, 1984 Taxable and economic incomes of non-financial corporations, Japan and USA, 1970-1993 Inheritance and gift tax rates, 1975-1987 and 1998 Trends of tax payments in the inheritance tax, 1958-1996 Inheritance tax revenues, 1950-1998 The inheritance tax burden, 1998 Redistributive effects of the inheritance tax, 1958-1996 Outline of land taxation, 1990 Effective tax burden of the land-holding tax, 1970-1995 The ratios of property tax assessment to official valuation of land price, selected years Outline of land transfer taxation, 1990 The property tax on agricultural land in Urbanization Promotion Areas, 1971-1991 The burden ratio of land value tax Percentage distribution of national indirect taxes, selected fiscal years Alcohol tax rates as a percentage of retail prices, 1987, 1992-2000 Tobacco tax burden on retail prices, 1987 and 1998

133 138 140 144 156 159 159 162 163 170 171 173 176 177 179 181 191

193 198 201 210 211 213 214 216 224 225 226 228 231 243 253 257 259

List of Tables Proportion of cigarette taxes in retail prices, selected countries, 1985, 1991, and 1997 14.5 Car-related taxes, 1998 14.6 Proportion of taxes on petrol in the retail price, selected countries, 1985-1986 14.7 Commodity tax revenues collected from the top ten major items, 1986 14.8 Rate structures of the commodity tax, 1986 14.9 The commodity tax as a percentage of retail price, major items, 1987 15.1 Past trends of VAT among OECD countries 15.2 History of tax reforms in Japan before 1989 15.3 Results of public poll for or against the introduction of the consumption tax by mass-media 15.4 Various tax rates applicable to different sales—the case of non-wholesale firms, 1989 15.5 Tax variations in the national and local tax system under the Takeshita tax reform 15.6 Trends of the consumer price index (CPI) in a nationwide area 15.7 Effects of VAT on prices of major commodities: results of the sixth price monitor 15.8 Extent of forward-shifting, May 1989 15.9 Extent of forward-shifting by firm size at the retail stage, fiscal 1989 15.10 Full-, over-, and under-shifting at the retail stage by business type, April 1989 15.11 Small and medium traders applying to the special simplified rule as a percentage of total taxpayers, 1989 15.12 Value added ratio by type of industry under the application of the special simplified scheme 15.13 Chronology of events after implementing the consumption tax 16.1 CO2 emissions by major countries, 1991 16.2 Tax burden per ton of CO2 emissions, 1992 16.3 Hypothetical types of carbon tax and carbon/energy tax, 1992 16.4 Implicit carbon taxes, 1988 17.1 Tax increases of national corporate income taxes, 1980-1986 17.2 Tax increases of national domestic indirect taxes, 1980-1984 17.3 Reform package of tax increases and reductions: tax bill proposed in July 1988 17.4 Revenue changes of income taxes, full-year basis 17.5 A modified flat rate structure of the income tax under the proposed tax reform plan, 1988 17.6 Basic rate on retentions % 17.7 Removal of reduced tax rates on dividends % 17.8 Changes of tax rates, national income tax, 1993-1999 17.9 Changes of major exemption and deductions, national income tax, 1993-1999 17.10 Changes of the national corporate tax rates, 1990-1999 18.1 Total annual revenues of local government, fiscal 1998

xxi

14.4

260 262 262 265 266 266 269 275 280 283 284 286 287 288 289 290 293 293 295 302 308 311 314 324 325 329 331 334 337 337 345 345 347 354

xxii 18.2 18.3 18.4 18.5 18.6 18.7 18.8 18.9

List of Tables Local tax collection, by source, 1999 International comparison of taxes on real estate, selected years Local transfer taxes, fiscal 1998 Fiscal equalization at the prefectural level per capita, fiscal 1998 Trends of the tax-sharing ratio, 1954-2000 Local allocation tax: receiving and non-receiving local governments, fiscal 1998 Specific-purpose grants from the national government, fiscal 1998 Proportion of support coming from special-purpose grants, selected items, 1984-1999

356 363 364 368 370 371 377 378

List of Abbreviations EPA

Economic Planning Agency

LDP

Liberal Democratic Party (the present ruling political party)

LPFP

Local Public Finance Programme

MITI

Ministry of International Trade and Industry

MOC

Ministry of Construction

MOF

Ministry of Finance

MOHA

Ministry of Home Affairs

MOHW Ministry of Health and Welfare NLA

National Land Agency

NTA

National Tax Administration

SCAP

Supreme Commander for the Allied Powers

TIN

Tax Identification Number

VAT

The value-added tax


Introduction

The Japanese tax system has undergone drastic changes during the postwar period. Immediately after the end of the Second World War, it was completely overhauled under the American influence, particularly by tax recommendations of the Shoup Mission in 1949. Thereafter, however, the process of building a Japanese-style tax system began by the government initiative in the 1950s and 1960s, when it was reorganized to promote the growth of the Japanese economy. Since the oil price shocks of the 1970s, there have been some further attempts to reform the tax system as a whole in view of the changed circumstances experienced by the Japanese economy. The Japanese tax system is characterized by several unique features in comparison with those of other major industralized countries. The most distinctive feature is that the tax burden is among the lower third in this group: in the ratio of total tax revenues of GDP, Japan comes third-to-last of the 24 OECD countries, the lowest ranking apart from Turkey and the USA. A second feature is the heavier reliance on direct taxes: nearly 70 per cent of total taxes is collected from individual and corporate income taxes. Third, there had been no broad-based consumption tax during the postwar period until 1989, and indirect taxes came solely from selective excise taxes. Japan was the only country that did not rely to some extent on revenues from the value added tax. A fourth feature is the centralized collection system of the tax structures under a unitary nation. More than 60 per cent of total taxes is collected by the national government, and as a consequence the workings of intergovernmental fiscal transfers become important. A principal aim of this book is to describe systematically these aspects of the developing Japanese tax system. Based on a great deal of empirical evidence, the main features of tax policy issues and reforms are clarified. Major issues are divided among four parts, each having three, four or five chapters. Part I provides an overview of Japan's tax development, dating back to the prewar period. Chapter 1 begins this review with a brief explanation of general characteristics that are thought to be of basic importance to an understanding of the Japanese tax system. In Chapter 2, the historical background is investigated in relation to an empirical analysis of tax structure development and the impact of the Shoup tax proposals. Chapter 3 sheds light on the strategies the Japanese government has adopted concerning tax policy during the postwar period. Chapter 4 was added in the second edition to attempt an in-depth analysis on major issues of tax administration in view of tax equity. Part II is devoted to an analysis of the fundamental framework and economic effects of the individual income tax. After a brief outline of basic structures is given in Chapter 5, four important aspects of policy issues relevant to the individual income tax are considered in Chapters 6-9 by using empirical analyses with statistical data: the phenomenon of tax erosion (Chapter 6), adjustment for inflation

xxvi

Introduction

(Chapter 7), the tax treatment of savings and investment (Chapter 8), and the effects on income distribution (Chapter 9). These studies constitute the primary part of the book. In Part III the main features of corporate income tax and capital taxes are explored. Basic principles of corporate taxation and its economic effects are dealt with in Chapters 10-11, focusing on differences between Japan and other major countries. Chapter 12 looks at unique aspects of Japan's inheritance and gift taxes with some statistical procedures. Chapter 13 is devoted to exploring the movement of both land issues and land tax reform, in which a new tax named the land value tax was created. Part IV considers several aspects of the indirect tax system. Chapter 14 treats past trends and the framework of the excise taxes which had important influences on past tax reforms. The process of introducing value added tax (VAT) under both the Nakasone and Takeshita administrations and its aftermath are carefully examined in Chapter 15, which covers the latest debates and modifications. Chapter 16 is newly added in this edition, and analyses the basic nature of environmental taxes with certain empirical results of a hypothetical carbon tax. Part V has been substantially rewritten for the third edition to clarify all issues relating to tax developments in the past several years. Chapter 17 begins with an explanation of the general background against which the tax system has been rebuilt with special reference to fiscal deficits and administrative reforms since the late 1970s. Problems of local taxation are described in detail with intergovernmental fiscal relations in Chapter 18. This chapter covers both sides of local government budgets in a broader scope. Finally, in Chapter 19 the book concludes with an appraisal of recent tax reforms, and of further reforms which might be worth pursuing in years to come. Observing the past trend of Japan's tax developments since 1993 when the second edition of this book was published, it can be seen that the Japanese tax system has come to a turning-point in its search for a better form. As a result of tax reforms undertaken in recent years, successive rounds of income tax cuts in relation to fiscal stimuli have substantially changed the tax structure. While reliance on direct taxes has been reduced, indirect taxes have begun to increase their relative share in total revenues. This trend might be continued with population ageing in the new century. Needless to say, there are substantial parts of the present tax system that do not work satisfactorily and could not be made to do so even with great difficulty. From a practical point of view, switching from direct to indirect taxes could represent the first of a series of steps towards a better tax system. In this respect, I am sure that this book will be of great help. In writing the first edition, I basically relied on information that was correct as of December 1988, when the Takeshita tax bill was hotly disputed between the LDP and opposition parties in the legislative process of the Diet. The second edition was written by using data available to me as of December 1992. For this edition I have used data as of March 1999.

Parti

Overview


1

General Characteristics of the Tax System The purpose of this chapter is to provide a brief overview of the tax system in Japan. In general, a tax system presents certain significant features which have a close bearing on the workings of the economy in which it operates. What, then, are the general characteristics of the Japanese system? To answer this question, three important areas must be covered before beginning to consider major issues. First, a brief description is given of the evolution of the prewar and postwar tax systems. Second, three main features in tax levels and structure are explored. Lastly, several notable aspects of tax process and administration are examined as a preliminary step for further analysis.

HISTORICAL BACKGROUND

Modernizing the Tax System The origin of a modern tax system in Japan can be dated to 1887, when the national government instituted an income tax. Although this took a truly modern form only in 1940, Japan is thus counted as a pioneer in the use of the income tax. However, since the Japanese economy was at that time still underdeveloped, the income tax played only a minor role in total national tax revenues (say, 1.5 per cent in 1888). Before the income tax became predominant in the tax system as a whole, the main revenues for the national government were raised first from a land tax, and then from indirect taxes. In fact, the land tax accounted for the largest share of national taxes until 1908, after which time the primary source was revenues from indirect taxes, mainly on alcoholic beverages and tobacco. Only after 1935 did income tax on individuals and corporations became the most important single source of total revenues. In short, before the Second World War the Japanese government relied mainly on indirect taxes, deriving more than two-thirds of its total revenue from them. In 1940 an overall tax reform was carried out to prepare for the wartime economy. The whole tax system was thoroughly overhauled, resulting in the modern tax system, based mainly on direct taxes. Although individual and corporate incomes had been taxed together by a single form of income tax, in 1940 separate taxes were imposed for each type of income. Since then, individual and corporate income taxes have coexisted in the tax system. The individual income tax was a schedular tax, under which different sources of income were levied by different tax rates. It was supplemented by a progressive

4

Overview

comprehensive tax which applied to an individual's aggregated income above a specific amount. On the other hand, the corporate income tax was imposed on corporate income at a flat rate of 18 per cent. Moreover, the commodity tax was introduced in 1937 mainly to collect revenue for wartime expenses, and the tax on alcoholic beverages was also simplified in 1940. The relative share of indirect taxes, however, began to decline as a result of the 1940 tax reform. Evolution of the Postwar Tax System The process of developing the tax system in postwar Japan was initiated by the USA. In 1947, several important reforms were undertaken under the influence of the US occupation authorities. The schedular tax on individual income was replaced by a unified tax on an aggregate basis with graduated tax rates. Furthermore, a turnover tax, which was levied on the basis of the sales amount at every stage of transaction at the rate of 1 per cent, was enacted in 1948 to collect necessary revenues. In 1949 a tax mission headed by Carl S. Shoup came to Japan with the task of reorganizing the tax system as a whole. The Shoup Mission recommended a tax plan intended to achieve a complete overhaul of the Japanese system. Essentially, the Shoup recommendations placed more importance on direct taxes, mainly income taxes on individuals and corporations. Thus, the entire tax system was fully reconstructed producing epoch-making change. However, the ideal tax system achieved by the initiative of US influences was of temporary duration: many of the taxes were abolished or modified soon after their enactment. Hence it is necessary briefly to describe the post-Shoup evolution of four major taxes.1 First, and most importantly, comprehensive income taxation has been replaced by a combination of a comprehensive tax and a schedular tax. This hybrid system was produced as a result of modifying the global income tax approach of the Shoup proposals. For example, instead of aggregating most incomes with progressive tax rates, some incomes (e.g. capital gains or interest income) are not now subject to global income taxation but are taxed at reduced flat rates, separate from other incomes. This special treatment is due to a number of tax concessions intended to stimulate saving and investment and to improve the welfare level among specific taxpayers. Second, the corporate income tax was a split-rate system until March 1990, but it has now become a uniform one in which a single rate is imposed on a whole corporate income. The old system was quite similar to that used in West Germany, in which retained profits and dividends are taxed at different rates. Also, numerous special tax measures have made the corporate income tax extraordinarily complicated. Third, an accession tax on transfer of wealth at death proposed by the Shoup Mission was replaced by a combination of inheritance and gift taxes. Fourth, in contrast to the general trend of modifying direct taxes, the whole system of indirect 1 For a general discussion, see e.g. Pechman and Kaizuka (1976), MO1; Tax Bureau (199)), and Aoki (1986).

Table 1.1

Source of tax revenues in Japan and seven major countries, 1995 (%) Japan

Income Individuals Corporate Social security contributions" Goods & services General Specific5 Death and gift Property Other Total

USA

UK

Canada

France

Italy

Sweden

36.6 21.4 15.2

45.8 36.3

36.9 27.4

45.9C 37.3

17.6 13.9

30.1 27.3

35.1 26.2

41.4 35.3

9.4

9.5

8.1

3.7

2.8

8.7

6.1

36.3 15.1

25.1 17.9

5.2 9.9 1.9 9.7 0.4

8.0 9.9 1.0

17.7 34.7 19.0 15.7

16.8 25.5 15.2 10.3

45.7 27.3 17.4

32.0 27.3 13.9 13.4

31.3 24.3 15.1

0.0 10.5

0.0

0.6 9.9 0.2

1.3

9.9 0.8 4.4 4.2

39.4 27.8 17.3 10.5

0.3 2.5 —

0.2 5.5 —

9.2 0.2 2.7 0.1

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

10.2

* Include taxes on payroll, if any. Include taxes on specific goods and services and those on use of goods and performance of activities. 0 Unallocable items are omitted. Notes: Tax revenues include social security contributions, as well as national and local taxes. Source: Calculated from OECD, Revenue Statistics ofOECD Member Countries, 1965-1996.

b

Germany

6

Overview

taxes has remained unchanged for a long time since the time of the Shoup recommendations. There had been no general consumption tax in Japan until April 1989, and indirect taxation depended mainly on selective excise taxes. As a result of the evolving postwar tax system, what significant characteristics can we observe from the present tax system of Japan as compared with those of other major countries? As seen in Table 1.1, the Japanese system has several features different from those of other countries.2 Japan depends more on revenues from taxes on income, and places greater emphasis on the corporate income tax, which is the highest among OECD countries. Social security contributions (equivalent to the payroll tax) play an equally important role in raising revenues, in spite of heavier reliance on income taxes. The relative share of social security contributions has increased very sharply during the past decade or so. Among major advanced countries, Japan is the only one that did not impose a general consumption tax until 1989, and so its relative share was at the lowest level in 1995 even after the adoption of VAT. Thus, taxes on goods and services derive from specific excise taxes (e.g. alcohol, tobacco, or petrol taxes). Minor sources of revenue are obtained from death and gift taxes, while the property tax, a main source of local government revenue, occupies a relatively higher share in the totals. In 1995, of total tax revenues collected in Japan, 36.6 per cent came from individual and corporate income taxes, 36.3 per cent from social security contributions, 15.1 per cent from taxes on goods and services, 9.7 per cent from the property tax, and 1.9 per cent from inheritance and gift taxes. The USA raised total tax revenues in relatively similar proportions from each source. Japan's tax structure and that of the USA, with their heavier reliance on income taxes and social security contributions, should be compared with those of many European countries where a much larger share of revenues is obtained from taxes on goods and services.

MAIN FEATURES OF TAX LEVEL AND STRUCTURE

Low Tax Burden It is difficult to judge whether taxpayers in any one country are burdened to a greater or lesser degree than those in other countries. Since many taxpayers believe that they are personally overburdened and overtaxed, we need some objective yardstick. Perhaps the best tool is an international comparison, which is frequently used by economists. The relevant question is, Is the Japanese tax system required to raise more revenue than are tax systems in similar countries overseas? 2 OECD (1987, 62) classified OECD member countries into three categories geographically, by using the percentage distribution in question: those in (1) northern and central Europe, which rely on general consumption taxes; (2) OECD non-Europe, which rely on income (individual and corporate) taxes; and (3) Mediterranean Europe, which rely on social security contributions and consumption taxes. In accordance with this classification, Japan and the USA belong to category (2).

General Characteristics of the Tax System

1

One measure of the tax burden is provided by tax levels i.e. the ratio of tax revenues (including social security contributions) to GDP. Table 1.2 lists this ratio for 24 OECD countries for the years 1970, 1980, 1990, and 1995. Although the level and timing of increases change from country to country, the highest tax levels are to be found, as expected, in the Scandinavian countries, followed by other European countries, such as Belgium, France, and Luxemburg. In this comparison, Japan came third to last of 24 countries in 1970. In fact, apart from Spain and Turkey, Japan was ranked lowest among the major advanced countries. The same holds for 1980. However, this ratio in Japan rose, exceeding that of the USA and Australia in 1990, but it was still ranked among the lower group around the 30 per cent level. In 1995, it had declined again to the 20 per cent level, reflecting successive tax reductions, although its rank remains the same.

Table 1.2 Tax levels in OECD countries: tax revenues" as a percentage of GDP, selected years

Denmark Sweden Belgium Finland France Luxemburg Netherlands Austria Norway Greece Italy Germany New Zealand Canada UK Spain Switzerland Ireland Portugal Iceland Australia Japan USA Turkey

1995

1990

1980

1970

51.3 49.7 46.5 46.5 44.5 44.0 44.0 42.4 41.5 41.4 41.3 39.2 38.2 37.2 35.3 34.0 33.9 33.8 33.8 31.2 30.9 28.5 27.9 22.5

48.7 55.6 44.4 45.4 43.7 43.4 44.6 41.0 41.8 36.5 39.2 36.7 38.1 36.5 36.4 34.4 31.5 34.8 31.0 31.4 30.8 31.3 26.7 20.0

45.5 48.8 44.4 36.9 41.7 42.0 45.2 40.3 42.7 29.4 30.4 38.2 33.0 31.6 35.3 24.1 30.8 33.8 25.2 29.2 28.4 25.4 26.9 17.9

40.4 39.8 35.7 32.5 35.1 28.0 37.1 35.7 39.3 25.3 26.1 32.9 27.4 31.3 36.9 16.9 23.8 31.0 20.3 27.0 24.2 19.7 27.4 12.5

" Social security contributions are included. Source: As Table 1.1.

8

Overview

As is evident from such an international comparison, the Japanese tax system imposes a lower burden on taxpayers, as does that of the USA, among major industrialized countries.3 Perhaps this is one of its most salient features, which deserves wider attention.

Heavier Reliance on Direct Taxes Tax structure is another important element of the tax system of a country. Past trends in the variation of tax structure are seen in Table 1.3, in which we compare the percentage distribution of tax revenues (excluding social security contributions4) by source at both national and local government levels. The dominance of income taxes since 1950 was maintained and even accelerated until fiscal 1990. In 1990 approximately 70.7 per cent of total national taxes was collected from individual and corporate income taxes, and 63.7 per cent of local taxes came from the same sources. In contrast, the relative importance of taxes on consumption has continued to decline for the past decade or so. Taxes on wealth, which are used more prevalently by the local governments, increased their relative share until 1990. However, substantial changes have recently taken place in both national and local taxes, switching from income taxes to taxes on consumption. The Japanese often use the ratio between direct and indirect taxes3 as a crude measure of investigating a change in tax structure (simply referred to as the 'direct-indirect taxes ratio') in Japan. In Figure 1.1 this ratio is depicted for national and local governments, respectively for selected years over a 60-year span. Obviously, the relative weight of direct taxes moved upward until 1990. In particular, the shift from indirect to direct national taxes in the 1980s progressed markedly since the prewar period. It was widely acknowledged in Japan that the reliance on direct taxes was excessive. By definition, taxpayers felt the burden of direct taxes more strongly than other taxes and were inclined to avoid and evade them. Consequently, many Japanese taxpayers had complaints about the fairness and reliability of the tax system. Such an excessive reliance on direct taxes was partly responsible for the tax reforms, and as a result the relative share in direct taxes began to fall in the 1990s.

3

Kay and King (1986, 224) also attempted to classify selected OECD countries into three group: (1) Holland and the Scandinavian countries, (2) other Western countries, and (3) a somewhat heterogeneous group including Japan and the USA. 4 Revenues from social security contributions are generally not included when studying tax policy issues in japan. In fact, tax data do not include social security contributions, mainly for two reasons: (1) social security contributions are not treated as part of the general account of the national government, and (2) they are administered mainly by the Ministry of Health and Welfare, not the Ministry of Finance. Thus, we are likely to neglect them in analysing tax issues because of the lack of combined data. 5 At the national level, direct taxes are composed of individual and corporate income taxes and inheritance and gift taxes, and the remainder are all classified as indirect taxes including miscellaneous ones The same holds true for local taxation.

Table 1.3 Sources of tax revenues, by source, selected years Fiscal year

National taxes Taxes on income Individual Corporate Taxes on consumption Taxes on wealth Total Local taxes Taxes on income Individual Corporate Taxes on consumption Taxes on wealth Total a

1985

1990

1995

2000a

69.5 38.1 31.5 25.2 5.3

70.1 39.4 30.7 21.9 8.0

70.7 41.4 29.3 22.0 7.3

60.5 35.5 25.0 29.4 10.1

56.5 36.9 19.6 36.9 6.6

100.0

100.0

100.0

100.0

100.0

100.0

55.6 20.1 35.5 24.5 19.9

54.8 26.3 28.5 20.0 25.1

57.1 27.6 29.6 19.2 23.7

58.1 28.7 29.4 17.1 24.7

63.7 31.1 32.6 12.3 24.0

52.7 30.5 22.2 14.7 32.6

45.9 28.5 17.4 21.7 32.4

100.0

100.0

100.0

100.0

100.0

100.0

100.0

1950

1960

1970

1975

54.4 38.8 15.6 43.4 2.2

53.6 21.8 31.9 42.2 4.2

64.3 31.2 33.0 30.9 4.9

67.2 37.8 29.4 26.8 6.0

100.0

100.0

100.0

44.9 38.9 6.0 18.1 36.9

49.6 15.4 34.1 23.3 27.1

100.0

100.0

1980

Preliminary figures. Note: The classification of taxes in this table is rearranged according to the OECD criterion, but social security contributions are excluded. Source: Tax Advisory Commission, Tax Report on a Sweeping Tax Reform (in Japanese), Oct. 1986, pp. 18-19; and MOP, Primary Statistics of Taxation (Zeisei Shuyo Sanko Shiryoshu), Feb. 2000.

10

Overview

Fig. 1.1 Trends in the relative shares of direct and indirect taxes, 1934-1998 Source: As Table 1.3.

Centralized Tax Collections Attention should also be paid to the relative proportion of national and local taxes. Japan is centralized to a degree comparable to that of the UK, and local governments are given only limited fiscal responsibility and powers over their fiscal activities. Under a centralized fiscal system, unitary taxation is generally prevalent throughout the country (see Chapter 18). This is quite different from the decentralized systems of federal nations such as the USA, Canada, and Australia. Table 1.4 summarizes the relative share of total tax revenues of the national and local governments for selected years since 1950. In 1998 the total tax ratio (not including social security contributions) amounted to 24.5 per cent, of which 14.5 and 9.6 per cent were allocated respectively to the national and local governments. More than 60 per cent of total taxes is collected by the national government, but its relative share has steadily declined for the last three decades. This ratio of tax share is of course not applicable to the allocation of public expenditures to different levels of government, in accordance with the workings of intergovernmental fiscal transfers. It is also interesting to compare collection costs (i.e. administrative costs in the public sector) at two different levels of the government (see Chapter 4 for an international comparison). As seen in Figure 1.2, collection costs of national taxes constantly moved at a much lower level than those of local taxes during the period 1965-98, although two lines show the declining trends over the long run until fiscal 1990. It is thus conjectured that tax administration at the national level has been performed more efficiently by enlarging automation and computerized procedures with an unchanged number of staff. The problem lies in the costs of collecting local taxes by prefectural and municipal governments. The collection of local taxes costs over twice as much as that of


11

Table 1.4 Tax revenues as a percentage of national income, by level of government, selected years Fiscal year

1950 1955 1960 1965 1970 1975 1980 1983 1984 1985 1990 1995 1998

Level of gover nment National

Local

16.9(75.1) 12.8(71.1) 13.6 (70.8) 12.4 (67.9) 12.7(67.5) 11.7(64.0) 14.2 (64.1) 15.0 (63.3) 15.3(63.1) 15.1 (62.7) 18.2 (65.2) 14.5 (62.0) 14.9 (60.7)

5.6 (24.9) 5.2 (28.9) 5.6 (29.2) 5.9 (32.1) 6.1 (32.5) 6.6 (36.0) 8.0 (35.9) 8.7 (36.7) 9.0 (36.9) 9.0 (37.3) 9.7 (34.8) 8.9 (38.0) 9.6 (39.3)

Total

22.5(100.0) 18.0(100.0) 19.2 (100.0) 18.3 (100.0) 18.8(100.0) 18.3 (100.0) 22.2 (100.0) 23.7(100.0) 24.3 (100.0) 24.1 (100.0) 27.9(100.0) 23.3 (100.0) 24.5 (100.0)

Notes: Social security contributions are excluded. Figures in parentheses are percentage distributions. Source: As Table 1.3.

Fig. 1.2 Trends in tax collection costs, 1965-1998 Note: Costs are calculated per ¥100. Source: National Tax Administration, Annual Reports (Kokuzeicho Tokei Nenpo).

national taxes. Why are costs so much higher in the collection of local taxes? One reason often mentioned is that tax offices of local governments perform less efficiently; tax revenues cannot be collected in proportion to the larger numbers of tax officials. Given this situation, it is often proposed that the two administrative

12

Overview

mechanisms be merged to improve the efficiency level of tax administration and to lower total costs of collecting taxes as a whole. In summary, the most salient features of the Japanese tax system are that (1) the tax burden is relatively lower among major industrialized countries; (2) heavier reliance is placed on direct taxes; and (3) tax collection is centralized.

TAX PROCESS The Formulation of Tax Policy The Japanese have 'generated a unique system to create a general consensus on tax policy. For convenience, we distinguish the Japanese type from the US-Western type in formulating tax policy. In Figure 1.3, the two types are drawn as upper and lower pathways. Each process starts from a government inquiry which is initiated when the government admits the necessity for promoting any tax policy or reform. In the US-Western type, a task force or committee is usually established by the president, prime minister, or other authorities; several tax experts are appointed as members, and they are headed by an influential chairman.6 The task force or commission of inquiry generally undertakes a review of the whole tax system for five or six years, and then publishes a report to recommend a plan of structural tax reform for specific goals of tax policy. This report is virtually independent of political powers, at least while it is in preparation. In this sense, it can be considered a public report of an independent task force or commission. After the report is published, the government begins the legislative process in the congress or parliament. It should be emphasized, however, that there is a low degree of implementation of inquiry recommendations by the promoting government (see Kay 1987,67-9). Needless to say, the proposals contained within the report tend to define both gainers and losers, which causes a great deal of controversy by a variety of vested interests. In particular, it is very difficult to win public support for potentially unpopular policies, including tax increases. In general, governments lose their inclination to tackle the political problems of structural reform, and in turn the initial objective is likely to be frustrated. Therefore, under a tax formulation of the US-Western type, the possibility of implementing the tax reports is much poorer than might be expected. In contrast, the Japanese government uses the tax commission on a more formal basis; that is, as a Tax Advisory Commission of the prime minister. After the prime minister submits the inquiry to the Commission, it starts the discussion with a review, public hearings, and analyses. In the case of long-range tax reform proposals, after an intensive study for a relatively shorter period, say less than three years, the 6 Typical examples include the Carter Commission in Canada in 1962, the Ashrey Committee in Australia in 1972, the Irish Commission in 1980, or the US Treasury case in 1984 (see Kay 1987). The Shoup Mission can be regarded as an example of this category.


13

Fig. 1.3 The taxation process under Japanese and US-Western systems * Liberal Democratic Party (in Japan).

Commission presents its recommendation to the prime minister, not to the general public. What is of primary importance in the Japanese method is that tax recommendations differ from those undertaken in the US—Western type. As will be explained shortly, the Tax Advisory Commission contains representatives from many vested interests. Despite serious conflicts between each group, reconciliation is conventionally made before the Commission recommends a tax programme, since political pressure groups can resolve differences. The Commission does encounter considerable difficulty and resistance in making reconciliations, and as a result a great number of political compromises are involved in the final tax proposals. This makes the contents of proposals difficult to understand readily. The greatest benefit from the Japanese method is that almost all inquiry recommendations can be relatively easily approved by the Diet, chiefly because political problems have already been resolved by the major interests. In short, the extent of implementation is much higher, in spite of many defects in the theoretical level of tax proposals. This process, however, has changed with the intervention of the tax committee of the ruling party, the Liberal Democratic Party (LDP), as is shown in Figure 1.3. The LDP has its own tax committee, which makes tax recommendations every year, and it has become much more influential politically. The Tax Advisory Commission's recommendations, which were adopted virtually intact two decades ago, have often been altered by the different views expressed within the LDP. Thus, the Tax Advisory Commission is inclined to exclude engineering features of tax proposals (e.g. percentage of tax rates or the level of deductions) in order to avoid any possible conflicts with the LDP Tax Committee. Usually, its recommendations are merely made as a fundamental argument for tax reforms without supplying any concrete figures necessary to construct tax structures. The detailed framework of tax changes is politically decided by the LDP, not the Tax Advisory Commission. A typical example will be seen in the case of the land holding tax (see Chapter 13).

14

Overview The Role of the Tax Advisory Commission

The Tax Advisory Commission has long played a central role in assisting the formulation of tax policy and reform, although it is often criticized that its role has tended to be less important. It is impossible to understand the Japanese tax process without a detail knowledge of the Commission itself. The Tax Advisory Commission was established by the government in 19537 to review the whole tax system and to formulate annual tax changes as well as long-run tax policy. One of Japan's most prestigious committees, it consists of 30 regular members and about the same number of supporting members, all of whom are appointed by the prime minister. Usually, tenure for members is for three years. Members of the Commission are selected from a variety of people, including academics, tax experts, tax lawyers, journalists, former government officials of each ministry, representatives of big and small businesses, labour union leaders, and so on. Because of this diverse membership, the Commission can arbitrate conflicting issues among the vested interests who greatly affect tax policy. The Commission's major objectives include the formulation of both short-run and long-range tax proposals. Annual tax revisions are prepared from a short-run standpoint by the Commission when the budget is compiled. Likewise, the Commission is responsible for the formulation of long-run policy on the structure of the tax system. Both types of tax proposals have had a substantial impact on general attitudes towards changes in the tax structure. High-level staff of the tax bureaux in both the Ministry of Finance (MOP) and the Ministry of Home Affairs (MOHA) are deeply involved in the decision-making process of the Commission as its secretariat. In fact, they appear to have much influence on the Commission's recommendations in two respects: first, they are responsible for the recruitment of members to assist the Commission and the government; second, the necessary data are for the most part limited to that provided by the tax bureaux. Therefore, the role of the staff of the tax bureaux is of great importance to the functioning of the Tax Advisory Commission. These staff members are the most enlightened and imaginative of government officials, and if necessary they could formulate tax proposals by themselves without the aid of any outsiders. In fact, before recommendations are actually made by the Commission, tax proposals are frequently discussed with the MOF and other ministries. Why do the government and tax bureau staffs seek the assistance of the Tax Advisory Commission? Generally speaking, a consensus is reached in the movement between the government and the Commission. While the Commission is sometimes seen as 'hiding behind the government',8 the government's views also have been 7 This was not a permanent establishment, but a temporary committee whose term lasted for a couple of years. After 1960 it was replaced by the long-standing committee that still operates today. 8 We use the term 'Kakuremino' in Japanese. The government sometimes uses the commission as a screen in order to divert attention from its real motives.


15

altered as a result of conferring with the Commission's members. Thus, tax recommendations prepared by the Commission, which signify the government's approval, are instrumental in smoothing the way for the enactment of tax reform. Without them, the government would encounter obstacles in having any tax proposals implemented, chiefly because it is difficult to obtain public support of tax changes. Strictly speaking, the Commission has not played a crucial role in short-run tax revisions for the past two decades or so. Attention should also be paid to the more influential role of the LDP tax committee, as argued before, which usually attempts to present tax proposals in the annual tax process. Around the end of the calendar year, the recommendations of the Tax Advisory Commission are transmitted to the prime minister for budgetary preparations; those of the LDP tax committee follow almost simultaneously. The final decisions are made by the cabinet based on these two recommendations. In the case of opposed views, it became the general tendency for the LDP Tax Committee to predominate over the Tax Advisory Committee in the tax process. Indeed, the powerful LDP committee often succeeded in significantly modifying the proposals of the Tax Advisory Commission. Consequently, as noted above, the Tax Advisory Commission has tended to limit its proposals to the basic arguments in favour of the desirable direction for the longterm tax system and not to raise politically subtle issues of tax policy (e.g. any judgement regarding the value added tax rate). Also, it is of great importance to note that deliberation in the Commission has become open to the press since autumn 1998. This is a result of the recent movement to make official information open to the public.

2

Background of the Japanese Tax System In order to study the main features of a nation's tax system, it is important to explore its basic background from an historical perspective. The main issues in this chapter are considered from two distinct viewpoints. The first approaches tax issues in terms of the long-run trend of tax structure development. The second examines the starting point of the postwar tax system, with particular attention given to the Shoup tax reform.

A LONG-TERM

VIEW OF TAX S T R U C T U R A L

DEVELOPMENT

Generalizations of Change If we take a long and broad view back into time, our attention is drawn to the salient features of the Japanese tax structure during the process of economic development. In other words, we are led to consider how tax structures appear to change during the transition from a traditional society to a modern one, and whether there is any theory to tie together common threads among tax systems in varying stages of development. Deriving an answer to these questions would be important to achieving an understanding of the basic framework of the present tax system and to elucidating long-term changes in the size and composition of tax revenues. In fact, many studies to date have attempted to investigate tax structure change from a similar point of view. Based on broad empirical findings with special emphasis on the size and structure of tax revenues over time, in past studies generalizations have made an attempt to incorporate these findings into a consistent framework (see e.g. Hinrichs 1966). The purpose of generalizations is to determine whether there is some economic law which, as Engels's law confirmed for consumption spending, reveals a relationship between tax revenues and the development process.1 The basic nature of this study is to examine whether or not such generalizations can be applied to the Japanese experience. What is of greater significance to this study is how the tax structure changes at different stages of economic development. Consequently, an ideal approach would be to examine the same countries at different levels of development, using time-series data rather than cross-sectional data.2 However, this approach presents This chapter is based partially upon Ishi (1978,1987). 1

The necessity of constructing such a law is stressed by Thorn (1967, 19—20). Most of the published studies, however, have focused on tax structure development in developed and developing countries, using cross-sectional data. This approach is the only one feasible in many 2

Background of the Japanese Tax System

17

several difficulties. For one thing, data for many countries are not available for the entire period of their economic development; for another, few countries have completed all phases of economic development. Nevertheless, there are some eases (e.g. the UK, the USA, Germany) for which one can analyse the entire process of tax structure development from a long-term standpoint.3 The present study represents an effort to extend past analyses of developed and developing countries through an examination of Japan's experience during the 113-year period of 1885-1997. Japan's experience is notable for two reasons. First, Japan is the only non-Western nation to have succeeded in attaining the level of economic development enjoyed by Western nations. This means that Japan has passed through all the levels of development within the past 100 years. Second, in the process of economic development, Japan has not oriented itself too closely to the European pattern. It can be argued that the economic development of Japan was different from that of Western countries in many respects. Thus, Japan should prove an illuminating case-study of tax structure change.4 We must now consider what empirical evidence in Japan's case supports the generalizations of tax structure development. There are two generalizations presented in past studies which will be investigated here: 1. that the size and composition of tax revenues tend to change over time, reflecting structural changes in the economy; 2. that forces (e.g. social, political, and cultural) other than changes in economic structure also govern the determination of tax shares. Here we seek the similarities or dissimilarities of Japan's experiences in terms of these generalizations.5 Fortunately, such an investigation is now feasible, through use of the long-term statistical data prepared by the Hitotsubashi University group.6 A Model of Tax Structure Change Based upon empirical analyses, past studies have developed a general theory to explain and predict the size and composition of tax revenues in the process of cases, chiefly because there is a lack of reliable historical series data on GNP, its components, price levels, and other related data in most countries. Yet a cross-sectional approach is necessarily very rough and appears to have several defects. See e.g. Oshima (1957), Martin and Lewis (1956), Williamson (1961), Lewis (1963). 3 Some useful hypotheses emanated from such works as Peacock and Wiseman (1961), and Musgrave (1969). 4 For past studies of Japan, see Hinrichs (1966, 49), Musgrave (1969, 137). Recently Chelliah (1986) has attempted to compare the Indian experience with that of Japan with special reference to Ishi (1978). 5 More detailed analysis has already been attempted in Ishi (1978). For a general discussion on the Japanese economy, see Minami (1986). 6 The Hitotsubashi group has been engaged in the lengthy project of estimating economic statistics of Japan from 1868; see Ohkawa et al. (1965-88). Fourteen volumes, including National Product, Capital Formation, and Government Expenditure, have finally been published. In particular, these three volumes

18

Overview

economic development. Generally speaking, it is difficult to generalize the development of tax structures in different countries and time periods because tax structure change at first sight appears as a multi-coloured fabric containing numerous patterns. However, a general theoretical pattern of tax structure change emerges from the empirical and historical observations. This pattern was presented in the form of an 'heuristic model' by Hinrichs (1966, ch. 6), which he derived from a crosssectional analysis of countries at different levels of development. Hinrichs uses an heuristic device to establish a typology and an average picture to which individual cases can be compared. It is helpful for the purpose of our investigation to compare the tax structure development of Japan with such a general pattern. According to Hinrichs's model, tax structure generally develops through the shifting of the relative weights of land taxes, indirect taxes (divided into taxes on foreign and internal sectors), and income taxes as the economy advances through each phase of development (traditional, transitional, modern). Figure 2.1 illustrates the pattern of tax structure development in Japan for comparison with Hinrichs's heuristic model (1966, 99).7 Let us first pay attention to the lower part of the figure. Here, three characteristics of Japan's experience are illustrated. 1. A shift from land taxes to indirect taxes, and further to income taxes, can be seen over time. This is almost identical to Hinrichs's ideal type of tax structure change. 2. However, taxes on foreign trade played no major role in the initial stage of Japan's development—a sharp contrast to the experience of many other countries. The chief reason for this is that tariff autonomy was not achieved until 1899, and even after that date only partial revision of tariff structure was undertaken (see Yamazawa 1975, 41). 3. There is one more dissimilarity: the trend of internal indirect taxes has been to decline, not rise, until recent years in postwar Japan. This reflects the sharp growth of individual and, until recent years, corporate income taxes in the rapidly growing economy. Next we turn to the ratio of taxes to GNP (T/Y) at all levels of government; this curve is depicted in the upper part of Figure 2.1. The level of tax share provides an important indicator of the role of government and taxation during development. It represents the fiscal capability of the government to meet the increased need for public services. Also, it measures citizens' power or capacity to bear the burden of taxation. (For more extensive discussion, see Bird 1964, 303-4.) The tax-GNP ratio reveals six major swings, although four of them are not perfect. The first swing, in 1885-95, shows a large tax share of around 8-12 per cent, are essential to my analysis. For other major statistical materials, see Bank of Japan (1966), Economic Planning Agency (1992). 7 In his model, both the expenditures line and the expenditure-revenue gap are depicted clearly, in addition to each line of individual taxes as a percentage of GNP.


Fig. 2.1

19

Changes in the level and composition of Japanese tax revenues, 1885-1995 (in

terms of GNP) Note: While tax shares cover all level of government, four ratios of each tax to GNP are limited to the scope of national government.

the same level as that reached in developing countries during the postwar period.8 The principal taxes were those collected on land and alcoholic beverages, and these were used to meet extensive government needs while the country was still at a low level of national income.9 The second swing peaks around 1910 and bottoms in 1918. New or increased taxes on income, alcoholic beverages, tobacco, sugar, textiles, and beverages, as well as custom duties, all contributed to the upswing. The third swing, falling between 1918 and 1932, is a plateau at a fairly low level, reflecting the depression in the 1920s. The main revenue sources at that time were taxes on alcoholic beverages and income. In the 1930s, the tax revenue share rapidly increased in the fourth swing with no downswing following. Direct taxes on individual and corporate incomes, including an excess profits tax instituted in 1935, rose tremendously during this period. Fifth, the highest peak emerges immediately after the end of the Second World War, followed by a relatively stable level of T/Y after 1952-3. After postwar reconstruction was complete, the economy's rapid growth and the government's decision to stress private-sector growth (partly through a tax reduction policy) lowered the tax share, and it remained between 14 and 17 per cent until about 1970. Lastly, from that time the tax share began to rise rapidly, apart from a sharp fall in the mid-1970s. This reflects the expansion of fiscal deficits without a tax cut during the 1980s, reaching its highest in 1990; subsequently it began to 8 9

For data on the developing countries, see Musgrave (1969, appendix table 6). Oshima (1965, 386—7) notes the high level of Japan's tax burden in international comparison.

20

Overview

decline as a result of the recent slump. Until the tax share moves upward in the future the six swing (trough-peak-trough) will continue. Tax Structure Change and Economic Development The major question to be answered here is, How have the size and composition of tax revenues been determined? More specifically, what kinds of factors are most important in explaining the variation of both tax level and structure? Various studies have shown that the ratio of tax revenues to GNP increases with economic development and that the structural change of taxation reflects different levels of development between developed and developing countries. In what follows, attention is given to tax structure change during development. Tax structure, of course, is greatly affected by institutional, economic, and socio-political factors.10 Indeed, it can be regarded as a product of the historical interaction between such forces. Although changes in tax structure can, in principle, influence these economic, social, and political forces over time, such changes have generally tended to be more a determined than a determining factor. If one were to emphasize the passive nature of the evolving tax system, one could even say that the major determinant of tax structure change is the structural change in the economy itself during the process of economic development. There are several specific variables which have been employed in past studies to explain the size and structure of tax systems. Among them, the following three factors are pertinent to the case of Japan: 1. y/N

2. Ag/Y

3. M/For (M+X)/Y

= per capita real GNP (y is real GNP in the prewar period at 1934—6 prices and in the postwar period at 1960 prices, and Nis total population); = agricultural products' share in GNP (Y is nominal GNP, and Ag is output of the primary sector, i.e. agriculture, forestry, and fishery); = openness of the economy (M and X stand for imports and exports, respectively).

Factors (1) and (2) are viewed as indices of economic development, while (3) measures the size of the foreign trade sector, which is used as an alternative to (1) and (2). All these indicators have been found to be significant variables in explaining variations in 777 and the composition of tax revenues, although a fuller explanation could be made by introducing additional factors. Table 2.1 shows the correlation of these three variables with tax shares. The simple correlation coefficients during the prewar period are as high as would be expected from past empirical observation of developing countries on a cross-sectional basis, and we find that the estimated results of the postwar era are reasonable, too. 10

For the socio-political aspect of the problem, see Deutsch (1961).


21

Table 2.1 Correlation of per capita real income (y/N), openness (M/Y, (M+X)/Y), and agricultural products' share (Ag/Y) with tax shares (J7Y), 1885-1997 Period

Sample size

Correlaticin coefficient y/N

M/Y

(M + X)/Y

Ag/Y

Prewar (1) 1885-1944 (2) 1885-1909

60 25

0.281* 0.337

-0.030 0.458*

-0.081 0.563**

-0.307* -0.377

Postwar (3) 1951-1997

47

0.674**

0.283

0.504**

-0.465**

* Significant at the 5% level. ** Significant at the 1% level.

This, therefore constitutes a rough sketch of the interdependence between the variables. The results may be summarized as follows. 1. There is a significant correlation between y/N, Ag/Y, and T/Y, with reasonable positive or negative signs in the prewar period of 1885-1944. 2. Negative correlation exists between Ag/Y and T/Y for the postwar period of 1951-97, and y/Nand T/Y are still significantly related. 3. Openness provides a better index than the other two variables mentioned above for the first 25 years of the prewar period. In addition, it becomes more significant in explaining the variation of T/Y for the postwar period. These findings are similar to the results of cross-sectional analysis that have been obtained from regression or correlation between T/Y and various development variables." Another approach can be taken to pursue the same analytical objective. In addition to the investigation of T/Y in Table 2.1, an equally important question to ask might be, Is there any systematic relation between the source of tax revenues and the level of income, or is tax structure influenced by institutional factors relatively independently of economic development? In this question, the focus shifts to changes in the composition of the tax structure. Tax revenues Tn are disaggregated into three sources: 1. land taxes—Te; 2. indirect taxes, including profits of government monopolies12—T,-; 3. income taxes on individual and corporate income—T y . 11 Another way of explaining the variation of T/Yuses the hypothesis that the existence of the E-R gap (i.e. the discrepancy between government expenditure and revenue) necessarily stimulates a concomitant increase in tax burdens. This hypothesis is tested by using statistical procedures in Ishi (1975,213-16). 12 If the share of foreign trade taxes is assumed to be high (which is not the case in Japan), it should be treated as one of the dependent variables, distinguished from domestic indirect taxes as in Lewis (1963).

22

Overview

These figures, however, are limited to national government tax revenues because the data for classifying local taxes in such a manner are lacking.13 Therefore, for the dependent variables, we let TtITn stand for the relative share of land taxes, TjTn for the indirect tax share, and Ty/Tn for the income tax share, respectively. (For the same procedure, see Williamson 1961, 51-2.) There are three reasons for stressing the relative importance of each tax share. First, it appears that the effects of economic development upon the tax structure are more a function of institutional change than they are inherently an economic matter. Thus, greater emphasis should be placed on the various sources of tax revenues, as their change reflects the institutional setting of the tax system, which is relatively independent of changes in economic structure. Our 'share-approach' seems to capture the effect of institutional factors. Second, there is a high correlation between T/Y and each share component of total national government taxes during the time period in which each constitutes the principal share of the total (e.g. land taxes for 1885-98, indirect taxes for 1899-1935, and income taxes for 1936-90).14 Third, the relatively poor results of Table 2.1 must be reconsidered; they may be the result of aggregating all the taxes and using GNP as the divisor in tax revenues. A simple regression model was constructed with regard to the development of tax structure. As shown in Table 2.2, almost all the coefficients of determination are statistically significant with a few exceptions. The regression coefficients of all the independent variables in all the equations are also statistically significant. The conclusions from these estimations are as follows. 1. As y/N (an index of development) increases, TtITn and Tj/Tn decrease. The relative importance of these two taxes tends to decline over time, a result found in other studies. 2. The declining trend of Tt/Tn is influenced by the decreasing share of AgfY. 3. 'Openness' (M/For M+X)/Yor can explain the variation in T,-/Tn in the prewar period,15 while it is not significant in the postwar era. 4. Ty/Tn is dominantly affected by y/N. Obviously, the relative share of income taxes tends to rise in the course of development.

13 In general, national government taxes dominated the tax composition of local taxes in the prewar period, since the latter was levied virtually as a surtax on national government taxes. A more complete coverage of tax revenues would not alter the conclusions presented here. This is inferred from data available for the prefectural level, not including the lower levels of local government (i.e. city and town). 14 The correlation coefficients between T/Y and each relative share of national government taxes are: 0.662 of land taxes for 1885-98, 0.512 of indirect taxes for 1899-1935, and 0.492 of income taxes for 1936-90. All the coefficients are significant at the 1 % level. 15 As is evident from the positive coefficient of the indirect taxes equations in the prewar period, the 'openness' of the economy expanded the tax base for indirect taxation, through spillover effects which stimulated consumption, commercialism, transportation, etc. On the other baud, the sign of the 'openness' coefficient in the postwar period is insignificant. It appears that 'openness' was no longer effective in increasing the indirect tax base at this level of economic development.

Table 2.2 Regression equations for tax sources as a function of economic development variables, 1885-1944 and 1953-1997 Prewar period (1885-1944)

Postwar period (1953-1997)

Land tax

Indirect tax

Income tax

The generalized least square (GLS) method was used to generate all these equations. R2 is the coefficient of determination adjusted for degrees of freedom, ** and * indicate significance at the 1 and 5% levels, respectively, DW is the Durbin-Watson statistic, and the values in parentheses are f-statistics.

24

Overview

Ay 5. Real growth rate — is added only in the postwar case. It is significant to explain y

the changes in Tf/Tn and Ty/Tn with opposite signs each. Negative interrelation Ay between TyITn and — appears to reflect the large scale of income tax reductio y in a growing economy. The purpose of the present study has been to examine the development of the tax structure in Japan during the period 1885-1997 in the context of generalizations made in previous published studies. The evidence of Japan presented here seems to be in full support of past generations. Since time-series analyses of tax structure change are relatively rare in the literature, Japan's case-study is especially important if it provides support for these empirical generalizations. Although the findings indicate some divergence from those patterns, many similarities can be found in Japan's experience. In particular, great emphasis should be put on the fact that Hinrichs's heuristic model fits Japan's case with only minor exceptions. Other Determinants of Tax Structure Development In addition to the economic factors underlying tax structure development, there is another key factor in determining the growing ratio of tax revenue to GNP and the 'proper' tax structure composition. In view of the results of various studies, attention should be directed towards the cultural-political preferences for adopting a specific size and composition of the tax system. When a country has reached a high income level and a large government sector share of GNP (say, between 20 and 40 per cent), these preferences appear to become much more important than at lower income levels. For instance, the level at which the government sector share settles between 20 and 40 per cent is likely to be determined by differing commitments towards 'security and defence' and/or 'welfare policy', rather than by change in economic structure (see Hinrichs and Bird 1963, 433). In postwar Japan, such ideological commitments are probably among the determinants of tax structure development. What is of great interest is the low level of the tax share in the postwar period. As we have seen in Table 1.2, Japan has the lower level of tax burden among major advanced countries during four selected years. Japan's low ranking remains unchanged today, even though the tax share has rapidly increased to reduce the gap with other countries. This feature, peculiar to postwar Japan, needs to be explained. Among explanatory factors, of most importance is the difference in the level of military expenses between the prewar and postwar years. If military expenses are shown relative to GNP (or GDP) during selected fiscal years (the relevant table is omitted), the ratio rises drastically during the war period (e.g., 8.44 per cent in 1894-5 (the Sino-Japanese War), 22.97 per cent in 1904-5 (the Russo-Japanese War), and 27.98 per cent in 1941-4 (the Second World War)). Attention should, however, be directed towards the very low figures for the postwar period in comparison with those of prewar Japan: the percentage of postwar military expenditures as a part of has been less than 1.00 per cent. The low level of military spending i


25

also apparent in international comparisons of the military expenses-GNP ratio: between 1971 and 1984 this ratio averaged 6.3 per cent for the USA, 5.1 percent for the UK, 3.5 percent for West Germany, and 3.8 per cent for France. The 1985-91 averages showed 5.7 per cent for the USA, 4.5 per cent for the UK 2.5 per cent for West Germany, and 3.1 per cent for France, while Japan's figure was only 0.97 per cent. During 1992-96 the ratios averaged 4.8 per cent for the USA, 4.5 per cent for the UK, 2.0 per cent for Germany, 3.6 per cent for France, and 1.2 per cent for Japan. Thus, the gap between Japan and other major countries had shrunk substantially. In addition to the absence of military expenses, reference may also be made to the low level of Japan's welfare commitments. The percentages of social welfare transfers to GDP in 1970 in various countries were: 16.8 per cent in France, 12.7 per cent in West Germany, 8.6 per cent in the UK, 7.6 per cent in the USA, and 4.7 per cent in Japan. These ratios tended to increase in each country: by 1990 the US ratio had increased to 11.1 per cent, the UK's to 11.6 per cent, Germany's to 15.2 per cent, and France's to 21.1 per cent, in comparison with Japan's 10.8 per cent. Japan still had the lowest ratio, although this showed a markedly faster rise during the previous decade. In 1996, Japan's ratio rose to 13.5 per cent, exceeding 12.9 per cent in the USA, but both still remained lower than those of major Western countries: 14.2 per cent in France, 18.5 per cent in Germany, and 23.4 per cent in the UK.16 Cultural-political factors also appear to have some influence on the pattern of tax composition in postwar Japan. It is widely acknowledged that there are typically two tax styles in the world, reflecting the cultural determinants of any tax system: one is the direct tax style (or at least, an even split between direct and indirect taxes), and the other is the indirect tax style. Japan has preserved the former style in the postwar era, to a large extent because of the American influence during the occupation period, i.e. the tax recommendations of the Shoup Mission in 1949. Had there been no Shoup Mission, Japan's tax system might have moved towards a different type of system with a greater share of indirect taxation. THE DAWN OF THE POSTWAR TAX SYSTEM:

THE

SHOUP TAX REFORM

The Shoup Mission As described above, cultural, political, and social forces are equally as important as economic ones in determining the nature of a nation's tax system. In this regard, we must stress the significant role of the 1949 Shoup Mission in shaping the style of the tax system in postwar Japan. (For a more expanded discussion, see Ishi 1987, Shoup 1989.) The mission, headed by Professor Carl S. Shoup, visited Japan in April 1949 at the request of the Supreme Commander for Allied Powers (SCAP).'7 They stayed 16

These figures are derived from OECD estimates. In regards to the process of inviting the Shoup Mission, see Moss (1948), MOF Tax Bureau (1977a). The Mission was composed of seven members, including S.S. Surrey, W.S. Vickrey, J.B. Cohen, H.R. Bowen, R.R. Hatfield, and W.C. Warren. 17

26

Overview

in Japan for about four months, investigating the Japanese tax system as well as its economic and social background. As a result of intensive studies, they presented to SCAP the Report on Japanese Taxation by the Shoup Mission in August 1949. The Japanese government then attempted to reorganize the country's entire system of national and local taxes in accordance with Shoup's recommendations, and the new system went into force with the next supplementary budget in 1949. What were the main reasons for the visit of the Shoup Mission? Two points should be stressed. First, conditions of the postwar Japanese economy were chaotic. Rampant inflation was wreaking havoc among business accounts, tax assessments, and revenue collection. The tax system was truly in a mess. Second, there was mutual understanding between the USA and Japan as to the necessity of overhauling the tax structure and its administration after implementing the 'Dodge Line'.18 The Shoup tax reform was not the first reform of the Japanese tax system during the occupation period, but earlier tax reforms had proved far from satisfactory (see Shavell 1948a,fo; Cohen 1949). Consequently, an authoritative plan was absolutely required to revise the tax system in order to attain greater equity and efficiency. It was not simply a political expedient. In the earlier years of the occupation, the tax system had developed certain defects which the Shoup Mission was asked to remedy. The most important defects were as follows. 1. Individual income taxes had become higher and more progressive, with low exemption and broad coverage. Heavier tax collection overwhelmed tax administration and weakened tax morale. Revenues were collected by the 'goal system', in which each tax office was assigned a goal or quota. 2. Corporate income and excess profits tax rates were high, and businesses were not permitted to adjust depreciation allowances to allow for drastic price hikes. 3. Tax sources remained concentrated at the national government, although a great volume of public functions was allocated to local governments (prefectures and municipalities) in the name of strengthening 'local autonomy'. In addition to trying to remedy these defects in the tax system, the Japanese government negotiated with SCAP to achieve a substantial tax reduction (especially of individual income tax) on behalf of taxpayers. SCAP, however, did not want to verify the necessity of such tax cuts.19 Thus, tax reduction became a crucial issue before the Shoup Report was published. Fortunately, economic conditions, which had been improving in the period before the beginning of the Shoup Mission, favoured its recommendations; inflation 18 The 'Dodge Line' is the name of the anti-inflationary programme conceived by Joseph Dodge, an American banker, who was invited by General MacArthur in 1948 to evolve a formula for arresting runaway inflation. Dodge aimed at stabilizing the yen value by establishing a true balance in the consolidated budget and eliminating the subsidies that had been the prime cause for the continuing growth of fiscal deficits. It was very successful in halting inflation, although a great depression ensued. See Cohen (1950), Yamamura (1967). 19 For a Japanese view, see Economic Stabilization Board (1949). However, Dodge completely disagreed with the request for tax reductions from the Japanese side; see SCAP (1949).


27

had been halted as a result of Dodge stabilization policy of 1948. Nevertheless, there were still a number of difficulties in the wake of the Shoup Mission. For instance, in view of anti-tax and anti-inflation sentiments in Japan, it was necessary for the Mission to include tax cuts without unbalancing the budget.

Basic Framework of the Shoup Report Like the Carter Report in Canada and the Meade Report in the UK, the Shoup Report has been highly evaluated by many tax experts and has received considerable attention over a long period of time, especially for its theoretical and logical consistency (see e.g. Hicks 1951). The Shoup Report had epoch-making significance in the history of Japanese taxation. In contrast to the tax reports noted above, most of the Shoup recommendations were put into practice in Japan, although a movement towards modifying them began very soon after implementation. The contribution of the Shoup recommendations to Japanese taxation should not be underestimated; throughout the postwar period, the report has served as the benchmark of a well designed tax system whenever Japan discusses tax reform. The Shoup Report is generally considered to contain new and advanced views long cherished by Shoup, Surrey, Vickrey, and other tax experts. The Shoup Mission attempted to reconstruct the Japanese tax system along lines generally familiar to American tax experts, and a number of novel features were designed to make the Japanese tax system 'the best tax system in the world' (Shoup Mission 1949, vol. 1, p. ii) in a so-called experimental manner. Three points characterize the Shoup Report as a whole. First, the fundamental aim of the report was to establish a permanent and stable tax system in Japan over the long term.20 Needless to say, the goal of the Shoup Mission was to create a modern tax system based on direct taxation. Alternatively, it would have been possible to choose another form of taxation, based upon the indirect tax. Indeed, the Japanese government preferred the indirect tax system, mainly because it had favoured indirect taxation during the prewar period.21 Second, the tax proposal was intended as a single integrated plan. The Shoup Mission felt very strongly that the whole plan would be destroyed if any parts were 20 At the time, the Shoup Mission seems to have thought that their proposed tax system should be preserved for more than ten years. Shoup referred to this point in retrospect when he came to Japan in 1972, see MOP Tax Bureau (1972). Furthermore, long-term tax reform was possible because there was not an immediate need for revenue. Accordingly, the Mission recommended a tax plan that would bear fruit only in the long run. 21 The Shoup Mission gave two reasons for not recommending an indirect tax system. (1) Such a system could raise the required revenue, but it would perpetuate gross inequities among taxpayers, dull the sense of civic responsibility, keep the local governmental units in uneasy financial dependence on the national government, and give rise to undesired economic effects on production and distribution. (2) Moreover, the difficulties in obtaining fair and efficient administration of the tax laws, and a high degree of compliance by the taxpayer in Japan, should not be seen as inevitable.

28

Overview

eliminated. On this point, the Shoup Report argued very strongly that 'What we are recommending here is a tax system, not a number of isolated measures having no connection with one another' (Shoup Mission 1949, vol. 1, p. ii). By referring to 'a tax system', they placed strong emphasis on the interlinkage among individual taxes. The Mission carefully considered the effect of individual taxes on one another. For example, individual income and net worth taxes, and succession and real estate taxes (the land and house tax), were investigated jointly. Furthermore, the principle of full inclusion of capital gains and losses in income taxes was closely related to the inter-relationship between individual and corporate income taxes. It is evident from the basic idea of the Shoup Report that the whole income tax structure would be seriously weakened without full inclusion of capital gains and losses. Third, among various tax criteria, most importance was placed on tax equity throughout the whole report. In the press interviews immediately after the Shoup Mission came to Japan (19 May 1949), Shoup himself emphasized the importance of restoring fairness in the Japanese tax system as one of five objectives for his tax reform. Thus, the basic philosophy in support of tax equity is repeatedly argued: A tax system can be successful only if it is equitable, and the taxpayers must realize that it is equitable.... We have often encountered surprise at the emphasis we place on the search for equity. But no one remains in the tax field for long without realizing that nothing he recommends will stand up unless it meets the test of fairness in the distribution of the tax burden. (Shoup Mission 1949, vol. 1, p. 16.)

Turning to major parts of the recommendations, several points are worth noting. First, of greatest importance is the fact that the progressive and broadly based individual income tax was retained as the mainstay of the Japanese tax system. In retrospect, the individual income tax proposed by the Shoup Mission really provided an ideal form of comprehensive tax base with a single 'progressive' rate system in the true sense of the term. Of course, this was the first time that anything of the kind had been attempted in an Asian country, although Japan had experienced schedular income taxes since 1887 on a smaller scale.22 Second, a major concern of the Shoup Mission's recommendations was the improvement of tax administration, especially of the income tax. For instance, withholding taxes at sources from wage and salary incomes and the self-assessment with universal filing of returns were recommended. Furthermore, the use of the 'blue form' for tax returns was especially suited to encourage the proper keeping of accounts, particularly in the case of small businesses. Obviously, these efforts to improve assessment and administration were indispensable for an effective and equitable tax system.23 22

U.K. Hicks argued that the income tax was introduced under rather primitive conditions: 'It is not the sort of economy in which one might, on a priori grounds, expect to be able to recommend a very large sphere for income and profits taxes' (Hicks 1951,200-1). See also Kimura (1952). 23 Tn addition, the Shoup Mission attached great importance to the provision of regular training for tax assessors and collectors, the establishment of training colleges, and the improvement of pay and conditions of the tax administration.

Background of the Japanese Tax Syste

2

Third, a general revaluation of all assets (i.e. land and fixed capital) was recommended as a prerequisite for the adoption of the Shoup tax plan. Since the value of the yen had depreciated on the order of 200- or 300-fold since the prewar period, such a process of revaluation would stimulate private capital accumulation. In addition, it was proposed that a tax of 6 per cent be imposed on the appreciation in the written value of all assets, although it would consist almost wholly of paper gains in terms of book value. Fourth, great emphasis was placed on the reform of local finance in order to educate the Japanese in democratic citizenship. The provision of a fiscal framework for 'local autonomy' was an important element of the Shoup proposal. The general recommendation of the Shoup Mission was that local powers and duties should be substantially increased; in particular, priority should be given to the lowest of the three levels of government (i.e. the municipalities). For this purpose, local governments were given new tax resources (e.g. property tax and value added tax), and at the same time intergovernmental transfers were overhauled to implement a new scheme for the equalization of local budgets called the Equalization Grant Scheme. Finally, the Shoup tax plan contained several novel fiscal experiments. The Mission suggested these experiments for Japan to try without the benefit of any large-scale applications in other countries. Special attention was given to three of these—the net worth tax, the accession tax, and the value added tax—although they were minor in size. The Aftermath of the Shoup Tax Reform It is rare in history that a tax report is enforced in practice. The Shoup Report was almost wholly enacted in both the 1949 supplementary budget and the 1950 budget. The Shoup tax reform is interesting to tax reform experts as a case-study of the accomplishments of a tax mission in a short period under ideal conditions. In seeking the necessary conditions for a successful tax reform, there seem to be a number of relevant factors to learn from the impact of the Shoup Mission. From the very beginning, however, some of the Shoup tax plans were criticized as being too theoretical to be carried out, given the state of socio-economic development in postwar Japan. No doubt, the Mission thought of tax reform primarily in terms of US practice and experience. This was apparent in such matters as the treatment of capital gains taxation or the emphasis on 'local autonomy'. Accordingly, modifications to the Shoup tax system were implemented shortly after 1950. Two tendencies emerged from these modifications. One was the revival of the old system: equity was sacrificed for the convenience of incentives and administration. The other was the reduction of the tax burden of firms, especially big businesses. The goal of this trend was to give priority to the restoration of the postwar economy and the promotion of capital accumulation. Tax equity, on which the Shoup Mission put utmost priority, began to be replaced by incentives as the criterion o

30

Overview

taxation.24 As time has passed, essential features of the Shoup plan have been 'eroded' or 'patched and tattered' by the later tax reforms of the Japanese government. (See Table A2.1 in the Appendix to this chapter.) The most symbolic modification of the Shoup system occurred with the repeal of full taxation on capital gains from sales of securities in 1953. It has often been pointed out that the Shoup Mission was well aware of the shortcomings of the US tax system in respect of capital gains taxation, and that they tried to introduce a better treatment of capital gains as an experiment in the Japanese situation. As noted earlier, the Mission repeatedly insisted on the need for capital gains taxation. In spite of their strong appeals, the actual capital gains tax mostly disregarded proceeds from security sales since 1953, partly because the difficulties of administration were great, and partly because the promotion of capital accumulation became a national goal. In addition, the innovative tax devices of the Shoup Report disappeared from the Japanese tax system after brief or no trials. The net worth and accession taxes were abolished in 1953 because of inadequate revenues. The value added tax was not even brought into operation; its enactment date was postponed twice, and it was finally repealed in 1954 (see Ito 1950; Bronfenbrenner 1950). When the Japanese government departed from the Shoup system, its departure was not in the direction of further experimentation, but towards a return to prewar traditions and practices which it considered particularly suitable to the Japanese economic situation.25 Thus, the tax innovations advocated in the Shoup Report were disregarded. Necessary Conditions for a Successful Tax Reform The modifications by the Japanese government of the Shoup proposals were drastic in their later consequences. Consequently, it may be argued that the Shoup reforms achieved only a partial success, largely because the modifications gradually made the tax system more inequitable and complicated. In spite of these drawbacks, the Mission's contribution in reconstructing the postwar tax system in Japan was considerable. Throughout the postwar period, the Japanese system has retained substantial features of the Shoup framework. Thus, the Shoup tax reform can be considered one of the most successful tax reforms in the world. 24 In general, these modifications of the Shoup tax reform were accepted as inevitable by the Japanese. For instance, Hanya Ito commented on this point: 'However, it is to be observed that the tax system in practice is a product of historical development depending on the social, economic, and political conditions of time and place. It would not be wise to condemn such a course of events merely from the standpoint of abstract theory'(Ito 1953,382). 25 See Bronfenbrenner and Kogiku (1957, part 1, 241). Ito (1953) also argues that, 'Judging from the development of tax reforms these last three years, Japanese taxation is showing a tendency to restore the old system which was in effect before the Shoup recommendation' (p. 358).


31

To conclude this discussion, we shall seek to explore the necessary conditions for successful tax reform, a rare event in history. Particular attention should be paid to the following three points. The first and most important point is that the foundations on which the Shoup Mission was to erect the new tax structure in Japan had experienced a complete break with the past by the events of the Second World War and postwar inflation. Prewar values had become irrelevant as a basis of postwar taxation, and any former injustices could be disregarded in view of the sweeping change that had affected all values. Furthermore, the changes recommended by the Mission were far less drastic than the overhaul of values that the Japanese had experienced during the war. These circumstances facilitated the work of the Mission and encouraged them to experiment with innovative tax reform. To use Feldstein's terms, a 'tax design', rather than 'tax reform', was implemented (Feldstein 1976, 77). The second point is that circumstances greatly favoured the Shoup Mission. Seldom has any advisory mission received instructions as broadly defined as in the case of the Shoup Mission.26 In addition, the Mission's arrival coincided with the introduction of a national programme for the redirection of the entire economy, which had become feasible since economically chaotic conditions had settled down to a considerable extent. Thus, the Mission was implicitly given leeway for wide and sweeping changes in forming the tax plan. Of utmost importance was the fact that the Mission was supported by SCAP and General MacArthur. Reflecting this support from the highest authorities, the recommendations of the Shoup Mission received high-priority consideration from the Japanese government. These circumstances put the Mission in an exceptionally favourable position as to the enactment of recommendations. The Japanese government and Diet acted with vigour in accepting nearly all of the proposals. Third, from a professional point of view, the Shoup Report itself has been rated as one of the best tax reports, and of the highest quality. The academic specialists of the Mission first followed the basic principles of taxation as developed in textbooks. Thereafter, they tried to link these theoretical considerations with the institutions of Japan, although they were handicapped by unfamiliarity with the Japanese tax system. It is often pointed out that the Shoup proposals not only are logical and well balanced in theory, but also can stand the test of practicability to some extent. In the case of tax missions to other countries, when results generally fall short of expectations, the host countries often tend to refuse to enforce the proposals wholeheartedly. As far as the Shoup proposals are concerned, this tendency was at a minimum in Japan. The Japanese people would not have accepted the proposals if the Shoup Mission had prepared a set of recommendations doomed to failure. Instead, the

26

Shoup himself mentioned in retrospect that General MacArthur did not interfere with the work of the Shoup Mission, and refrained from making any special requests or issuing any orders when they were writing the Report; see MOF Tax Bureau (1972).

32

Overview

Japanese understood that they were to benefit from some of the best thinking on tax issues which the Shoup Mission provided.27 It is obvious from the above discussion that the conditions under which the Mission created a 'tax design' for Japan were exceptional. They would never reappear in the future. 27 For the discussion of the relationship between the Shoup Mission and Japanese taxation, see e.g. Sundelson (1950); Bronfenbrenner and Kogiku (1957a, 1957, pt. 2); Dodge (1949); Moss (1949).


33

APPENDIX

Table A2.1 Shoup Mission recommendations and their modifications (major items only) Shoup recommendation I. Individual income tax (1) Type of tax: to be single on an aggregation basis, not schedular

(2) Top bracket rate, to be lowered to 55% from 85% with 8 income brackets (3) Exemption, deduction and credit personal exemption, dependants, and earned income deductions to be changed (4) Capital gains and losses: to be included in or deducted fully from income, and treated as form of fluctuating income with averaging system (5) Interest income: source collection (separately from other income) to be abolished

Japanese legislation (1949-56) Carried out Subsequent Japanese moves towards schedular income tax, e.g. special treatment of bank interest, dividend, retirement income, etc. Carried out Top bracket rate raised to 65% (1953) when net worth tax repealed Carried out Medical deduction (1950), life insurance payment deduction (1951), and social insurance payment deduction (1953) Carried out Both gains and losses from security sales disregarded (1953); instead, security transfer tax introduced Carried out, but old system revived at 50% rate (1951); rate cut to 10% (1953); bank interest income made tax free (1955)

II. Corporate income tax (and asset revaluation) (1) Corporate income tax rate: not to Carried out be increased above 35%; no Raised to 42% (1952); lowered to progression to be imposed 35% on first ¥500 000 of income, 40% on the remainder (1955) (2) Excess profit tax. to be replaced Carried out (3) Revaluation procedures: land and Four revaluations carried out depreciable assets to be revalued (1950, 1951, 1953, 1954) as of 1 July 1949 Last revaluation made compulsory for depreciable assets of large-scale corporations; farm land not to be revalued until sold (4) Tax on revaluation gain: to be set at Carried out 6% of gain; payable in instalments over Repealed in connection with 1954 3 years for depreciable assets; payable for revaluation non-depreciable property at time of sale III. National indirect taxes (1) Turnover tax. to be repealed as soon as revenues permit (2) Textile consumption tax: to be repealed

Repealed as of 1 Jan. 1950 Carried out

34

Overview

Table A2.1

cont'd

Shoup recommendation (3) Alcoholic beverage excises: (a) Rates to be raised to pre-May 1949 level, with further increases as local alcohol taxes are repealed (b) Alcohol consumption tax to be repealed

(4) Tobacco taxes (monopoly profits) prices of cheapest (rationed) cigarettes and cut tobacco to be reduced (5) Commodity taxes: rates to be reduced (6) Minor excises to be repealed: Soft drinks Travelling (3rd class) Registration and stamp taxes

Japanese legislation (1949-56) Partially carried out

Carried out

n Tobacco prices increased by local tobacco consumption excises (1954) Substantially carried out Never carried out Soft drinks included in items subject to commodity tax Carried out Never carried out

IV. Local taxes and intergovernmental fiscal relations (1) Value added tax: enterprise tax to Never carried out become an income-type VAT Effective rate postponed annually exclusively at prefectural level through 1953; tax repealed (1954). Enterprise tax remained in effect (2) Inhabitants' tax (local income tax): (a) Allocation: to be reserved for Carried out municipalities Made partially prefectural (1954) (b) Variable element: to be based Carried out on income alone, not on National income tax used as property or social status as standard, takes form of 18% surtax formerly. Base may be (i) income tax, (ii) taxable revenue, or (iii) the difference (ii) — (i) (c) Corporation to be exempted Carried out Corporations made taxable (1951), tax taking form of 15% surtax on national corporate income tax; surtax lowered to 12.5% (1952) (3) Property tax: (a) Allocation: to be reserved for Carried out municipalities Made partially prefectual (1954) (b) Coverage: to be extended to Carried out depreciable assets as well as real property, but not to inventories

Background of the Japanese Tax System Table A2.1

35

cont'd

Shoup recommendation (c) Assessment: to be based on capital rather than on rental values of property

(4) Equalization Grant. (a) To be established as replacement for shared taxes and partial subsidies, but approximately double the total amount of former (b) Distribution among local units to be based on needs for major activities (c) Distribution element of income taxes to be eliminated

(5) National subsidy of local activities: Methods to be changed: (a) 100% subsidies to be replaced by national government performance (b) Partial subsidies to be replaced by Equalization Grant (except for promotional purposes)

Japanese legislation (1949-56) Carried out, with capital values determined by selling prices and volume of business Rate originally set at 1.6% of base, subsequently lowered twice, in 1955 to 1.4% Carried out Equalization Grant system abolished (1954); instead, new tax shared programme introduced Carried out for 90% of grants (1951-4) Carried out 20% of personal income, corporate income, and alcohol taxes distributed to local governments (1954); percentage raised to 22% (1955)

Partially carried out (in some cases, subsidy reduced instead) Never carried out.

Note: This table is partially based on information provided by Bronfenbrenner and Kogiku (1975).

3

Two Strategies of Postwar Tax Policy

The postwar Japanese tax system emerged from a blueprint proposed by the Shoup Mission, but the blueprint was promptly altered by the Japanese government. In the 1950s and 1960s, when the Japanese economy saw very rapid and sustained growth (see Patrick 1970; Patrick and Rosovsky 1976), a Japanese brand of tax system gradually developed as part of postwar growth policy. As a result many of the Shoup tax proposals were replaced by Japan's own tax measures. What is of particular interest to outside observers is how significant the contribution of postwar tax policy was to the remarkable record of economic growth. This chapter is devoted to the study of two aspects of the performance of tax policy since economic reconstruction: (1) tax incentive policies and (2) annual tax reductions. In addition, attention is paid to a renewed movement aimed at restoring tax equity and neutrality in the second half of the 1980s, instead of relying on the older growth policies. The Japanese experience during the postwar period provides an interesting and important case-study, since it was different from those of many other advanced countries.

TAX INCENTIVE POLICIES

Two Strategies

Soon after the new tax system initiated by the Shoup Mission was enacted, modifications began to be proposed, mainly for two reasons. First, the new tax system proved too idealistic and far-reaching for the relatively conservative government. Second, the government wished to enable Japan to revive its postwar economy more rapidly; thus, it behaved in a more growth-oriented manner and used the tax system more effectively to promote private economic growth. It is not easy to describe briefly the special character of Japan's tax policy in the 1950s and 1960s.1 Special conditions have certainly existed in Japan for which the government has been able to create unique tax policies. In this connection, there are 1 Note that 'some of its particular features seem strange to a visitor from a country in which economic conditions and social attitudes are quite different. Japan's high rate of growth and moderate government expenditures (when compared with other developed countries) permit the Japanese to adopt tax policies that can well be envied elsewhere' (Pechman and Kaizuka 1976, 323).


37

two strains of postwar tax policy that are of interest to the outsider: 1. an incentive tax policy to achieve specific policy goals; and 2. a successive tax-cutting policy to maintain a lower level of tax burden. These two strategies were gradually adopted by the government to complement the actual performance of the economy and were sustained at least until the outbreak of the oil crises in the 1970s. If we extend our view to cover government expenditures, a third strategy of sound public finance should be added to clarify the whole image of government fiscal activity (see Ishi 1973, 59-60). Special Tax Measures Let us begin with the first strategy of creating tax incentives. Especially before the outbreak of the first oil shock in 1973, there was wide agreement in Japanese government and business circles that the tax system should be actively employed to promote economic growth. Based on tax incentive policies, a number of special measures were formulated to promote exports, private saving and investment, housing, technological development, environmental quality, etc. These usually included tax exemption tax-free reserves, and accelerated depreciation for individuals and corporations. Linked with Japanese-type industrial policies, such measures deserve attention as contributing factors to the rapid economic growth in the 1950s and 1960s. It is, however, very difficult to ascertain the effectiveness of using tax incentives for specific policy purposes. In fact, the evaluation of these policy effects has been controversial, subject to multiple difficulties. These difficulties are due mainly to the scarcity of quantitative studies concerning tax incentives on economic activities. In some cases, it is impossible to quantify the effect of these policies. We shall begin our discussion by considering the performance of tax incentive policies that the Japanese government adopted in the postwar period. Although there are a number of notable contributions in this field (see Komiya \966b; Pechman and Kaizuka 1976), it is worth reviewing the development of tax incentives in the system, covering a more recent period (see Kaizuka 1984). In the postwar period, numerous tax measures were introduced to stimulate economic growth and other related policy targets that had a high national priority.2 Such measures are included in the Special Tax Measures Law, which is separate from the ordinary income tax laws. This special law was formulated to prepare a list of most (though not all) of the incentive provisions applying mainly to individual and 2 The origin of special tax measures dated back to the prewar period when the Temporary Tax Measures Law was enacted in 1937. During the war these measures were substantially expanded in co-ordination with the war effort, and were continued in the form of the new Special Tax Measures Law after the war. Naturally, the number of special measures was sharply curtailed when the Shoup tax reform was carried out in 1950. Only a few measures survived at that time, including a tax-free reserve for bad debts and tax-exempt interest income of specific types.

38

Overview

corporate income taxes. However, it must be noted that the same type of incentive measures are often found in the ordinary income tax laws as well. In fact, at present, a number of significant tax benefits are contained in the ordinary income tax laws, rather than in the Special Tax Measures Law. Thus, it is rather difficult to distinguish those tax provisions that are considered special from those that are not. In what follows, I shall use the term 'tax incentives' chiefly to refer to those taxes that the Ministry of Finance (MOP) calls 'special tax measures'. These can also be termed 'tax preferences'. Each item included in the 'special tax measures' is drawn from both the Special Tax Measures Law and the ordinary income tax laws, in accordance with the specific definition of policy incentives adopted by the MOF.3 There is no comprehensive list of 'tax expenditures' for Japan comparable to that contained in the US budget. The scope of'special tax measures' seems to be narrower than that of'tax expenditures' in the USA in terms of tax incentive policy (see Office of Management and Budget 1986). One major aspect of the special tax measures was their proliferation, coming as a result of the political bargaining that plays an important role in the process of Japanese decision-making. Once a particular incentive was approved, people with similar needs requested similar tax preferences for themselves. This process was repeated many times. For instance, the introduction of a deduction on life insurance premiums emerged initially from the needs of life insurance companies. Once they received special treatment, marine and fire insurance companies also demanded a special deduction for fire and other casualty insurance premiums to maintain equal positions in the market. In this way, special tax measures proliferated in the 1950s and 1960s. (For a more detailed discussion, see Ministry of Finance 1991.) Significance of the Tax Incentives Is there any quantitative factor that can indicate how significant the tax incentive measures have been so far? For this purpose, a series of revenue losses caused by the special tax measures, which are compiled annually by the MOF, are of great use. Table 3.1 shows the trend of estimated revenue losses from the special tax measures in individual and corporate income taxes. From these estimates, it is obvious that the revenue loss resulting from the special tax measures was 13.2 per cent of total income tax revenues in the late 1950s, fell to 10.3 per cent in 1960, rose to 12.0 per cent in 1965, and then declined to 4.0-5.0 per cent in the 1980s. In the 1990s, however, it rose to 5.0-6.0 per cent once again. In a word, a long-term declining trend can be observed clearly throughout the postwar era. This trend reflects the 3 The definition of special tax measures was officially announced when the Tax Advisory Commission presented their tax report in 1976. In both the Special Tax Measures Law and the ordinary income tax law, a great deal of effort was taken to draw the distinction between incentive measures relating to specific policy purposes and the ordinary tax structure. The tax staff at the MOF assert that the distinction between the two is clear-cut, given their professional knowledge and understanding of the income tax law (see Tax Advisory Commission 1976, 3-4).

39

Two Strategies of Postwar Tax Policy Table 3.1 Comparison of estimated revenue loss from special tax measures and total individual and corporate income tax revenue, 1958-1998 Fiscal year

1958 1960 1965 1967 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 1997b 1998°

Total individual and corporate tax revenue (¥b.)

586.3 964.0 1 897.5 2597.6 4014.3 4995.3 5445.6 6718.3 9850.2 11 166.6 9610.2 11004.5 12144.6 15665.9 16657.9 19722.3 20802.8 21980.1 23 467.3 25404.0 27455.7 29917.8 33 247.9 36391.9 40374.8 44379.1 43344.4 36945.0 35824.4 32780.6 33250.5 33448.2 34288.0 35829.0

Revenue losses from special tax measures" Amount (¥b.)

% of income tax revenue

77.6 99.1 228.2 255.4 357.7 434.5 532.7 580.4 645.0 727.0 796.0 759.0 840.0 933.0 959.0 1 026.0 1 121.0 1149.0 1 193.0 1281.0 1525.0 1 562.0 1 528.0 1 694.0 1 836.0 2121.0 2355.0 2183.0 2097.0 2034.0 1 923.0 1 799.0 1 994.0 1 864.0

13.2 10.3 12.0 9.8 8.9 8.7 9.8 8.6 6.5 6.5 8.3 6.9 6.9 6.0 5.8 5.2 5.4 5.2 5.1 5.0 5.6 5.2 4.6 4.7 4.5 4.8 5.4 5.9 5.9 6.2 5.8 5.4 5.8 5.2

a

Includes only the items listed as special tax measures by the Tax Bureau, excluding revenue gains from the curtailment of corporate special and entertainment expenses. Preliminary figures. c Estimated on the basis of budget data. b

Source: tax revenues, from Ministry of Finance Tax Bureau (1998); revenue loss from special tax measures, from data presented to the Budget Committee, National Diet by the Tax Bureau, MOF.

40

Overview

fact that many of the special tax measures for specific purposes have been eliminated in the past decade to make the tax system more neutral and equitable. Objectives of the Tax Incentives The MOP classifies the special tax measures of individual and corporate income taxes into six different objectives: (1) promotion of saving; (2) promotion of environmental quality and regional development; (3) promotion of natural resources.; (4) promotion of technological development and modernization of industrial equipment; (5) strengthening the financial position of firms; and (6) other incentives. This classification, however, is not useful for economic analysis. Some attempt must be made to rearrange these into more meaningful categories. Table 3.2 summarizes the percentage distribution of revenue losses from the special tax measures, dating back to 1958, for which the relevant data are available. The first category, the promotion of individual saving and housing, is tied exclusively to the individual income tax. The other categories are related mainly to the corporate income tax. Throughout every time period, the first category has maintained the highest share. In addition, it has recently shown a trend towards an increasing relative share, because the other tax incentives in the corporate category have been substantially eliminated. The second-largest share of revenue loss arising from tax incentives is related to the promotion of business saving and investment. The tax devices used to promote these activities include tax exemptions an credits, accelerated depreciation, and tax-free reserves. These devices are generally used in particular industries or specific activities, and in many cases two or three measures are employed to promote the same objective. Third, the promotion of export and foreign investment has greatly declined in importance, although in earlier periods it was a major consideration. In the 1950s and 1960s, devices to promote export and foreign investments were prevalent as special deductions of export income from taxation and accelerated depreciation for export-oriented firms. These tax provisions were finally eliminated by the early 1970s. Today, remaining provisions for exports and overseas investment are very minor. Consequently, their relative share in Table 3.2 has continued to fall over the long term. The fourth category includes measures concerned with the promotion of environmental quality. In the early 1970s, the Japanese economy began to experience the side-effects of rapid economic growth, such as rampant inflation, urban congestion, and environmental pollution. Popular attention was focused on the prevention of air and water pollution in order to preserve the quality of the environment. Since that time, the share of tax devices relating to pollution control sharply increased, reaching a peak of 9.2 per cent in 1975. Typical measures include special additional depreciation and tax-free reserves for pollution control activities.

41


Table 3.2 Percentage distribution of the estimated revenue loss from special tax measures, by type of incentive, 1958-1997 (%) Fiscal year

1958 1960 1965 1967 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

Promotion of individual saving a housing (1) 49.5 45.6 59.9 61.0 54.6 47.0 44.5 44.7 49.6 49.4 44.6 47.0 48.2 48.1 49.7 53.6 57.4 58.0 57.3 54.6 59.0 59.8 57.4 59.3 58.2 60.3 61.8 61.2 61.7 61.8 66.0 62.3 63.1

Promotion of business s investment (2) 25.8 35.9 18.7 20.9 18.2 18.1 19.9 27.5 25.1 21.4 22.2 18.8 18.9 18.3 22.4 20.6 20.2 21.1 23.4 26.0 24.2 24.2 29.1 26.4 28.7 22.8 21.2 22.0 22.4 19.7 18.5 18.3 18.2

Promotion of export and f investment (3)

Promotion of environment q

Others

(4)

(5)

18.0 12.6 12.9 10.9 14.8 19.3 17.0 4.7 3.8 5.5 6.5 5.5 4.2 2.8 2.1 1.9 1.7 2.3 1.6 2.4 2.2 1.7 1.9 1.4 1.0 0.8 0.6 0.4 0.6 0.4 0.8 0.9 0.7

0 0 0.3 0.8 1.5 1.6 5.8 6.9 7.3 8.3 9.2 6.7 4.6 4.4 5.5 3.6 4.2 4.3 4.1 4.2 4.1 4.1 2.9 3.2 2.8 3.5 4.0 3.9 3.9 3.5 3.5 3.2 3.2

6.7 5.9 8.2 6.3 10.9 14.0 12.8 16.2 14.2 15.4 17.5 21.9 24.0 26.4 20.2 20.4 16.5 14.4 13.7 12.7 10.6 10.2 8.7 9.7 9.3 12.7 12.4 12.5 11.4 14.6 11.2 15.3 14.8

Note: The official classification was rearranged into five items as listed above. Source: As Table 3.1.

42

Overview

Finally, other miscellaneous measures have accounted for 10-20 per cent of revenue loss since the 1970s. Mostly, these consist of measures such as the special deduction of physicians' fees, which are not related to the corporate income tax. ANNUAL

TAX REDUCTIONS

Past Trends Turning to the second strategy, let us focus on the annual tax-cutting policy. The growing economy of postwar Japan automatically generated large increases in annual tax revenues. Given the fact that nominal GNP rose by an average of 15 per cent a year in the 1950s and 1960s, tax revenues were growing annually at a rate of more than 15 per cent (because the income elasticity of total taxes usually exceeds 1.0). If the tax system had remained unchanged, the government would have obtained fiscal resource increases annually in excess of 20 per cent. In Japan, these abundant revenues were never used to expand the size of the government sector as they have been in many other countries. In fact, the Japanese government adopted a tax-cutting policy to keep the ratio of total tax revenues (including both national and local taxes) to national income constant (i.e. at 20 per cent), rather than simply to increase public expenditures. This guideline was followed especially strictly during the period between 1955 and 1965, and as a result the ratio of total taxes to GNP remained relatively constant during those years. The government therefore had the luxury of providing successive annual tax reductions. It is widely believed that, if annual tax reductions had not been implemented, the individual income tax would have unduly overburdened taxpayers. Thus, tax reductions were very popular with Japanese taxpayers during the high-growth period, incurring envy in many other countries. Table 3.3 summarizes estimated annual tax changes since 1950. These estimates of tax decreases or increases are calculated as the effects of tax actions, using official economic forecasts, that were enacted at the beginning of each year. Before 1975, total tax revenues were reduced annually at both national and local government level with only minor exceptions. This reflects the fact that until the late 1970s a strong sentiment prevailed not to increase national and local taxes even as potential revenues increased. Interestingly enough, annual tax policy since 1976 has adopted the opposite stance towards tax increases in accordance with the new direction of fiscal consolidation. The trend of tax increases, however, was ended in fiscal 1987, and thereafter tax reduction policy was initiated once again by using increased revenues due to the 'bubble boom' until 1991. In the 1990s after the collapse of the bubble, fiscal stimuli packages led to successive rounds of large tax reductions to buoy up the depressed economy. Cuts in individual income taxes in a growing economy prior to the early 1970s occupied the lion's share of annual tax reductions. In comparison with the continuous long-term reductions in individual income tax, both the corporate income tax

43

Two Strategies of Postwar Tax Policy Table 3.3 Estimated annual tax changes, fiscal years 1950-1998 (¥bn.) Fiscal year

National taxes Total

1950 1952 1954 1956 1958 1960 1962 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 a

Local taxes

-206.8 -89.5 -16.9 -1.5 -37.3 6.6 -116.4 -118.2 -115.1 -310.6 -94.0 -68.0 -181.0 -205.5 -108.5 -4.6

-378.1 -1 115.0 -372.0 383.0 81.0 463.0 634.0 367.0 1531.0 308.0 7.0 157.0 187.0 131.0 -599.0 -1725.0 -298.0 -306.0 -4.0 473.0 101.0 -4386.0 36.0 -1 586.0 -164.0 -2092.0"

Individual income tax -135.8 -112.7 -31.4 -22.6 -6.3 0.0 -50.3 -74.5 -65.4 -158.3 -92.5 -125.1 -183.0 -288.7 -206.9 -32.2 -375.2 -1783.0 -186.0 0.0 -141.0 -12.0 80.0 22.0 -13.0 24.0

-1.0 -770.0 24.0 -30.0 -464.0 -1838.0 -30.0 -173.0 0.0 -7.0 -32.0 -3845.0 2.0 - 1 400. -82.0 -1 494.0

Corporate tax

-24.4 19.1 2.6 14.4 21.5 0.0 -1.3 -58.6 56.6 98.7 30.3 0.0 2.4 75.2 12.1 30.6 26.8 352.0 -6.0 115.0 110.0 49.0 218.0 371.0 640.0 324.0 36.0 414.0

194.0 51.0 -512.0 —1 112.0 -39.0 -83.0 1.0 29.0 38.0 88.0 42.0 26.0 16.0 259.03

Other direct taxes

13.9 -2.4 -2.9 0.0 -3.3 0.0 -2.4 -4.5 -0.5 -15.0 -3.1 0.0 0.0 0.0 -6.6 -12.1 -39.7 0.0 -298.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 -6.0 0.0 -2.0 0.0 -584.0 0.0 0.0 0.0 371.0 -150.0 -3350.0 -4.0 -182.0 0.0 -134.0

provisional figures

Source: MOF, Primary Statistics of Taxation (Zeisei Shuyo Shiryoshu), Feb. 1992.

Indirect taxes

-60.5 6.5

20.0 6.7 -6.3 6.6 -62.4 19.4 7.4 -38.6 31.9 57.1 -0.4 8.0 92.9 9.1 10.0 316.0 118.0 268.0 112.0 426.0 336.0 -26.0 904.0 -40.0 -28.0 519.0 -31.0 112.0 377.0 1 809.0 -229.0 -50.0 -5.0 80.0 110.0 21.0 -4.0 -30.0 -98.0 -205.0a

n.a. -27.7 -26.3 12.3 20.0 -11.8 -39.9 -55.7 -8.7 -53.4 -19.3 -20.7 -94.7 -96.6 -87.0 -98.1 -145.7 -113.4 -466.8 34.7 -53.4 77.7 181.5 213.2 153.7 54.8 56.0

31.4 89.4 42.3 117.5 -2214.9 -31.8 -107.8 -722.3 4.8 263.4 - 1 564. 19.2 -640.5 2.6 -806.3

44

Overview

and indirect taxes show varying changes over the years. Corporate income tax was increased more frequently, with successive tax increases beginning as early as 1969. Indirect taxes were also often lifted until 1970, chiefly because most of them were levied on a specific basis and periodically needed to be adjusted for the rise in commodity prices. This tendency, however, has been reversed since mid-1970. Other direct taxes, including inheritance and gift taxes, showed no significant variation as a whole except for 1975. Increases in revenue from corporate and indirect taxes were small relative to income tax reductions in the 1950s and 1960s. Consequently, tax policy during the high-growth period was characterized by net annual tax reductions, owing to the major impact of individual income tax reductions in total tax collections.

High-income Elasticity of Tax Revenues As mentioned, one of the most notable features of Japan's tax policy in the 1950s and 1960s was the government's annual tax reductions. Even though a substantial amount of tax revenues was cut annually, however, this did not reduce the size of the government sector. This fortunate circumstance can be explained largely by the rapid growth in revenue, which financed both tax cuts and expenditure increases during the postwar period. At this point, therefore, we must determine whether there are any significant features of the Japanese tax system that render its revenues especially responsive to income growth. The tremendous growth of Japanese tax revenues has been achieved automatically, owing to the highly elastic income taxes and the high rate of nominal GNP. Thus, frequent tax reductions, as shown in Table 3.3, were required periodically to avoid overburdening taxpayers by large revenue increases. To make this mechanism clear, it is necessary to investigate how responsive the tax system has been to income growth. In general, the income elasticity of tax revenues is used to measure the responsiveness of tax revenues to the growth of nominal GNP. The formula for elasticity

where Y= nominal GNP, and T = tax revenues. From the elasticity formula (3.1), we obtain the growth rate of tax revenues:

This rate is indispensable for estimating tax amounts each year in the process of compiling the annual budget. The necessary amount of taxes can be calculated by multiplying ET, if this value is definitely determined by the specific rate of A17Y, which is based on economic forecasts by the Economic Planning Agency.


45

The response of taxes to a change in nominal GNP logically takes place during two stages: (1) the response of the particular component of national income—i.e. taxable income—on which the tax is based to changes in GNP, and (2) the response of the tax yield to a change in the tax base. In order to observe these factors more closely, let us divide (3.1) into the two stages as follows. Thus, we obtain

as the new expression for ET, where B stands for the tax base,

The elasticity of the tax yield ET can be represented as the product of the elasticities of (1) the tax yield with regard to changes in the tax base Et and (2) the tax base with respect to changes in total income Ef,. Thus, Er varies directly with the elastici

t and Eb. The first 'tax-rate elasticity' relies heavily on the statutory-rate for mula of a progressive schedule of income tax rates. The second 'tax-base elasticity', however, is determined primarily by the way in which the private sector works during boom and slump periods. The value of ET tends to vary considerably from one tax to another. In the case of individual and corporate income taxes, it is generally considered to be high, depend ing on the rates and bases of the taxes. The terms for Et and Eb given in expression (3.3) could be the major reason for the high value of Ej- in the direct taxes. A progressive individual income tax should produce proportionately greater swings in tax revenues than in income. This is a typical characteristic of progressive rates, and consequently Et>\. On the other hand, the high ET of the corporate income tax seems to indicate that £;&> 1, since it shows the volatile changes in the level of corporate profits over the cycle. As regards the indirect taxes, it appears that the values of Ej, are lower with few exceptions. Neither Et nor Eb can rise to any significant degree because of their proportional or regressive rates and stable tax bases (see e.g. Musgrave 1959k, 506-7; Lewis 1962, 28; Maxwell 1955, chs. 12-14). Since the concept of elasticity used above deals with the automatic aspects of tax response, the calculation of ET requires estimates of that part of the change, in actual recorded data, which results from automatic rather than discretionary actions. Therefore, the practical measurement of ET raises some serious difficulties The most serious is that the data on tax collections in postwar Japan reflect not only an automatic response to changing income, but also tax reductions enacted in the tax code almost every year. This does not allow for a separation of the effects arising from changes in the tax code. In general, the tax revenue T may be assumed to depend on income Y, the minimum level of exemptions and deductions e, and statutory tax rates t. In other

46

Overview

words, T can be shown to function in the following way:

In measuring T for ET, T should be calculated on the assumption that e and f are fixed by the institutional setting. This entails using the tax revenue as reflected in income changes arising from the 'fixed tax system'. However, it is very hard to assume the tenability of a fixed tax system in the long run. As mentioned earlier, many tax changes have been enacted in postwar Japan, and these make such an assumption unreasonable. Under such circumstances, we cannot hope to measure any elasticity in the strict sense of the term. Actual tax revenues must be adjusted for any changes in tax laws. An attempt has been made by the Tax Bureau in the Ministry of Finance to develop a method for adjusting revenues for tax changes. It is a bold and crude attempt, but it is the only one that has yet been made and whose details are presently available in Japan. The Tax Bureau adjusts tax revenues for statutory changes that are made in the tax code each year. Actual taxes are converted to estimated tax accruals assuming the absence of tax code changes; i.e. tax rates, exemptions, and deductions remain unchanged. In calculating the value AT in (3.1), estimated tax accruals in the current year are compared with actual taxes in the preceding year. In other words, the calculations are based upon estimated tax receipts before the annual tax reductions.4 These estimates provide measures of the year-to-year elasticities of national taxes to changes in nominal GNP (or GDP) on the basis of the annual figures. Table 3.4 lists them for each category of major national taxes during the period 1953-98. Despite their uncertain accuracy,5 these elasticities are very important because the Japanese government has been using them as an indicator for forecasting tax amounts in the budget process every year. As can be seen from each column in Table 3.4, the elasticities show a high degree of variability from year to year, reflecting mainly volatile, short-run factors. The approximate magnitude of the long-run elasticities can be inferred by averaging the annual figures. To observe the long-run tendency of elasticities, it is informative to distinguish between the periods before and after the first oil shock.6 Over the period 1965-74, the elasticity of the whole tax system averaged 1.32, while the same figure fell to 0.85 for the period 1975-85. However, it jumped up to 1.94 for the second half of the 1980s, due to the outbreak of the 'bubble economy'. Once again it declined to 0.28 for 1991-8 after the bubble boom was terminated. 1 Lewis (1962, 29-31) made the adjustments simply by multiplying actual recorded tax accruals by the ratio of pre-change to actual tax rates. Pechman (1956,144) also made the adjustment for tax revenues in terms of an 'index of tax rates', which was based on estimates prepared by the Treasury Department. 5 There has been some criticism of the accuracy of the method used by the Tax Bureau series. Ishi (1968) attempted to recalculate relevant elasticities, selecting specific periods of'no tax change' on a quarterly and annual data basis. 6 Pechman and Kaizuka (1976, 349) tried to average year-to-year elasticities for a five-year period by using a specific formula in order to remove the erratic fluctuations in the annual data.


47

Table 3.4 Year-to-year income elasticity of the national tax system, 1965-1998 All taxes

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 Ave. 1965-74 Ave. 1975-85 Ave. 1986-90 Ave. 1991-98 a

0.67 1.19 1.51 1.27 1.41 1.58 1.35 1.38 1.93 0.92 -0.34 0.99 1.13 1.01 1.80 1.25 0.56 0.82 1.37 1.10 0.95 1.91 3.33 1.96 1.48 1.03 0.04 -5.21 -1.30 5.75 1.09 0.14 0.67 1.08 1.32 0.85 1.94 0.28

Individual income tax

1.75 1.48 1.76 1.85 1.94 2.25 3.29 1.96 2.52 1.55 1.64 0.99 1.41 1.23 2.14 1.84 1.97 1.39 1.67 1.25 1.44 1.78 2.69 1.57 2.80 2.59 0.63 -7.05 1.50 5.50 -1.82 -1.11 1.67 1.21 2.04 1.54 2.29 0.07

Corporate income tax

-0.19 1.17 1.82 1.37 1.41 1.59 -0.15 0.91 2.37 1.10 -2.96 1.22 1.20 0.83 1.71 1.70 -1.29 0.08 1.37 1.46 0.51 1.70 4.98 2.88 0.59 0.15 -1.48 -9.00 -12.40 8.25 5.91 1.64 2.22 2.17 1.14 0.53 2.06 -0.34

Indirect taxes

0.22 0.89 1.01 0.67 0.87 0.83 0.71 0.96 0.87 -0.28 1.00 0.69 0.71 0.88 1.55 0.01 0.51 0.51 0.60 0.34 0.46 1.87 2.02 1.04 0.79 0.06 0.11 1.26 3.30 8.25 1.41 1.18 -1.00 0.33 0.68 0.66 1.16 1.85

Alcohol tax

-0.09 0.73 0.57 0.23 0.49 0.63 0.44 0.73 0.71 -0.11 1.21 -0.11 0.82 0.40 1.09 -0.29 -0.20 0.96 0.53 -1.81 0.42 0.50 1.15 0.85 0.55 0.58 0.36 0.26 -0.40 5.75 -1.64 0.00 0.78 0.08 0.43 0.27 0.73 0.65

Indicates the average of the 1957-74 period.

Note: Income elasticity is the ratio of the percentage of tax revenues, based on estimated tax liabilities before the annual tax reductions, to that of GNP (1953-85) and of GDP (1986-98). Source: Unpublished data from the MOF Tax Bureau.

48

Overview

The responsiveness of Japanese tax revenues to income growth is due primarily to the high elasticity of the individual income tax, showing 2.04, 1.54, and 2.29, respectively, for the three periods defined above. On the other hand, the elasticity of the corporate income tax reveals more volatile figures. In fact it conspicuously fell to 0.53 in the second period, while it greatly increased to over 2.06 in the third period. Corporate income tax elasticity responds closely to sharp drops in business conditions and even shows negative figures during periods of slump, such as in 1965, 1971, 1975, and 1981. Indirect taxes and the tax on alcoholic beverages are less responsive to income fluctuation, and their elasticities are less than unity, as was pointed out in the theoretical discussion above. While income growth slowed down during 1991-8 due to the collapse of the bubble boom, the elasticities of individual and corporate income taxes have indicated substantial drops to 0.07 and —0.34 from the higher values of 2.29 and 2.06 in the second half of the 1980s. These facts imply how deeply income tax revenues can be affected by economic growth. In contrast, both indirect taxes and alcohol tax show relatively stable values of elasticity. Biased Estimates of Tax Revenues We must now turn our attention to the important concept of the 'natural increase' in tax revenues. Taxes are usually estimated at the time when the budget is being drawn up at the end of each calendar year. During each fiscal year in the 1950s and 1960s large 'natural' tax increases were realized, providing a substantial amount of new financial resources for the current and following year's budget. As noted above, part of these tax increases were appropriated to the financing of new government programmes, while the remainder were devoted to the financing of tax reductions. Therefore, this phenomenon of 'natural' tax increases bore a close relation to the size of the annual tax reductions. The natural increase in tax revenues seems to have been created intentionally, since the government consistently underestimated the tax amounts necessary to finance government expenditures at the first stage of budget planning. Natural tax increases, therefore, occurred during the fiscal year as a result of biased estimates of tax revenues. Some tax revenues, such as individual income taxes, naturally register increases as the tax base expands with the growth of the economy, even if there are no changes in the tax rate or exemptions. The higher the rate of economic growth, the larger is the amount of natural increase in tax revenues that can be expected. A question is raised about the estimation of the natural tax increase. This is based largely on the anticipated rate of economic growth, which is usually computed five or six months before the beginning of fiscal year. As an illustration, let us suppose that the government believes that GNP will expand 12 per cent in the following year. Based upon this anticipated rate, the MOF usually estimates projected natural tax increases. When doing so, some non-economic biased factors can easily be introduced into the calculation of the anticipated rate of economic growth. Frequently, the GNP growth


49

Fig. 3.1 Actual and anticipated rates of economic growth, 1955-2000 Note: In calculating the actual rate, the SNA data are used after 1965. The GNP series is published at three different stages, depending on the time of estimation: Y" stands for preliminary GNP, Y' for interim GNP, and Y for realized GNP. Anticipated rates of economic growth in n periods are defined as

n-rn i)/i"n i, while actual rates are as (Yn-Yn i)/Yn__l. Source: National Budget (MOF), Annual Report on National Accounts (EPA).

rate is intentionally underestimated in order to decrease the expected amount o natural tax increase used as a financial resource at this stage of budgetary preparations. Thus, since at the end of each fiscal year the realized rate of growth is always much higher than the anticipated rate, an enormous natural increase in tax revenue materializes after the implementation of the new budget. Figure 3.1 illustrates both actual and anticipated rates of economic growth from 1955 to 2000. Actual rates of economic growth markedly exceeded anticipated rates before 1974, but this relationship was reversed for about one decade after 1975. Hence, the gap between the anticipated and actual rates of economic growth substantially contributed to the production of large national tax increases in the 1950s and 1960s. In contrast, the government has occasionally suffered from a slowdown in revenue growth for a decade since the late 1970s. The large-scale revenue shortages in 1981 and 1982 mainly reflected the slowdown of income growth (see Ish 1982). Once again, anticipated rates have surpassed actual ones, generating a number of revenue losses in the 1990s.

TOWARDS TAX E Q U I T Y AND NEUTRALITY

The Emergence of Fiscal Deficits

To a considerable degree since the late 1970s, the basic strategy of tax policy has shifted from tax incentives to tax equity and neutrality. The main reason for this

50

Overview

Fig. 3.2 The gap between government expenditures and tax revenues, 1960-1998 Notes: Tax revenues include non-tax revenues, and all figures are related to the scope of general account in the national govenment. Source: MOF Budget Bureau, Fiscal Statistics (Zaisei Tokei).

is that fiscal deficits have begun to grow, reflecting a conspicuous slowdown in economic growth caused by the two oil crises (see Ishi 1986a; Noguchi 1986, 1987a; Lincoln 1988). To curtail the growth of fiscal deficits, the Japanese government terminated the annual reductions in individual income tax, and gradually began to increase corporate taxes and indirect taxes (see Table 3.3). As a result, the tax incentive policy necessarily came to a close. Traditionally, the Japanese government has followed a balanced budget policy. The roots of this policy lie in a conservative view of'sound' finance which developed in reaction to the extravagant government spending and inflationary pressures during the wartime period and immediately afterwards. A balanced budget was maintained in the true sense of the term until 1965, when national bonds were first issued in the postwar period. Although national bonds have been floated periodically since then, fiscal deficits posed no serious problems for a while.7 But they did become serious after the mid-1970s. Figure 3.2 depicts government expenditures and tax revenues during the period 1964-98. The gap between the two lines, which corresponds roughly to fiscal deficits, began to expand rapidly at the outbreak of the 7 Needless to say, the concept of fiscal deficits differs in accordance with the scope of the government (the 'general account' of either the national government or the 'general government', which includes all levels). In this regard, general account deficits are generally used when focusing on policy issues in lapan. For more detailed discussion, see Ishi (1986fo).


51

first oil shock in 1973. The situation has worsened since then. Since the collapse of the bubble boom in 1991, the gap has continued to expand to a great extent, reflecting revenue shortages in recession. What are the main causes for the sharp rise of fiscal deficits since 1973? Many factors can be put forth, but two are the most important. First, there was a conspicuous slowdown of Japanese economic growth caused by the two oil crises. Indeed, the real growth rate declined sharply, from 10 to 5 per cent in 1973, and to 3 per cent in 1979 when the second oil shock struck the Japanese economy. These experiences resulted in a sharp reduction of revenue growth, which in turn contributed greatly to the increase in fiscal deficits during the early 1980s. The second factor concerns the expanded role of government in income maintenance, medical care, public pensions, etc. In the early 1970s, important institutional reforms were undertaken in the social security system with a slogan, 'Construct the welfare state in Japan'. The target was to catch up with the Western level of social welfare programmes. Previously, it was often pointed out that Japan had lagged behind Western countries in the development of social welfare policies. Ironically, it was in 1973 that new social welfare programmes were launched and expanded to the same size as those in Western countries. At the time, it had been expected that the high rate of economic growth would continue and would thus generate the additional resources needed to finance higher public expenditures. Unfortunately, in the decade that followed, the rate of growth fell considerably in Japan as in most industrial countries. Growing Emphasis on Equity and Neutrality Because of the huge debt accumulation, the Japanese government decided that a substantial tax increase could not be avoided in the future in order to curtail fiscal deficits. At first, the government intended to resolve the fiscal deficits problem by instituting tax increases, rather than by making expenditure cuts. The introduction of a value added tax (VAT) was attempted in 1979, but resulted in a complete failure. As a consequence, it became politically difficult to introduce tax increases sufficient to reduce the level of national debt, and therefore the government changed its policy stance from advocating tax increases to promoting expenditure cuts.8 At the same time, the Japanese people demanded that the government should take the initiative in revising the inequitable burden of income taxation as a prerequisite to future tax increases. Because of these requirements, postwar tax policy was forced to shift towards putting greater emphasis on tax equity and neutrality. These trends gave rise to two important issues. One concerns the manner of reducing the number of special tax measures for tax incentives; the other concerns the perception of increasing unfairness of income taxation. As for the former, there 8 As regards the introduction of VAT and subsequent fiscal performances, see, for an expanded discussion, Ch. 15.

52

Overview

has been a growing realization that special incentives have eroded the fairness and neutrality of the tax system, and have made it increasingly complex. As noted earlier, Japan introduced a tax system after the Second World War which had relatively few special exemptions, credits, or similar provisions, under the influence of the Shoup Mission. Gradually, however, an interventionist approach to taxation had resulted in the enactment of various special tax measures designed to promote specific policy goals. The appearance of a renewed emphasis on tax equity and neutrality, with restricted use of special provisions, indicates that there has been a reversal of the interventionist trend in recent years. This trend is illustrated in Table 3.1 by a sharp decline in revenue losses from special tax measures in the 1980s. Concerning the latter issue, the Japanese government has been under increased pressure to address perceived inequities in the tax system. Complaints from taxpayers have centred round three issues and have concerned mainly the individual income taxes paid by salaried workers. First, there have been widespread complaints about the problem of 'bracket creep', which refers to the phenomenon by which salaried workers are pushed into higher brackets when they receive wage hikes for inflationary adjustment. Second, the problem of bracket creep is exacerbated by the fact that a large number of the self-employed and farmers have been able to escape their tax burden. In general, non-wage-earners, such as farmers and the selfemployed, can manipulate their declared taxable income to reduce their tax burden, whereas salaried workers cannot because their income tax is withheld at the source; as a result, they are considered to be subject to the utmost taxation by the tax office. Third, tax inequities among various income groups have been aggravated by the income-splitting practices available only to business owners and the self-employed. These groups are allowed to split their income by paying salaries to family members and even themselves, thus lightening their tax burden. Furthermore, they are allowed to deduct their business expenses to a greater extent than salaried workers, who are permitted only a standard deduction. As can be seen, salaried workers have major complaints about the present income tax system in terms of fairness.

4

Tax Administration and Tax Equity The aims of this chapter are twofold. One is to analyse how efficiently the Japanese tax system is administered from an international perspective. A major concern is with the well-established withholding tax system which may be the main factor enhancing the efficiency of tax administration. The other is to shed light on the reverse side of the withholding system and to explore the views of different taxpayers on the unfairness of the burden from a standpoint of equity. 'Tax gap' phenomena in Japan are also estimated using the available data to provide some empirical evidence. The plan of Chapter 4 is as follows. First, empirical findings which support the efficient aspects of Japanese tax administration are presented in comparison with those of the other major countries in the second section. Secondly, the third section analyses the well-established framework of the withholding system in Japan in institutional settings. Thirdly, empirical evidence of the 'tax gap' phenomenon between different sources of income with a discussion on policy is presented in the fourth section. Lastly, the discussion is concluded by proposing a desirable tax structure.

E F F I C I E N T TAX C O L L E C T I O N : A N I N T E R N A T I O N A L COMPARISON

Two Types of Tax Collection

Generally speaking, tax collection established two principles which have remained fundamental to tax administration in the leading industrialized countries: (a) withholding at source, and (b) self-assessment on a tax return basis. Income is divided into two categories: wages, salaries, interest, and dividends withheld at source; and other incomes required to file a return. It is widely acknowledged that the withholding tax system improves efficiency in collecting tax revenues by saving costs. Payers of income subject to taxes withheld at source are responsible for collecting the tax and paying it over to the government on behalf of each taxpayer. By contrast, under the self-assessment system each taxpayer is required personally to assess taxable income, file a return, and pay the tax due to the tax office. From an administrative point of view, tax collection is much more expensive under the self-assessment system because most of the collection work must be done by the revenue staff. This chapter is based on Ishi (1992fc).

54

Overview

Under these two different methods, taxpayers with income withheld have no freedom of manipulating taxable income, while self-assessed income-earners have a substantial margin to manipulate their income for tax purposes. Given the current state of tax collection, the former must feel that the basic rule of horizontal equity, that equals should be treated equally, is not being observed. For example, income tax of employees is administered under the withholding system, and unless the employer co-operates in evasion it is difficult for an employee to pay less than the tax due. However, in any country, avoidance and evasion on self-employment income including agricultural income are much more common; with the exception of a few abnormal cases, only a minority of taxpayers making tax returns are subject to tax inspection. There seems to be a general belief that income-earners who file their own taxes pay less tax than those whose taxes are withheld at source. Thus, different collection systems tend to induce a potential conflict between efficiency and the fairness of tax administration. This is obviously true in Japan. The main cause behind this fact is that the Japanese tax system has developed the withholding system to a greater extent over a broad area of primary taxes. Administrative Costs It is very difficult to investigate in quantitative terms how efficient tax administration is. To some extent, this analysis may be pursued by the measurement of collection costs, although this measure is far from satisfactory. Needless to say, costs of various kinds arise from the collection of individual taxes and the existence of the tax system itself. In what follows, particular attention will be paid to public sector costs (i.e. administrative costs), chiefly because necessary data are not available to cover compliance costs. To explore efficient aspects of tax administration, account must be taken that compliance costs in the private sector play an equally important role in collecting taxes. Compliance costs are less transparent than administrative costs, but they are significant in the process of administering the tax system as a whole. In particular, since there is a substantial degree of transferability from administrative to compliance costs, the government often tends to cut the former at the expense of the latter. Therefore, consideration of compliance costs should not be neglected for the purpose of our argument, but unfortunately there are no reliable data at hand.1 Our empirical analysis will therefore be based on only one aspect of collection costs. Let me begin with a preliminary discussion on the relative size of administrative costs compared to tax revenue. In general, administrative costs are officially calculated by the revenue department of a country, such as the Internal Revenue Service (IRS) in

1 Japan's tax authority is still cautious in exploring the actual situation of compliance costs. Thus we cannot obtain any relevant data in Japan at a similar level to the attempt by Sandford et al. (1989).

Tax Administration and Tax Equity

Fig. 4.1

55

Administrative cost as a percentage of tax revenue, selected countries, 1960-1995

Note: Each figure is calculated as a percentage of costs per $100, £100, or ¥100. Sources: USA: Internal Revenue Service, Annual Report, 1989 and 1990. Canada: Revenue Canada Taxation, Inside Taxation, 1975 and 1989; Supply and Service Canada, Report of the Department of National Revenue Customs Excise and Taxation, 1977-9 and 1981. UK: Board of Inland Revenue, Report for the Year, 1970,1972-5,1977-85, and 1987-90; Customs and Excise, Report of the Commissioners of Her Majesty's Customs and Excise, each year. Japan: National Tax Administration, Annual Report of Statistics, 1970, 1985, and 1990.

the USA. This includes wages and salaries of staff, accommodation costs, travel, postage and telephone, computing, and equipment costs. Therefore, based on official data, an approximate international comparison becomes feasible in terms of administrative costs as a percentage of tax revenue. Figure 4.1 shows the movements of such a ratio at the central government level in the USA, the UK, Canada, and Japan mainly during the period 1960-97. Unfortunately, it is not possible to obtain a detailed breakdown of data by individual taxes except for the UK where direct and indirect taxes are separated. In view of the possible different coverage of costs in each country, a strict comparison is not possible, but two interesting points are worth noting in the case of Japan. First, the ratio of costs to revenue until the early 1970s was higher in Japan than in the other countries. In particular, it was at a much higher level than that of the USA where the ratio has been kept very low. On this point, it seems that Japan's tax system was administered less efficiently in terms of the costs-revenue ratio. Second, however, the ratio in Japan declined sharply in the mid-1970s, and was lower than that of Canada. Thus, Japan now seems to lie between the USA and Canada. This comparison is a common procedure as the first step to present a measure of the efficiency in administering the tax system or of the relationship between input and output (i.e. 'productivity') in tax offices as a whole. It is, however, necessary to interpret the results derived from the data used in Figure 4.1 with care. Particular

56

Overview

attention should be paid to (a) nominal income growth and (b) changes in the tax system.2 The growth of nominal GNP generates additional tax revenue which tends to diminish the cost-revenue ratio. This was true until the early 1970s during which rapid economic growth was still prevalent in the Japanese economy. By contrast, the peak of the ratio around the mid-1970s implies that recession caused by the oil shock tended to decrease tax revenues and in turn led to an increase in the costrevenue ratio. If we emphasize the effects of income growth on changes in the ratio, the investigation of relative ratio would tell us nothing about the efficiency of tax administration. Tax changes are another and more important factor in exploring administrative efficiency in any time-series data. Tax reductions had been repeated almost every year before the outbreak of the oil shocks, producing a substantially decreased revenue, but towards the 1980s a deliberate tax-cutting policy was terminated to secure finance to make up for debt accumulation (see Ishi 1986«). Thus, changing the tax system led to automatic increases of revenue in the economy which resulted in a sharp decline in the cost-revenue ratio in the 1980s, as seen in Figure 4.1. However, it began to rise after 1990, reflecting the decrease in tax revenues. Another Measurement The next step is to make a more significant comparison by using per-staff basis data. Tax statistics in four countries provide us with three series of data: (1) the number of personnel in tax administration (N), (2) administrative cost (C), and (3) tax revenues (T), all of which are similar in the scope of coverage. Particular attention is paid to the gap between T/N and C/N, which implies a proxy of output and input per staff, respectively. If T/N is larger than C/N, the tax system is administered more efficiently and vice versa. Figure 4.2 shows the trends of these related data in four countries in terms of the base year= 100. Although the period covered and the scale of the index on a vertical axis substantially vary from one country to another, it seems that certain interesting facts can be derived from the comparison. Most importantly, great stress should be placed on the unique phenomenon of Japan's case in which T/N is consistently greater than C/N. On the other hand, the cases of the USA, Canada, and the UK turn out to be quite the opposite; the movement of C/N is in general over and above that of T/N, although the discrepancy between the two lines sometimes gets wider or narrower. In Japan, T/N has increased

2 See, for an expanded discussion, Sandford et al. 1989, pp. 19-20. In addition to these two points, the scope of revenues is also important. For instance, one of the main reasons why the US cost—revenue ratio has been kept much lower must be due to simultaneous collection of social security taxes with other taxes. By contrast, in Japan, the social security contribution is not included in tax revenues because it is collected by the Ministry of Health and Welfare.


57

more rapidly than C/N since the late 1970s, expanding the gap to a great extent. As stated above, the gap may be considered as evidence of the 'productivity' of tax staff. This being the case, the increased 'productivity' may indicate that the efficiency of tax administration has been improved in Japan as compared with that of other countries. One reason behind this is that the total number of national tax staff in Japan is much smaller. In fact, in 1996 the total was 55 108 in Japan, while the corresponding figures were 111 543 for the USA, and 93 138 for the UK in 1990. In addition, tax staff in Japan have increased by only a small margin: from about 52 000 in the 1960s and 1970s to about 55000 in recent years.3 In contrast to such a relatively stable number, tax revenues have enormously expanded, say, by more than 100 times during the same period. Needless to say, computer technology has enabled tax authorities to collect increased tax revenues with the very limited number of staff.

MAIN

F E A T U R E S OF A W E L L - E S T A B L I S H E D WITHHOLDING TAX SYSTEM

Current State of Tax Collection One of the distinctive factors explaining the efficient tax administration in Japan is obviously the withholding tax system which is firmly built into the basic structure of individual income tax. Historically, individual income tax was first introduced in 1887, and then the withholding system was adopted for interest income in 1899. Income withheld at source was widely expanded to cover employment income and dividends in 1940 when a sweeping tax reform was attempted by the government. Since the Shoup Mission recommended the overall reform package in 1949, the Japanese tax system has placed increasing importance on withheld income tax. In principle, the individual income tax is paid on a self-assessment basis, in which taxpayers themselves compute their income tax liability on the basis of their annual income and file a final return to the district tax office. As noted below, however, a major proportion of taxable income is now subject to being withheld at source. Taxes on such incomes as wages and salaries, interest and dividends are calculated by payers, deducted from the relevant source of income, and transferred to the tax offices on behalf of income-earners. At present, taxable income in the individual income tax is divided into ten categories, which are in turn classified into four types, depending upon tax collection methods. The relationship between taxable income and different types of collection are summarized in Table 4.1. Employment and retirement incomes (Type 1) are collected from payers on the basis of the withholding system on behalf of taxpayers—in its strict form. Such income is first computed comprehensively and progressive tax rates are strictly 3 The recent increase in staff reflects the fact that both the consumption tax (Japan's VAT) and the land value tax were introduced in 1989 and 1992.

58

Overview


59

Fig. 4.2 Comparison between per-staff tax revenue (TIN) and administrative cost (C/N) Notes: Temporary employees are included, and tax revenue excludes reimbursement. b Staff do not include temporary employees. Tax revenue is calculated by excluding revenues of the Canada pension plan and unemployment plan from total collections. c Casual staff are not included after 1968. '' Casual staff are included. 0 Okinawa regional tax offices are added after 1972. a

Source: As Fig. 4.1.

applied. Types 2 and 3 are kinds of interim measures, so far as the basic nature of comprehensive income taxation is concerned. In Japan, however, separate taxation at source has gradually become the practice in the case of specific incomes, mainly because of administrative considerations.

60 Table 4.1

Overview Tax collection methods by taxable income in 1996 (¥bn.)

Income

Withholding (Type 1)

Employment Retirement Interest Dividends Capital gains Stock etc. Others Real estate Timber Business Agricultural and self-employed Others Occasional Miscellaneous

262 660 11379

Subtotals

274039 (85.9)

Total

Separate taxation at source without final returns (Type 2)

with final returns (Type 3)

Filing returns (Type 4)

1 1 808 255

4003

1 010 6488

437 30 6345

28 7720 3407

297 3 136 13073 (4.11)

17034

14857

(5.3)

(4.7)

319003 (100.0)

Note: Figures in parentheses are percentage distribution. Interest and dividends contain some amounts that corporations have earned as well as individuals. Source: Calculated from NTA (1997).

Interest, dividends (some portion), and capital gains on the sale of stocks (Type 2) are separated from other incomes and tax is withheld at source by applying a flat rate of 20 or 35 per cent.4 Moreover, these incomes are excluded from the tax base in filing a tax return. Another type of taxation is applied to certain other dividends, capital gains on the sale of land and building, some business income, and miscellaneous income. Once tax on these incomes is withheld separately at a lower rate of tax,5 the incomes are required to be filed later as taxable income. The necessity of 4 All interest income and capital gains on the sale of stocks are withheld at source at the tax rate of 20 per cent (including 5 per cent of local tax) (see Ch. 8) while dividends attract 35 per cent. 3 In Type 3, dividends may alternatively be subject to the rate of 20 per cent at the taxpayer's option, but he/she must be required to file a final return. Capital gains on the sale of land and buildings are taxed separately on a self-assessed basis, depending upon the holding period of the relevant assets (see Ch. 17). Some portions of business income and miscellaneous income, which are mainly composed of fees, royalties, and remuneration paid to professionals, are withheld, usually at the rate of 10 per cent, as an advance taxation.


Fig. 4.3

61

Relative share of withheld income tax, 1960-1998

Note: Calculated from MOF, Primary Statistics of Taxation (Zeisei Sanko Shiryoshu), 1970, 1980, 1991, and 1998.

final returns distinguishes one type of separate taxation (Type 2) from another (Type 3). The amount of tax withheld under Type 3 is regarded as an advance payment of tax due, with final adjustment for the tax payable being made by incomeearners themselves as taxpayers. The remaining income (Type 4), from other categories of capital gains (e.g. the sale of valuable assets, paintings, or jewels), real estate, timber, and occasional, agricultural, or self-employed incomes, are levied under a self-assessment method on a calendar year basis. A tax return is required for annual income, to be filed not later than 15 March of the following year and be paid to the tax office at the same time. The relative share of taxable income collected under the present withholding system in 1996 is calculated in Table 4.1. The pure method of withholding occupies about three-quarters of total taxable income, while the share of filing returns accounts for only 4.7 per cent. Since separate taxation can be grouped into a withholding category, the importance in the tax system of non-filing returns increases further. It is difficult to estimate corresponding figures for international comparison, but the coverage of withheld tax collection in Japan seems to be much broader than in any other country.6 As argued earlier, individual income tax is paid in two ways: withholding or filing returns. In order to make clear the trend of relative weight in withholding taxation, Figure 4.3 shows two ratios of withheld income tax relative to national taxes and the 6

Rough information can be derived from OECD 1990 (Table 4, pp. 30-1). As far as this table is concerned, it seems that the source of income to which withholding tax is applied in Japan was the largest among the OECD countries.

62

Overview

individual income tax for 1960-98. Particular attention is paid to the upward movement of relative share in national taxes from 16.3 per cent in 1960 to 27.2 per cent in 1998. Meanwhile, the highest level of 33.1 per cent was reached in 1993. A sharp fall can be observed for the years 1986-90, but that was due to the abnormal state of the 'bubble economy'7 in which stock and land price hikes generated a great amount of tax revenues in the form of non-withholding taxes, such as the corporate tax, the security transaction tax, registration, and licence tax. No doubt, this induced a drop in the relative share of withheld income tax, A similar pattern is, more or less, shown in the upper line of Figure 4.3, although the increasing trend is less sharp. These facts imply that the withholding tax system has been entrenched in the framework of income taxation. How is Employment Income Taxed? There are two points to be emphasized about the significant role of the withholding tax system unique to Japan with particular reference to employment income (Type 1). First of all, the withholding system plays an important role in reducing the number of 'direct taxpayers' to pay taxes due. The number of taxpayers who are paying withheld income tax on employment income in 1996 is estimated at 52009683. By contrast, the number of withholding agents (i.e. 'direct taxpayers'), who have the obligation to withhold the tax at the time of income payment, accounts for 3 893 483: merely one-thirteenth of employment income-earners. It is easy to understand the efficient mechanism of tax collection through withholding agents if we compare the small number of withholding agents with the large number of self-assessed income taxpayers for filing returns: i.e. 8 239 8588 in 1996 (see NTA 1997). Second, great stress should be placed on how to withhold employment income at source. For this purpose, both the withholding tax table and the year-end adjustment are prepared to calculate the tax payable and enable it to be paid to the tax offices almost simultaneously. The amount of withheld tax on employment income is estimated every month, based on elaborate withholding tax tables (see MOP 1997, p. 66). These tables are prepared by the National Tax Administration to take account of many factors, such as progressive tax rates and a variety of exemptions and deductions. Since the Japanese companies generally pay salaries monthly, employers easily withhold the tax due on income, based on a 'withholding tax table for monthly salary payments'. Similarly, tax on a bonus, which is equivalent to a few months' salary in summer and winter in accordance with the Japanese wage custom, is also calculated by using a 'withholding tax table for bonuses'. 7 One of the most remarkable phenomena in the second half of the 1980s was a 'bubble economy', mainly caused by easy monetary policy. Excess money risked in stock and land markets, produced the abnormal hike of asset prices. During the period 1985-90, nominal rate of GNP growth accounted for 6.0 per cent. 8 This figure contains some taxpayers, subject to withholding tax, whose income exceeds ¥15m. per year or is earned from more than two sources. These taxpayers are obliged to file a final return in addition to an advance payment of withholding tax.


63

However, the use of such withholding tables provides merely a provisional calculation of the tax liability. Thus, at the end of the year, employers are requested to calculate annual income and the tax due as a whole, and to adjust for the difference between the annual tax liability and the tax amount already withheld. This is a 'year-end adjustment'. When such an adjustment is made every December, certain special deductions not considered in the monthly withholding table are added to recalculate total tax liability.9 As a consequence, the year-end adjustment plays an equal role in filing a final return. Since most employees usually have no other income of Type 4 (see Table 4.1) or no income higher than ¥15 million, they do not need to get direct access to the tax offices.10 Given the two factors mentioned above, the withholding tax system on employment income is characteristic of the following three points (see Ozaki 1991). 1. A great number of taxpayers who are employment income-earners have no relation with the tax offices. Indeed, only 3.5 million people out of 46 585 000 taxpayers paid income tax on employment income by tax returns in 1991. 2. Withholding agents perform the same job as the district tax offices. If they fail to collect the tax at source and to pay it to the tax offices, unpaid taxes will be collected directly from the agents, not income earners. Even in the case of tax delinquency, additional taxes and interest on them are paid through the agents. 3. Taxpayers are requested to present their personal data for tax purposes to their employers, not the tax offices. Therefore, taxpayers need to follow minor procedures which require a much lighter work load than that for self-assessed taxpayers. Likewise, other incomes under separate taxation, such as interest and dividends (Type 2), are equally withheld at source. Since such capital income is paid very widely to an unspecified number of people by financial institutions, withholding taxation is absolutely necessary to collect the tax on such items of income adequately. In particular, because there are no Tax Identification Numbers (TIN) at present, it is impossible to tax interest, dividends, and capital gains on the sale of stocks adequately in the strict sense of the term. Additional evidence can be derived from Figure 4.4 to show the efficient aspect of withholding taxation. Data on tax delinquency is available by categories of national taxes, and the ratios of tax delinquency to relevant tax revenue are depicted for 1960-96 for four selective taxes. Among them, the withheld income tax has kept the lowest level of the tax-delinquency-revenue ratio as compared with other taxes and the average of total national taxes. Obviously, the withholding tax system has been very effective in securing necessary revenues, leaving behind no possible delinquency of taxation. 9 For instance, deductions for accidents, medical expenses, life insurance premiums, etc., are taken into consideration for a year-end adjustment. 10 The method of withholding taxes on employment income in Japan is substantially similar to the UK cumulative PAYE system applied to Schedule E.

64

Fig. 4.4

Overview

Tax deliquency as a percentage of revenue, 1960-1996

Source: Calculated from National Tax Administration, Annual Report of Statistics (Kokuzeicho Tokei Nenposho), 1991.

What is of great interest is the sharp rises of each ratio after 1991 when the Japanese economy plunged into prolonged recession. In particular, self-assessed income tax began to increase the tax-delinquency ratio sharply, reflecting the bad time of business and the accumulation of bad loans. Even withheld income tax tended to lift the relevant ratio.

EMPIRICAL EVIDENCE OF THE TAX GAP A TEST OF THE 'KU-RO-YON' PHENOMENON 'Ku-ro-yon' Ratios Turning to the other side of administrative efficiency, we shall focus on the unfair tax burden among different income sources in relation to horizontal equity. Horizontal equity is often related to administrative practices. In principle, horizontal equity is frequently impaired when administrative arrangements are not satisfactory. This is the case in Japan. As noted earlier, it is widely believed, especially among salaried workers, that there are large divergences in the identification of taxable income among different classes of taxpayers. Since salaried workers are taxed at the source of income under the withholding system, their income is almost fully (90 per cent) identified by tax authorities. On the other hand, the self-employed (including practising doctors and solicitors/barristers) and farmers file their own income returns. They are not taxed fully at source and can easily dodge tax liability by underreporting their income. Reputedly,


65

only 60 per cent of the incomes of the self-employed and 40 per cent of farmers' incomes are caught by the tax office. These percentages (90-60-40) are used so often in describing the present unfair situation in the Japanese tax system that a special term, 'Ku-ro-yon', has been coined. 'Ku-ro-yon' is a portmanteau word of Japanese numbers—9 (ku), 6 (ro), and 4 (yon). This term is, by and large, used in the same way as the 'tax gap' among different tax sources is referred to in the USA. It is very difficult to test the 'Ku-ro-yon' ratio statistically. One possible method, which I have tentatively attempted, is to compare the scope of taxable income quoted in tax statistics (TS) with that which appears in national income statistics (NIS). The most difficult task is to make the necessary adjustments for obtaining a common base against which comparisons can be made. In Table 4.2, differences in concept definitions are enumerated in employment, self-employed and agricultural incomes. Y stands for income recorded in national income statistics (i.e. NIS-based income), while y denotes taxable income in tax statistics (i.e. TS-based income). The relation between the two income concepts is basically expressed as follows:

The term 'a is considered as an unreported or underreported portion of taxable income, presumably due to tax evasion and avoidance (for general discussion, see Goode 1981; Roth et al 1989; Webley et al. 1991). If we regard NIS-based income as a reference level11 for comparison with taxable income recorded in tax statistics, the tax gap ratio (8) can be defined as follows. Both sides of (4.1) are divided by Y,

8 indirectly indicates the magnitude of unreported or underreported income which may be considered to be another interesting measure. The tax gap ratio is calculated for each income source,12 and 81( 82, and 83 are linked with that of employment, self-employed, and agricultural incomes, respectively. 11 Of course, this assumption would not be plausible. Not all the value of economic activity is recorded in national income accounts and some is presumably untaxed because of the so-called underground economy. However, since there would be no other reliable alternative, we rely upon NIS-based income as a reference to which reported taxable income is matched. 12 The tax gap is defined by income sources, not income earners, because income earners tend to have other incomes apart from income from their major occupations. In Japan, for example, farmers frequently earn both agricultural and employment incomes by working at nearby factories. Thus, it is more accurate to define the tax gap in terms of income sources.

66

Overview Empirical Results

The matching of TS-based income to NIS based income can be justified by the fact that the two incomes are obtained statistically from two different data sources. For instance, employment income in NIS is estimated by the Economic Planning Agency, mainly based on the Monthly Report of Working and Wage and other related statistics of the Department of Labour, while the corresponding figure in TS practically is collected by the National Tax Administration for tax purposes. The same holds for both self-employed and agricultural incomes.13 Such a comparison may make sense to obtain evidence of the tax gap. Procedures for tax gap estimates are rather complicated, based upon data processing by a number of different statistics.14 Thus, it should be emphasized that the estimates are subject to large potential errors. In particular, this is true in the case o self-employment income, because income below minimum taxable level unrecorded in tax returns must be estimated with bold assumptions. The matching for employment income is the easiest procedure among the three cases, producing the most reliable result. Basic tax statistics13 provide us with almost all the necessary data pertinent to taxable income in TS, including both b and c in Table 4.2. Therefore, it is not necessary to attempt troublesome estimates of these two items, and only minor adjustments need to be made for trivial differences in concept definitions; say, bonus for company executives which is originally included in company profits. On the other hand, both self-employed income and agricultural income require some additions to taxable income in TS. As is seen from the notes in Table 4.2, the minimum taxable level under the income tax law is first set at a specific income level each year, for a standard household (couple and two children), and then income below such a threshold is estimated by using data on the income distribution by income classes from Employment Status Survey (Statistics Bureau, Management and Coordination Agency).16 Procedures for these estimates must essentially become crude, reflecting less reliable data sources. Furthermore, special deduction for wages paid to family employees should be included in tax returns. The whole process of estimation concerning the cases of self-employment income is done by attempting

13 NIS-based self-employed income is estimated on the basis of Statistics of Non-corporated Business Offices (Jigyosho Tokei Chosa) and Economic Survey of Non-corporated Business Offices (Kojin-kigyo Keizai Chosa) (Statistics Bureau, Management and Coordination Agency). Similarly, NIS-based agricultural income employs farmers Economic Survey (Nohka Keizai Chosa) and Agricultural Census (Nogyo Chosa) (Ministry of Agriculture, Fishery and Forestry). 14 For example, in order to obtain detailed classification of incomes in question, I must use worksheet-level data at the Economic Planning Agency which are fortunately available to me. 13 Basically, Annual Report of the National Tax Administration and Statistics on Private Wages and Salaries (NTA) is employed. 16 Employment Status Survey (Stingyo Kozo Kihon Chosa) has not been published every year. In the past, data were only available in 1974, 1979, 1982, and 1987 for the period to cover our estimates. Thus, any single-year data were extended to make multiple-year estimates before or after the limited years.

Table 4.2 Conceptual adjustment between TS-base and NIS-base incomes3 Taxable income subject to tax code

Unreported or underreported income

Income below minimum taxable level*3

y

a

b

Non-taxable income due to additional deductions0 c

income in TS

?

including in

including in

y\

y\

b2

C

Employment income

y\

Self-employed income

the same Vi

Agricultural income

the same

yi

?

?

b3

2

C3

Income excluded in TS, but included inNIS d

Income Difference included in TS, in coverage but excluded inNIS e /

special deduction for wages paid to family employees d2

capital gains in inventory

the same d3

the same

e

2

bonus for company executives /i livestock farming and fishery incomes h

«3

Notes: TS—Tax Statistics, NIS—National Income Statistics. Minimum taxable level is calculated in the case of a standard household (couple and two children, including basic exemption, exemption for dependants, exemption for spouse, deduction for social insurance premiums, and special deduction for blue/white return). c Deduction for medical expenses, deduction for life insurance premiums, deduction for fire and other casualty insurance premiums, and deduction for small-scale enterprise mutual aid premiums are added to five items of minimum taxable level. a

b

68

Overview

Fig. 4.5

Tax gap among three income sources, 1970-1990

separate procedures for blue and white tax returns (see below).17 Certain adjustment is also needed to change the coverage of income in the case of self-employed income (i.e. the term'/' in Table 4.2). Final empirical results for 1970—90 are depicted in Figure 4.5 where we can find two kinds of line for each category of income. One line is the tax gap ratio itself;1;8

2, and 83, and the other is a matching ratio of original income data before concep tual adjustment: y\IY\, y2/Y2, and y^/Y3. Since a substantial number of estimat procedures are included to obtain the final tax gap ratio, original matching might be of use to ascertain the accuracy of the tax gap in question. Major fact findings are shown in the following four points. First of all, Sj, and y\IY\ in the case of employment income substantially remained stable in the band of 90-100 per cent during the period 1970-90. It can be conjectured that there would be almost no difference between NIS-based and TS-based incomes, and that the 'ku' ratio is justifiable by empirical data. The incomes of wage and salary earners are fully captured by the tax offices under the withholding system. Second, self-employed income varied the level of 82 to some extent, but it can be pointed out that the 'ro' ratio did not come into existence until about 1985. As will be argued shortly, it is necessary to explain the upward movement of 82 in the late 1980s. The other line of y2/y2 moved more smoothly with a regular margin agains

2, and the adequacy of our estimates would be reinforced by the original matching ratio. Third, the growth of 83 and y^Y3 in the case of agricultural income is differen from the other two cases. 83 moved below the 40 per cent level for 1970-84, but it 1 Statistics on the Self-assessed Income Tax (NTA) is employed to obtain necessary information.


69

turned upward sharply after 1985. Also, y^Y3 followed a similar pattern to 83 durin the whole period. The 'yon' ratio may not exactly be justified, but it is noted that 83 and 73/^3 were kept at the lowest level with a minor exception as compared with corresponding figures. Fourth, judging from our estimates, it seems that the tax gap among different income sources in Japan may be empirically tested to some extent in conjunction with the 'Ku-ro-yon' phenomenon. Particularly, this would be the case for 1970-85. Other studies of the tax gap also support my empirical evidence.18 Recent Development of Tax Gap The tax gap, as stressed previously, has been produced by different tax collection between the withholding system and tax returns. Given the existence of the tax gap found above, the well-developed withholding system has obviously strengthened taxpayers' perception of the unfair tax burden of wage and salary earners, while it has made tax administration more efficient. As a consequence, the conflict between efficiency and equity of tax administration in the Japanese tax system has been induced to a great extent by developing the withholding collection. It seems, however, as seen in Figure 4.5 that the tax gap narrowed up to 1990. As far as this evidence is concerned, the conflict between the two objectives of tax administration begins to be mitigated to some extent. This being the case, it would be necessary to explore the factors necessary for increasing the 'ro' and 'yon' ratios of self-employed and agricultural incomes. Most importantly, great emphasis should be placed on the improvement of tax return methods. The self-assessment on a tax return basis is generally divided into two systems, i.e. (a) blue returns and (b) white returns. When the Shoup Mission proposed tax recommendations in 1949, the blue return system was introduced to improve book-keeping and to promote honest self-assessment by taxpayers subject to income tax returns. Since then, the National Tax Administration has considered the 'blue return' as the fundamental requirement for efficient tax administration. The main aim of the blue return is to encourage small- and medium-sized business to keep a minimum set of accounting records, but in addition certain significant advantages are offered to individuals and corporations by the tax offices (see MOF 1991,63-4). The major advantage is that taxpayers filing a blue return are not subject to reassessment as long as errors cannot be found in their accounting books and records. Moreover, they are allowed to deduct reasonable amounts for wages paid to family members working in the same companies and to use special tax-free reserves 18 Following my own estimate in Ishi 1983fc, three studies have so far tried to testify to the 'Ku-ro-yon' ratio: i.e. Homma et al. 1984; Hayashi 1990; Okuno et al. 1991. These estimates lead, more or less, to the same empirical results, although statistics and procedures used are quite different in each. For example, Hayashi estimates 101.3-52.5-13.3 in 1979, 99.4-58.6-14.3 in 1982, and 101.4-61.7-20.7 in 1987 as the 'Ku-ro-yon' ratio. Likewise, Okuno etal. find 104.8-60.4-27.6 in 1985 (all figures are percentages).

70

Overview

(e.g. reserves for bad debts, losses due to price fluctuations, etc.). By contrast, taxpayers, who are not filing a blue return, are not given these advantages for tax purposes, but they are not obliged to keep books and records. This case is usually called the 'white return' system. Traditionally, most farmers do not file blue returns, and the tax offices estimate their income on the basis of their crops. This is considered to be a major factor leading to understatement of agricultural income for tax purposes. In fact, blue returns as a percentage of total filing returns were very low for farmers until the early 1980s; 2.5 per cent in 1970, 7.1 per cent in 1975 and 10.2 per cent in 1980. However, the blue-return ratio has sharply increased to higher values since then; 17.8 per cent in 1985 and 32.0 per cent in 1990. The sharp rise in using blue returns is evidently thought of as the most important factor to explain the narrower tax gap in agricultural income after 1985. The same reasoning may be applied by and large to the case of self-employed income, but it is not so clear-cut as the agricultural case. The corresponding percentage using blue returns merely increases from 48 per cent in the early 1970s to 51-3 per cent in the 1980s. More importance should be put on increasing exemptions and deductions applicable to self-employed taxpayers.

SOME POLICY IMPLICATIONS

In this chapter, the 'Ku-ro-yon' ratio does indeed seem to be approximated by these statistical procedures, although the results are far from satisfactory. A tax gap between the three income sources would probably arise from both evasion and avoidance. Obviously, it would be impossible to draw borderline distinctions statistically. It is necessary to observe with care changes in the tax gap, as seen in Figure 4.5 and to identify whether or not these phenomena would be temporary in the future. At the moment, however, let us assume that the 'Ku-ro-yon' ratio is still prevalent in tax administration. A high proportion of popular complaints about the present tax system emerges from inequities of this kind. These complaints have arisen among a majority of wage and salary earners. The national atmosphere vis-a-vis inequitable income taxation has attracted wide attention among the general public. Without doubt, this is one of the inherent features built into the Japanese income tax system. It is important to note that this atmosphere among the general public potentially supports the need to increase reliance on indirect taxes in the tax system. In the past, two sweeping tax reforms by both the Nakasone and Takeshita cabinets attempted to introduce the value added tax in the Japanese tax system and were finally successful (see Chapter 17). No doubt, this success was primarily due to the support of wage and salary earners who had major complaints about unfair tax burdens because of the 'Ku-ro-yon' phenomenon.


71

It is still uncertain whether recent tax reforms will solve the conflict between efficiency and equity of tax administration. It seems, however, that efficient administrative practices have essentially led to the adoption of VAT in Japan in order to mitigate inequities of tax burdens caused by the well-developed withholding tax system. The argument concerning the 'Ku-ro-yon' phenomenon was most popular in the 1980s in relation to the introduction of VAT. It seems, however, that the atmosphere against it virtually cooled down towards the 1990s for two reasons. For one thing, the long-term decline in the agricultural sector has structurally been remarkable in Japan, reflecting the steady decrease of farming families from 5.3m. in 1970 to 3.6m. in 1994. The other is that smaller-size farms have tended to be merged into large ones with modernized business management.


Part II

Individual Income Tax


5 Basic Structures

Similar to practically all other developed countries, Japan depends upon the individual income tax for a significant portion of its tax revenues. Throughout the postwar period, the income tax has reigned supreme in the Japanese tax system. In more recent years, tax experts have begun to pay attention to certain of its defects, particularly from the standpoint of those favouring an expenditure tax. In my view, however, the income tax will continue to be widely regarded as one of the best methods of taxation yet devised. In Japan, the tax on individual income is used simultaneously by the national, prefectural, and municipal governments. Our major concern here is with the national income tax; local income taxes will be discussed in Chapter 18. In this chapter we consider the basic structures of the individual income tax at the time of writing (April 1999) from three angles. First, it is necessary to define a few important terms used in computing the income tax base, such as 'income for tax purposes', 'exemptions and deductions', the 'tax unit', and the 'treatment of fringe benefits'. Second, the structure of progressive tax rates is considered with special reference to the application of separate tax rates on certain incomes. Lastly, attention is focused on some salient features of the present income tax, including a hybrid of an income tax and an expenditure tax.

THE

TAX BASE

Computation of Taxable Income In principle, the Japanese tax system still maintains a global system of individual income taxation, although in actual practice separate taxation methods have been introduced as exceptional cases. So long as this primary principle is followed, it is relatively easy to compute the final tax liability for each taxpayer. Each incomeearner first adds up his income from all taxable sources, then subtracts allowable exemptions and deductions to attain the amount of taxable income. After applying a single progressive rate schedule to such taxable income, the taxpayer may use certain tax credits to arrive at the final amount due. Taxable income is classified into ten categories: (1) interest income, (2) dividends, (3) real estate income, (4) business income, (5) employment income, (6) retirement income, (7) timber income, (8) capital gains, (9) occasional income and (10) miscellaneous income. Most of these categories permit specific exemptions and deductions,

76


Table 5.1 Minimum taxable level of the individual income tax (wage-earners), selected countries, 1998 Japan a (¥1 000)

USA ($)

National Local

Federal

Single

1 107

1 053

Couple

2095

1 857

Family of 3

2698

2380

Family of 4

3616

3031

9800 (1166) 12500 (1488) 17869 (2126) 20569 (2448)

UK (£)

Germany France (DM) (F)

4045 (789) 5418 (1056) 5418 (1056) 5418 (1056)

17714 (1187) 32996 (2211) 45416 (3043) 55784 (3738)

Stateb 7500 (893) 13000 (1547) 14000 (1666) 15000 (1785)

72771 (1455) 116442 (2329) 138268 (2765) 160112 (3202)

Note: Figures in parentheses are those expressed in terms of ¥1000, using 1998 exchange rates: $1 =¥119, £1 =¥195, DM1 =¥67, and Fl =¥20. a b

Includes the deduction for social insurance premiums. New York State income tax.

Source: MOF, Tax Bureau, Primary Statistics of Taxation (Zeisei Shuyo Sanko Shiryoshu), Dec. 1998.

except for interest and dividends. Fundamental personal exemptions for average households consist of the basic exemption, the exemption for spouse, and the exemption for dependants: this came to a flat ¥350 000 each for the taxpayer himself, his spouse, and other dependants in 1992.' Besides these personal exemptions, wage and salary workers are permitted to subtract two special deductions for earned income and social insurance premiums from their employment income. The MOF includes these five exemptions and deductions in calculating the minimum taxable level (the tax threshold) of each household. Table 5.1 shows the minimum taxable level of wage-earners among five selected countries. It is very difficult to obtain comparative figures since the income tax law admits different structures of exemptions and deductions in each country. Japan reveals relatively higher levels for the tax threshold, as do Germany and France. This implies that the actual tax paid by taxpayers at various income levels is rather modest, particularly in comparison with the UK and the USA. Allowable Deductions and Credits In addition to personal exemptions, which are widely applied to taxable income as a whole, numerous deductions from different income sources are allowed for

1 In addition, a special exemption for spouse (¥350000) was allowed from Sept. 1987 in the case of one-earner couples to balance the tax burden between wage-earners and others.

Basic Structures

77

specific reasons. Some of the major deductions are explained in the following paragraphs. Employment income, as already noted, is given a special deduction instead of being entitled to deductions of actual amounts of personal expenses.2 The main aim of this deduction is to put wage and salary workers, the only categories to whom this deduction is available, on a more equal level with the self-employed, who can treat their personal expenditure as business expenses. This special deduction for employment income is considered to be rather generous, although it can be used only on an across-the-board basis. In fact, the ratio of the amount used for this deduction to average earnings per worker accounted for 34.1 per cent in 1985, 31.6 per cent in 1990, and 31.7 per cent in 1996. A variety of special treatments is also involved in the determination of business income. Generally speaking, net income is subject to taxation after all necessary expenses are subtracted from gross receipts. Taxpayers filing a blue return were permitted to deduct the cost of preparing the tax returns up to a maximum of ¥100 000 until 1993. A special deduction for wages paid to family employees is also allowed in lieu of deducting them as necessary business expenses from the proprietor's business income. Even the proprietor's own remuneration is deductible from his business income, and this can also be applied to the special deduction for employment income, if the firm elects to be treated as a quasi-corporation for tax purposes (i.e. as a 'deemed corporation'). As a result of tax reform in 1992, this 'deemed corporation' system was abolished from 1993. Instead of this, however, the special deduction for filing a blue return was increased from ¥100000 to ¥350000 on condition that good record-keeping is observed, and it was raised to ¥450 000 in 1998. Retirement income, timber income, and capital gains are each granted special deductions. For retirement income, the deduction is ¥400000 per year for up to 20 years of employment or ¥700 000 per year for anything over 20 years of employment in excess of the amount accumulated during the first 20 years. For timber income, a special deduction of ¥500000 is deductible from gross receipts minus necessary expenses. Capital gains on the sale of securities and real estate are taxed separately from ordinary income tax after allowing for special deductions. Besides the special deductions permitted to certain taxable incomes, there are nearly 20 deductions involved in the present structure of individual income taxes.3 To compute final tax liability, several tax credits may also be taken, such as credit for dividends, credit for foreign taxes, credit for incremental research and experimental expenditure, credit for acquisition of a house. Some of these are linked to the promotion of specific policy goals. 2 'Employment income' includes not only wages and salaries, but bonuses, pensions, and other allowances of a similar nature. In 1999 the deduction was 40 per cent of the first ¥1,80m., 30 per cent of the next VI.80m., 20 per cent of the next ¥3.Om., 10 per cent of the next 3.7m., and 5 per cent for amounts in excess of ¥10m., but the minimum deductible level is set for ¥6.5m. 3 Typical categories of such deductions include deductions for casualty losses, medical expenses, donations, physical handicap, etc. For a comprehensive list, see MOF Tax Bureau (1998).


78

The Tax Unit In prewar Japan, the family unit was adopted as the basic tax unit, reflecting the belief that the head of the household (usually the husband) was responsible for filing a tax return of joint income on the grounds that the wife and children were dependent on him. The Shoup Mission recommended that the family be replaced by the individual as the primary tax unit. Since then, the Japanese tax system has maintained a system in which the individual is the unit of income taxation, although some exceptions have been allowed. According to this system, two-earner couples are taxed as separate individuals. For one-earner couples with several dependants, the head of the household is treated as a single taxpayer. The tax unit adopted by the Japanese is similar to that of Australia and New Zealand in which the individual is still predominant. It is in sharp contrast, however, with the USA, the UK, and France. (For a more detailed discussion, see Kay 1987.) The USA uses the family unit as the basis of taxation with distinctive tax rates for single persons and couples under an income-splitting approach. Similarly, the UK employs a unit basis of assessment in which the income of a working wife is aggregated with that of her husband after relief is taken for the wife's earned income. France has a unique quotient system that splits household income among the members of family. Given such diversity of existing practices regarding the tax unit, the Japanese tax unit does not seem to raise any serious problems. Accordingly, married couples in Japan with the same total income pay different taxes, depending on whether there are two wage-earners or one. Table 5.2 shows Table 5.2 Different income tax liabilities among one-earner and two-earner couples with a combined income of ¥10 million, 1998 Tax liability (¥000) Husband

Tax burden rate (%) Wife

Total

One earner Husband Wife

¥10 m. 0

720

0

720

7.2

Two earners Husband Wife

¥8m. ¥2m.

520

70

590

5.9

Husband Wife

¥6.5 m. ¥3.5 m.

294

165

459

4.6

Husband Wife

¥5m. ¥5m.

177

273

450

4.5

Note: Tax liabilities arc calculated on the 1998 tax code, assuming that each couple is a wage and salary worker with two dependants, excluding temporary tax cuts in 1998.

Basic Structures

79

Table 5.3 Income tax liability and income-splitting 1998 Income scale ¥m. 3 4 5 8 10 20 30 50

Present system

Income-splitting

Tax (¥000)

Tax burden rate (%)

Tax (¥000)

Tax burden rate (%)

0 28 101 368 720 3309 6822 15785

0.0 0.7 2.0 4.6 7.2 16.5 22.7 31.6

0 28 101 349 525 2366 4929 11392

0.0 0.7 2.0 4.4 5.3 11.8 16.4 22.8

Note: Tax liability under income-splitting is simply computed as ((taxable income X 1/2) X present tax rate} X 2.

the tax liabilities of various married couples with the same combined income of ¥10 million. Married couples with only one wage-earner pay higher income taxes than those with two wage-earners. To the extent that total income can be earned more equally by couples, they tend to pay less tax. Given the present structure of the individual income tax, the Japanese tax system benefits two-earner couples with evenly split income, although an exemption for a spouse is allowed for a one-earner couple. What happens if the income-splitting approach is introduced to one-earner couples in Japan? In Table 5.3, the tax liabilities of married couples with one wage-earner are computed as if they were two single taxpayers with their total income divided evenly between them. The effect of income-splitting would evidently lower the tax burden at higher-income classes. For instance, at the top income of ¥50 million, the tax burden would be reduced from the present rate of 31.6 per cent to 22.8 per cent. If children are recognized as independent earners, they may be taxed separately on their own income. However, few children have incomes large enough to be subject to taxation. Their financial reliance on their parents enables the family to take exemptions for them as dependants. Despite the fact that the tax unit is based upon the separate individual, the investment income of the spouse and other dependants used to be taxed as if it were received by the head of the household. The main purpose of aggregating income from investment was to prevent tax avoidance by allocating such income arbitrarily among family members.4 ^ The aggregation tax method of investment income was established as an exception to the individual tax unit in 1950, but it was repealed in 1951. In 1957 this method was re-established and was maintained until 1988. It has now been repealed.

80


The problem of the tax unit in Japan has some bearing on the special treatment of business incomes. As noted earlier, business proprietors are allowed to split their incomes by paying wages and salaries both to other members of their families and to themselves. Obviously, this income-splitting lightens their income tax burden, and as a result this practice is greatly criticized by wage-earners as being inequitable. Fringe Benefits In Japan, tax advantages for fringe benefits have not received any more attention than in other major countries. Apparently, the existing tax treatment of fringe benefits is more favourable than that accorded to ordinary income. Since this affects mostly employment income, it would be unfair to omit this aspect of taxation in relation to the definition of taxable income. Under the individual income tax schedule in Japan, cash income is taxable, but non-cash benefits given to the wage-earner and the business executive either are not taxable at all or are only partially taxable. Although no clear boundary can be demarcated between ordinary income and fringe benefits, there are several kinds of tax preferences for fringe benefits worth noting here. The most important of these benefits is the exclusion of the value of subsidized housing from taxable income. Most large companies, as well as the government, provide housing to many of their employees at much cheaper rents than market prices for comparable facilities. Companies also subsidize loans for the purchase or construction of private residences, so that employees can obtain mortgages at a very low rate of interest over ten or more years. The subsidized portion of such rents and interest payments is excluded from taxable income. Furthermore, the system of employee compensation includes free or cheap provision of recreation facilities and other welfare benefits. Such payment in kind is non-taxable, and greatly benefits Japanese wage-earners, particularly those working for large companies. For the majority of business executives, a compensation package is provided by their companies: a car, a large residence, an expense account for entertainment, etc. It is not common, however, for corporate executives in Japan to hold stock in their own companies, because of the lack of stock option plans. Of greatest importance is perhaps the expense account, which is by and large used for inviting business customers to expensive restaurants, clubs, and golf facilities. Thus, many people find it easy to see why their employers are generous in paying out large sums of money for entertainment purposes: all of these benefits go untaxed in their personal incomes.5 Needless to say, the tax advantage of fringe benefits aggravates the inequities between the treatment of taxpayers who receive only a cash income and those who 5 At present, total expenditures for entertainment expenses are not deductible from corporate income. However, small firms with ¥10-50m. of capital are permitted to deduct ¥3m. a year (but 20 per cent of the portion if less than ¥3m.), and those with less than ¥10m. are allowed ¥4m. (likewise, 20 per cent below ¥4m.). In the past decade or so, such expenses have been taxed more strictly.

Basic Structures

81

receive fringe benefits as well. Traditionally, major tax reports written around the world have reached a clear agreement on the appropriate taxation of fringe benefits (see Kay 1987, 74-5).

TAX RATES

Progressive Rates

All taxable income in excess of the minimum taxable level is subject to tax at progressive income tax rates. Table 5.4 illustrates the schedule of statutory tax rates at all levels of government in 1999. National income tax rates start at 10 per cent of taxable income up to ¥2.0 mil lion, rising to a top rate of 37 per cent above ¥1.8 million. By contrast, local income tax rates, consisting of both prefectural and municipal rates, are relatively simple. Prefectural income tax rates are 2 per cent on the first ¥7.0 million of taxable income, and 3 per cent above it. The rates for the standard municipal income tax begin at 3 per cent and rise to 10 per cent on taxable income above ¥7.0 million. Local income taxes cannot be deducted from the national income tax base. Thus, the combined marginal tax rate seems to have a significant effect on the attitude of taxpayers. At the top income bracket, it reaches 50 per cent which was a punitive level of 65 per cent until 1998.6 Table 5.4

Statutory rates of income taxes at all levels of government, 1999

Taxable income (¥m.)

under 2.0 2.0-3.3 3.3-7.0 7.0-9.0 9.0-18.0 18.0 and over

Tax rates (%) National

Prefectural

Municipal

Total

10 10 20 20 30 37

2 2 2 3 3 3

3 8 8 10 10 10

15 20 30 33 43 50

Note: Since the exemption levels are different for all three income taxes, the taxable income for any specific taxpayer is not exactly the same in each bracket. Source: As Table 5.1.

6 When the top rates were 70 per cent in the national income tax and 18 per cent in the local income taxes (i.e. a combined rate of 88 per cent), in 1986, a maximum amount of income tax was set to mitigate the total tax burden at the effective rate of 78 per cent. The adjustment was made in the prefectural and municipal income taxes to maintain that level. This method, however, was repealed after Sept. 1987. In addition to the income tax based on the progressive rates, there are per capita taxes levied by local governments. For a detailed discussion, see Ch. 18.

Table 5.5 Major changes of tax rates, national income tax, 1949-1999 1949 Tax rates %

1962

1950 Taxable income classes (¥000)

Tax rates %

Taxable income classes (¥000)

Tax rates %


8 10

100 300

Tax rates %

10

600

12 14

600 900

16

1 200

18

25

20

40

20

25

50

80

70

30

100

30

1800

35

100

35

120

35

2500

40

150

40

150

40

4000

200

45

200

500 50 50 250 55 500 55 300 60 500 65 700 70 1000 75 2000 5000 80 86 5000 No. of income brackets 14 8

45 50 55 60 65 70 75

1988


Tax

rates %


10.5 12 14

500 1 200 2000

10.5 12

1500 2000

16

3000

17

3000 20

5000

Tax rates %

1989

1995


Tax rates %


21 24

2000 2500

27 30 34

3000 3500 4000

38

5000

42

6000

46 50 55 60 65 70 75

8000 10000 20000 40000 60000 80000 80000

6000 10000 20 000 30000 45000 60000 60 000

15

10

3000

10

3000

1 500

1 200

30

45

Tax rates %

800

20

25

1987

Tax rates %


Tax rates %


500

15

20


1984-86

19

21

4000

25

6 000

25

6000

30

8000

30

8000

35

10000

35

10000

40

12000

40

12000

45

15000

45

15000

50 55 60 65 70

20000 30000 50000 80000 80 000

15

50 55 60

30000 50000 50000

12

20

6000

20

6000

30

10000

30

10000

40 50

20000 50 000

40 50

20 000 20000

60

50000

6

Source: MOF, Primary Statistics of laxation (Zeisei Shuyo Sanko Shiryoshu), Feb. 1998 and MOF, Outline of the 1999 Tax Reform.

5

10

3300

10

3300

20

9000

20

9000

30

18000

30 37

18000 18000

40 50

30000 30000

b

4

Basic Structures

83

Tax rate structures have varied greatly in Japan throughout the postwar periods. Major changes in national income tax rates are tabulated in Table 5.5. In 1950, a drastic change was enforced to make the income tax less progressive than it was under the influence of the Shoup proposals. The mitigation of progressiveness was carried out in conjunction with the adoption of the 'net worth tax' supplementary to a comprehensive income tax. After the repeal of the net worth tax in 1953, however, the rate structure again became more progressive with an increased number of income brackets. This trend continued until 1984, when the reverse movement began to emerge from general support for tax structures that have a smaller number of rate bands and lower marginal tax rates, particularly at higher-income classes. As a result of flattening tax rates in line with the movement of world tax reform, Japan's rate structure was simplified with five brackets from 1989 and further reduced to four brackets in 1999. Figure 5.1 shows the marginal tax rates implied by the statutory tax schedules in five major countries in 1998. The diagram depicts the rates that apply to a wage earner with a family of four and an income level of up to ¥10 million. Clearly, rate structures up to ¥10 million are by and large the same in the UK, the USA, and Japan in comparison with Germany. In particular, the long, basic rate band in the UK is conspicuous, so that the basic marginal rate applies to most taxpayers. The USA and Japan show similar patterns, although the basic bands are much narrower. In contrast, in Germany a non-linear rate begins to rise sharply after the basic rate band. The flatter rate structures in the USA and Japan were introduced by tax reforms. As a matter of fact, in 1986 the initial tax rate was lower and started rising more or

Fig. 5.1 Marginal rates of income tax (excluding local taxes) up to an income of ¥10 million for a family of four with a single wage-earner, selected countries, 1998 The US figure includes the state income tax for New York (but excludes the New York City tax). Exchange rates used for the conversion of yen are those for 1998: $1 =¥130, £1 =¥215, and DM1 =¥72. Source: MOF Tax Bureau, data presented to the Tax Advisory Commission, with some modifications.

84


less continuously. The main features of Japan's rate structure included a higher level of minimum taxable income, and an excessive number of marginal tax rates with narrow rate bands, but the progressive tax structure has recently been flattened to a great extent. Application of Separate Tax Rates So long as we maintain a global tax system of comprehensive income taxation, it is not entirely necessary to separate certain income sources from others for tax purposes: all incomes should be aggregated and taxed at progressive rates. This was actually realized under the aggregation approach proposed by the Shoup Mission. Today, however, many incomes are taxed separately at specially reduced rates. These separate taxation methods have been developed mainly to promote policy objectives. The majority of the ten taxable incomes described above are subject to forms of separate taxation. The most important of these is the application of a separate flat tax rate on interest derived from taxable personal savings. We shall consider these points in greater detail in Chapter 8. A similar procedure applies to dividends. Retirement income and timber income are also taxed separately under progressive tax rates. Only half of retirement income, after the special deduction for retirement is applied, is subject to taxation, and this is separated from other incomes. Timber income is handled by a sophisticated procedure of averaging; after allowance for the special deduction, taxable income is first divided by five and taxed at the rate applicable for that one-fifth; the tax amount thus computed is multiplied by five to obtain the total tax due. Likewise, capital gains on the sale of real estate are taxed at flat reduced rates favourable to all taxable long-run gains. Fluctuating and extraordinary income, which are usually classified as 'occasional income', are eligible for the 'averaging taxation' method. The Income Tax Schedule The progressiveness of the individual income tax is the most important means of securing vertical equity in the income tax burden distributed between individuals whose circumstances are different. This is also concerned with the redistributive implications of income taxation among the rich and the poor. How should the tax burden be distributed among people in different economic circumstances? To answer this question, it is important to formulate a progressive income tax schedule which relates tax liability to income received. In general, progressive income taxation is defined as taxation in which the proportion of income expropriated as tax rises as income increases. This implies that the average tax rate must increase with income, although the marginal rate of tax need not also increase with income: a flat marginal rate with a basic allowance can produce substantial progressivity (see Kay and King 1986, 211).

Basic Structures

Fig. 5.2

85

An exposition of Japan's income tax schedule

Figure 5.2 is an expository diagram of the income tax schedule that has been adopted in Japan during the postwar period, as derived from Table 5.5. The marginal rate of tax tm rises sharply at lower income levels, and begins to increase more gradually as it approaches the top rate t at income OA. At incomes above OA, tm maintains a constant level of f . The average rate of tax ta also shows a gradual upward movement, but it consistently maintains a lower level of tax rate than tm. Therefore, the income tax in Japan is progressive on the grounds that tm is higher than ta. Effective Progressivity The progressive income tax schedule illustrated in Figure 5.2 is based on the statutory tax rate schedule used thus far. In actual practice we may see a different progressivity of income taxation from that suggested by a statutory rate, because the income tax liability is not determined only by a statutory schedule of progressive tax rates. Now an attempt is made to seek empirical evidence of the effective progressivity involved in the income tax system. As a point of departure, let us construct a simple cross-section model. Tax statistics yield taxable income Y, number of taxpayers N, and taxes T by income class. Let us postulate a relation among 7;, N,, and T, for income class i:

86


A priori, (3 > 1 is hypothesized in (5.1), because the present income tax system is progressive. Expression (5.1) shows that taxes per capita are related to income per capita in a progressive tax function. Now, setting 11= TIN and y= Y/N, (5.1) is simplified to

The value of |3 is an indicator of tax progressivity.7 The required data, Y, N, and T, are available from two sources prepared by the National Tax Administration; one is the self-assessed income tax on the selfemployed, agricultural, and professional incomes, and the other, the withholding income tax on wage and salary incomes. (For a more detailed explanation, see the appendix to Chapter 6.) I estimated (5.2) for each of two types of the income tax for the period of 1951-96, depending upon two different tax sources. My empirical results are highly significant in statistical terms. (They are omitted because of space limitations.) In Figure 5.3, the values of p show the level and variation of income tax progressivity over three decades.8 As hypothesized above, (3 > 1 is verified. It should be noted first of all that the values of (3 are not at all stable. In other words, the tax function itself has changed. It is probable that these values would have been much more stable had the tax system been fixed in the postwar period.9 The tax function has been volatile, partly because of changes in the income distribution and other factors, but mostly because of the influences of annual tax changes. In other words, these changes in the tax function measure the effects of tax changes on the income tax structure. In the period under study, there is only one year in which no actual tax change was imposed: 1960. This is because the same income tax system was applied in both 1959 and 1960. We should therefore get approximately the same coefficient estimates for these two years. The values for (3 are relatively stable in the years 1959 and 1960 in Figure 5.3; thus, the theoretical expectation that the tax function is stable when the tax system is fixed is substantiated. (3 did vary sightly between the two years, perhaps because of changes in the income distribution and other factors. Except for these two exceptional years, (3 varied considerably from year to year. Since 1951, the income tax on wages and salaries has shown many variations, dropping by about 20 per cent, from 2.09 in 1957 to 1.71 in 1962, and then rising 7 The value of (3 also indicates the income elasticity of tax revenue, as discussed in Ch. 3. From this standpoint, Ishi (1976) calculated p in a similar way. 8 Of course, (3 in (5.2) is not due exclusively to the effect of income tax progressivity: changes in income distribution and other factors also influence (3. If we are to isolate the influence of income tax progressivity, we should use Y and T from the data of the statutory tax schedule. However, here we are interested in measuring the tax structure in actual practice. Thus, 1 have decided to use actual data derived from tax statistics. 9 Factors other than tax changes influence (3. Thus, even without tax changes, (3 would not be stable, in view of the nature of the data used in this study. However, the postwar US tax function, estimated from the same type of data, shows that (3 is virtually unchanged with the fixed tax system; see Snowborger and Kirk (1973, 242-3).

Basic Structures

87

Fig. 5.3 Variations of progressivity of income tax, 1951-1997: (a) self-assessed income tax; (b) income tax on wages and salaries but not returning to the high level of the late 1950s. In the 1970s and 1980s, the level of P moved within a narrow band of 1.9-2.0, registering many ups and downs. However, it began to decline sharply below 1.9 in the 1990s due to successive tax reductions. In contrast, the self-assessed income tax has revealed a somewhat different pattern of change. The values of P stagnated in the late 1950s and the early 1960s, fell briefly, and then quickly rose for a while. This pattern differs somewhat from the variations in income tax on wage and salary incomes since 1970. One important fact that has emerged from my estimates is that the progressiveness of the income tax in Japan has been manifest in terms of the effective progressivity,10 and 10 It is necessary to explain that (3 for the self-assessed income tax is rather stable and its level is generally lower. There are two reasons for this worth noting. For one thing, tax avoidance and evasion can be used more easily to assess income liability in the self-assessed income tax, although it is difficult to prove

88


that the tax progressivity maintained a relatively stable level from the late 1970s until about 1990 without any change in income tax structure. Are the variations in p observed in Figure 5.3 related to the direction of tax rate changes? Looking over the changes in income taxes for 1951-97, no changes occurred in the progressive tax rate schedule in 27 of these years, while tax rates were reduced in 17 years and were increased in 3 years.11 Taxes were reduced mostly in the late 1950s, the early 1960s, and the early 1970s. These variations in the progressive tax rates correspond to the fall in p. On the other hand, tax rates were changed in increasing revenues in 1967, 1968, and 1984, and they remained unchanged in most other years. These rate increases seem to have raised p and elevated the progressivity of income taxes. A flattening tax rate structure in the 1990s seems to have reduced the value of p, particularly in the case of income tax on wages and salaries. However, systemic changes in tax rates and variations in p do not always show exact correspondence. For example, 3 reached a peak in 1957 for the income tax on wage and salary incomes even though that year's tax reduction was one of the largest. This lack of correspondence may be due partly to changes in allowable exemptions and deductions and partly to factors other than tax changes influencing p. Tax Progressivity by Income Class

We have charted the value of p in relation to tax progressivity. Cross-section estimates of p based on Figure 5.2, however, are for all income classes taken together. Since not all income classes have the same P, it is necessary to examine differences in p among income classes. As an illustration, we take 197012 and calculate p for each income class under the self-assessed income tax and the income tax on wage and salary incomes. By definition, p is an indicator of income tax progressivity. Slitor (1948) suggested another indicator, denoted by 8, which is defined as

where ta = average tax rate, tm = marginal tax rate, and Y = average income of an income class. Income tax progressivity increases as the difference between tm and ta increases. 8 can be computed when Y is given an appropriate measurement unit. A comparison of 8 of different income classes enables us to determine their degree of progressivity. Since p is defined as tm/ta, P is closely related to 8. this accurately. For another, the conceptual difference of taxable income between the two types of income taxes appears to explain the different level of pi. For a more expanded discussion, sec Ishi (1976). 11 Tax reductions arising from rate changes were undertaken a number of times, but their amounts were much smaller than those of exemptions and deductions. 12 Other years show more or less similar results.

Basic Structures

Fig. 5.4

89

Tax progressivity by income class: self-assessed income tax, 1970

8 and (3 are shown by income class in Figures 5.4 and 5.5. The shaded area shows tm — ta] when the area widens, tax progressivity increases. The two diagrams lead to some interesting observations. 1. For the self-assessed income tax, neither tm nor ta shows an upward trend, even at high levels of income. Both are already stabilized at an income of ¥13.5 million and even decline slightly towards the highest income class (average income, ¥42.9 million). The distance between tm and ta does not widen with the increase in income. 2. The income tax on wage and salary income, on the other hand, shows rising tm and ta, with the gap between tm and ta widening. At the income level of ¥13 million the upward slope flattens, and the distance between tm and ta starts to narrow. 3. 8 and (3 also decline with increases in income in the case of the income tax on wage and salary incomes. However, they are higher than the comparable figures for the self-assessed income tax. Overall, the low level of (3 and its sharp decline at high-income classes for the self-assessed income tax are quite conspicuous. In particular, at the highest-income

90

Fig. 5.5


Tax progressivity by income class: taxes on wage and salary incomes, 1970

class, which includes individuals with incomes above ¥20 million, (3 is somewhat below 1. The low progressivity at high incomes is due mostly to the particularly favourable tax treatment of high-income individuals, such as special tax measures on interest, dividends, and capital gains.13 13

We note that both average and marginal tax rates are very changeable and are high over the three and four bottom income classes. This holds for other years, too. The changeable and high rates for individuals in these lowest income classes arc not representative of the tax burdens they pay over longer periods of time than a year, because in these classes there is a heavy concentration of people with temporarily low incomes. In order to eliminate this variability in income classes below ¥700 000,1 have computed averages of their Y and T weighted by numbers of taxpayers and then calculated their average

Basic Structures

91

GENERAL FEATURES OF THE PRESENT INCOME TAX

A Hybrid Income Tax: Income v. Expenditure as Tax Bases A great deal of controversy still rages over the choice between income and expenditure as alternative tax bases.14 It seems, however, that past discussion offers no decisive conclusions on either theoretical or practical grounds. As a great deal of literature is available on this debate,15 we need not concern ourselves with the various arguments here. As has been evident from previous chapters, the Japanese tax system has traditionally taken the view that income is the best index of an ability to pay. This bias towards income originated with the introduction of a comprehensive income tax at the inauguration of the postwar tax system. At the initiative of the Shoup Mission, the income tax base was defined to include almost all forms of income, with no major exceptions. In actual practice, Japan adopted a comprehensive tax base that expanded the economic concept of income as widely as possible (see Pechman 1986, chs. 4-6). Furthermore, such broadly based taxable income was taxed strictly at progressive tax rates under a global tax system. During the years following the war, the individual income tax began to be eroded considerably by numerous special provisions. It is interesting to note that the eroding process of the comprehensive income tax base has advanced mainly in the area of income from investment and savings. Most capital gains from the sale of stock were removed from the individual income tax base, and a large portion of interest income became non-taxable under a system of tax-exempt savings. In addition, dividends, capital gains on real estate sales, and taxable interest are taxed favourably at a separate flat tax rate. How should we interpret the present structure of income taxation? Japan seems to have deviated from the comprehensive income tax and phased in a hybrid income tax in which different sources of income are taxed at different rates. It can be concluded that schedular income tax has appeared which is much like the prewar individual income tax of 1940. Moreover, it is possible to justify the basic nature of the present individual income tax in the following way. The distinction between an income tax and an expenditure tax depends upon the treatment of savings. Under an expenditure tax, either savings or interest is excluded from the tax base. As a consequence, savings cannot be subject to taxation. Similar effects of an expenditure tax on consumption over the life-cycle can be achieved by excluding interest from the tax base. This and marginal tax rates; that is, 8 and (3 are computed for an aggregate of the four or five lowest income classes. Musgrave (1959fl, 2223) also stresses this point. 14 As the third tax base, wealth is equally important; but generally, taxes on wealth are rejected as a major tax source, particularly as a substitute for income or consumption taxes. 15 See e.g. Andrews (1974), US Treasury (1977), Institute of Fiscal Affairs (1978), Kay and King (1986), Pechman (1980), Aaron and Galper (1985), Pechman (J986).

92


argument was clearly explained by the US Treasury (1977), which preferred a pre-paid type of expenditure tax (i.e. taxing savings in advance but leaving interest tax-exempt). Either way, an expenditure tax does not discriminate against a pattern of consumption expenditure over a stretch of years, and the 'double taxation of savings' does not take place. Proponents of an expenditure tax often cite this as an advantage of choosing consumption expenditure as the tax base. The treatment of savings or investment income in Japan is almost the same as that which would be applied to all savings under an expenditure tax as described above. In effect, the Japanese individual income tax has been partially transformed into an expenditure tax. It can therefore be considered a hybrid of a comprehensive income tax and an expenditure tax,16 However, this hybrid has developed spontaneously, without any special attempt to avoid the double taxation of savings. Because of its split nature, the Japanese tax system has two distinctive options available for change in the future. One is to return in the direction of a comprehensive income tax by enlarging the tax base. The other is to move in the opposite direction towards some form of expenditure tax. Either direction can be defended theoretically. To be sure, as the third option we are also capable of preserving the current mixture of income and expenditure taxes (see Chapter 19). Income Tax Burden How do Japanese taxpayers bear the burden of the income tax? Table 5.6 provides some comparative information on the income tax burden among selected OECD countries. The highest income tax burdens are observed, not surprisingly, in Denmark, Sweden, and Finland, followed by Belgium and Canada. Both the USA and UK have relatively high ratios of taxes to the income of average taxpayers (APW), while their burdens relative to GDP are not ranked as high. What is of interest is that all these countries fell into two different patterns with respect to the change in the ratio of personal income taxes relative to GDP between 1970 and 1995: tax ratios sharply increased in Denmark, Belgium, and Italy, while they decreased in Sweden, Norway, and the Netherlands. Japan doubled the income tax burden during the quarter-century, like the first groups. Nevertheless, Japan shows the lowest figure as well as France, viewed from either a macro or a micro level of tax burden. Since France is a country in which indirect taxes predominate, it is easy to understand its rank as the lowest among major advanced countries. It is however interesting to learn that income tax burdens in Japan are the lowest by international standards, even though income tax yields Japan's major revenues. Perhaps the most important characteristics of the Japanese income tax system is the light burden that it imposes on the taxpayer. This is certainly true if we look at 16

The British tax system is more or less similar to the Japanese hybrid income tax; see Kay and King (1986, chs. 2-5).

Basic Structures

93

Table 5.6 Taxes on personal income3 as a percentage of GDP and at the income level of an average production worker (APW), selected countries As % of GDP

Denmark Sweden Finland Belgium Canada Australia Italy Norway Germany Switzerland OECD total USA UK Austria Netherlands France Japan

1995

1970

27.7 17.5 16.1 14.5 13.6 12.3 10.8 10.7 10.7 10.6 10.4 10.1 9.8 8.8 8.3 6.2

19.6 19.8 12.8 8.7 10.1 9.0 2.8 12.3 8.8 7.5 8.5 10.0 11.7 7.2 9.9 4.2 4.2

6.1

As % of the APW income 1985

34.3 33.9 25.3 17.7 10.2 16.8 14.0b 15.0 10.9 6.4 15.3 17.9 7.6 8.1 0.4C 2.8

* Includes taxes on income, profits, and capital gains (OECD classification 1100). 1984 figure. c 1983 figure. b

Source: OECD, Revenue Statistics of OECD Member Countries, 1965-96, 1997; Taxation in Developed Countries, 1987.

the low tax burden on the income of average workers. Individual income tax rates in Japan are not so low as those of other countries on a statutory basis (see Figure 5.1), but the actual tax paid by each taxpayer is extremely modest by international comparison. One of the most important reasons for Japan's low tax burden is related to major structural features of the Japanese income tax. In addition to a high minimum taxable level, generous exemptions and deductions for various incomes are allowed, and property income is favourably treated for tax purposes. Attention must now be directed towards the phenomenon of the erosion of the individual income tax in relation to the moderate tax burden.

6

The Erosion of Individual Income Tax For many years, individual income tax in Japan has comprised nearly 40 per cent of total national tax revenues. Because of its progressive character, most Japanese have accepted it as the most equitable means of distributing the tax burden among the people. In recent years, however, it has come to be greatly criticized, ironically from the standpoint of tax equity. This is because the Japanese individual income tax has been eroded by a number of special tax measures. These special provisions have resulted in the erosion of both the tax base and the tax yield, and tax erosion clearly impairs the equity of the existing income tax system. The main objective of this chapter is, first, to estimate the extent of income tax erosion by income class for 1972, 1975, 1983, 1984, 1987, 1991, and 1996 from Japanese tax data, and then, using the results, to choose a standard by which to measure the distributional effects of tax erosion and examine the extent of tax equity in the Japanese income tax system. The major conclusion of this study confirms similar studies made in the USA.

THE RELATIONSHIP BETWEEN INCOME TAX EROSION AND TAX EQUITY

Equity Considerations Tax experts view tax equity or fairness in two dimensions, vertical and horizontal. Vertical equity is concerned with the distribution of the tax burden among different income classes. Progressive taxation of all incomes on an all-inclusive basis is an essential condition in allocating different tax burdens fairly between the rich and the poor. On the other hand, horizontal equity is concerned with equalizing the tax burden among people in similar economic circumstances. This requires that the income tax base be as broad as possible. The Japanese income tax system has violated the principles of both vertical and horizontal equity to a considerable degree. This is due mainly to the fact that many special provisions have been introduced into the income tax law benefiting select groups of taxpayers. For instance, the special application of reduced tax rates to specific types of investment income favours higher-income classes, impairing vertical equity. Likewise, a number of special exclusions, deductions, and tax credits This chapter draws heavily upon Ishi (1979, 1989a).

The Erosion of Individual Income Tax

95

diminish the income tax base and create large differences in tax burdens among people with equal incomes. Obviously, the concept of tax equity and the phenomenon of erosion are closely interrelated. A study of the extent and the distribution of tax erosion would enable us to judge whether or not a specific income tax system can be regarded as equitable. Since tax erosion arising from special tax measures is actually disguised payment to individuals, it is often referred to as 'tax expenditures'. For more than two decades, 'tax erosion' and 'tax expenditures' have been frequently quoted terms in academic discussions on tax issues, and have attracted the attention of tax experts in many countries.1 There seem to be two reasons for this world-wide interest. First, there is a general trend to re-examine present tax systems in the light of tax equity. Second, increasing attention has been paid to comparing the roles of tax expenditure with those of direct subsidies. Eroding Aspects of the Present System

The fundamental structure of the present Japanese tax system was established upon the recommendations of the Shoup Mission in 1949 in the form of a progressive and broadly based income tax. The Shoup proposal dictated that a progressive tax rate structure be applied to a comprehensive definition of'taxable income'. In principle, all types of income were included in the comprehensive tax base, with no exceptions. For instance, capital gains from the sale of securities, or occasional income (e.g. winnings from horse races, prizes won in competitions, etc.), were fully subject to income taxation. Such a broad definition of 'taxable income' rendered the original postwar tax system all-encompassing. The major motivation for altering this system was to stimulate the achievement of specific policy goals, such as capital accumulation or the improvement of social welfare. In so doing, the government has often maintained that the deviation from equity is in the national interest, and that special provisions in the tax law are the most efficient way of achieving specific national goals. Income tax reforms enacted in the postwar period have included special measures such as 1. the full or partial exclusion of certain items of income from the tax base; 2. the separate application of special reduced tax rates to certain items of income; and 3. an increase in the scale and scope of deductions, exemptions, and tax credits. All three categories fall under the special tax measures, which were designed to give preferential treatment to specific sources of income and were motivated by 1 See e.g. Surrey (1973, 1981), Surrey and McDaniel (1985). The 1976 annual congress of the International Fiscal Association dealt with the topic, 'Tax Incentives as an Achievement of Government Goals', and, similarly, the 1977 annual conference of the International Institute of Public Finance considered the issues, 'Subsidies, Tax Reliefs, and Prices'. Also Surrey and Sunley (1976), reporting to the 1976 congress of the IFA, state that the official data published by the Japanese government is incomplete.

96


policy goals. At the time of enactment, these changes seemed appropriate to both the Diet and the government. However, the special tax measures tended to eliminate huge portions of the tax base and set a precedent for enacting still more measures to satisfy special interests. In the postwar period, therefore, the proliferation of special tax measures has eroded the tax base to a substantial degree. Although some of these special provisions have been phased out, a number remain in all three categories. Typical in principle (at the time of writing in 1999) of the first type is the full exclusion of a certain type of interest income for aged groups, although a flat rate on interest income was uniformly introduced in April 1988. As an example of the second type of special tax measures, income from interest, dividends, lump-sum severance pay, and the sale of timber, as well as capital gains from the sale of securities and land, were taxed separately from other incomes at lower tax rates. This provision for separate taxation clearly deviates from the spirit of the Shoup tax system, in which all incomes were aggregated for the purpose of income taxation, regardless of their source. There are numerous existing examples of the third category of special tax measures. While the 1950 tax law permitted only several exemptions, such as the basic exemption, the exemption for dependants, and the deduction for earned income, the present income tax law institutes more than 20 different exemptions, deductions, and tax credits. The introduction of such special tax measures has contributed to the erosion of the tax base and has greatly reduced income tax revenues. Enactment of these measures has also been detrimental to the equity and fairness of income taxation in Japan. In what follows, it is assumed that it is most important to terminate the erosion of the present income tax system and to achieve income tax reform which would promote greater vertical and horizontal equity. Every departure from the comprehensive income tax has narrowed the tax base, and keeps tax rates higher than they need to be. Also, the present system violates the principle of tax equity, and makes the income tax structure more distorted and complex. The major concern of this study is to measure how much the income tax base has been eroded, thereby causing revenue loss.

P R O C E D U R E S FOR THE CALCULATION

OF TAX EROSION

Definition of a Comprehensive Tax Base Estimates of tax erosion measure the difference between the income that is subject to taxation under the current law and the tax base that would result from an elimination of special provisions. Examination of tax erosion thus depends on the calculation of a comprehensive tax base. As a basis for this calculation, it is important to 2 Of course, we can choose a different direction of tax reform if the expenditure tax is preferred to the comprehensive income tax.


97

have a comprehensive definition of income.3 Past studies have traditionally adopted the definition of 'consumption plus the net increase in the value of assets during the year'. This concept of income has been defined to correspond as closely as possible to the economic income concept of Haig and Simons (see Simons 1938). It is impossible, however, to achieve a measure of income that corresponds exactly to this comprehensive concept. Difficulties in measuring such components as accrued capital gains, income in kind, and imputed rent, for example, mean that we cannot compute a truly comprehensive measure. Moreover, available data, particularly those arranged by income class, prevent us from including items which in theory should be counted as income. Thus, while applying the 'economic' income concept, we are forced to modify it in our measure of comprehensive income. In practice, it is conceptually problematic to define how far we should go in eliminating the special tax provisions in broadening the tax base. It is very difficult to decide which special provisions to retain and which to eliminate. My own preference is to approximate the comprehensive tax model as much as possible and to derive a modified concept of comprehensive income from such a model. Eroding Provisions

Erosion of the tax base has resulted from two provisions in the income tax law: those allowing for (1) exclusions, and (2) deductions and exemptions. Amounts excluded in accordance with these two provisions are added to the existing tax base to arrive at a measure of the comprehensive income tax. The specific types of income that came under these categories in 1984 are as follows: 1. Exclusions (a) interest exclusions: interest accruing from postal savings, bank current deposits, and government bonds, each not exceeding ¥3 million in principal or in total face value; interest income or distribution of profits for the formation of employees' assets not exceeding ¥5 million;4 (b) dividend income: non-taxable portions of distribution of profits from securities investment; (c) half-exclusions (or half-taxation) of aggregate capital gains5 and occasional income; (d) special preferential deductions: ¥500000 per taxpayer for timber or occasional income, and aggregate capital gains; ¥1 million for long-term separable capital gains; and (e) special deductions for retirement income (see Chapter 5). 3 For a discussion of this issue, see Royal Commission on Taxation in Canada (1966), Bittker et al. (1968), and Surrey (1973,1981). 4 For a more expanded discussion, see Ch. 8. ' This category of capital gains excludes those gains treated with the separate taxation method under the special tax measures law.

98


2. Ordinary deductions and exemptions. About 20 kinds of ordinary deductions and exemptions are encompassed in the income tax law. Typical examples are the basic exemption, exemptions for spouse and dependants, deductions for medical expenses, and deductions for charity contributions or donations. The latest estimate in 1996 was made by a slightly different category of exclusions, mainly reflecting the fact that taxation on interest income and capital gains on securities was drastically changed in 1988 and 1989 respectively (see Chapter 8 for a detailed discussion). Briefly, the scope of interest exclusions was restricted to groups over 65 years of age and disabled people, leaving the limit applicable to non-taxable treatment unchanged (i.e. ¥3.5 million for bank deposits, post office saving and public bonds or ¥5.5 million for savings for the formation of employees' assets). In connection with these modifications to the current tax base, it is necessary to explain two particular parts of my calculations, because some aspects of my computation of the comprehensive tax base might generate conceptual debates. First, I consider all ordinary deductions and exemptions as tax erosion from the comprehensive tax base. Had I not done so, it would have been necessary to classify all items into 'rational' or 'irrational' sources of erosion (see Pechman 1984). However, such a classification poses the problem of borderline cases in which arbitrary judgement intervenes. To avoid the problem of such an arbitrary classification, I chose a broader definition of tax base erosion (in terms of all deductions and exclusions) than has been utilized in other studies.6 Second, transfer payments are not included as items contributing to the erosion of the tax base for two reasons: (1) there is a lack of reliable data by income class; and (2) most recipients would not be subject to tax. The next step is to calculate the erosion of tax revenues. Although this definition is far from satisfactory, the comprehensive income tax is defined as the actual revenue plus the additional revenues that would result from a larger tax base. There are three sources of additional tax revenues: 1. tax revenues on the additional income items listed above—on exclusions, deductions, and exemptions; 2. tax revenues that would be collected if separate taxation of interest, dividends, retirement income, and capital gains on land7 (including securities in the 1991 and 1996 cases) were replaced by progressive taxation on an aggregate basis; 3. tax revenues raised from the elimination of the following tax credits: credits for dividends, foreign taxes, incremental R&D expenditures, and acquisition of a dwelling house.8 6 In fact, Pechman (1959, 251} excluded personal exemption, which is equivalent to the basic exemption in Japan from the items that promote erosion. 7 During the years of my estimation, four kinds of income were taxed separately from other income sources under the special tax measures. For a detailed explanation, see Ch. 5. 8 A tax credit for savings to purchase a house was admitted until 1982.


99

ESTIMATES OF TAX EROSION BY INCOME CLASS

Two Effective Rates of Tax The procedures described above were used to estimate the comprehensive tax base and yield for Japan in 1972, 1975, 1983, 1984, 1987, 1991, and 1996. Because of space limitations, only the results for 1996 are discussed here.9 Since the data available are scattered and relatively poor, painstaking efforts are necessary to attain a final estimation. (See the appendix to this chapter for data.) Tax erosion is analysed in terms of the effective rate of tax by income class. The 'effective rate of tax' is here defined as the ratio of tax liability to the comprehensive tax base. Comparison is made between the effective rate of tax under a comprehensive income tax (after elimination of all special provisions) and the actual effective rate of tax (the ratio of actual taxes to a comprehensive income). This comparison reveals the extent of erosion caused by the special provisions of the income tax law. Since the effective burden of statutory tax rates is altered by the existence of tax eroding provisions, the usual practice of examining the movement of nominal tax rates by income class does not clarify the true distribution of the tax burden. An examination of the difference between comprehensive and actual effective rates of tax over income classes reveals the distribution of the benefits of tax erosion, and provides a better tool for judging the equity of the tax system. It is generally believed that tax-eroding provisions have greater impact at higher income levels. In particular, it is expected that the exclusion of interest and dividends and of the separate taxation of capital gains will primarily affect taxpayers in the highest brackets. Thus, tax erosion is assumed to benefit higher-income individuals more than taxpayers in the lower brackets, and the entire tax system is thought to be much less progressive than the statutory tax rates imply. Table 6.1 contains the comprehensive tax base and yield for 1996. Using these estimates, two effective rates of tax can be derived. A comparison of these two rates is displayed in Table 6.2. Major Findings The results presented in Figure 6.1 to some extent confirm our expectations. The changes in effective tax rates resulting from the adoption of the comprehensive income tax for taxpayers are illustrated at different income levels. The top line of the graph indicates the effective tax rates that would be paid under the comprehensive income tax, while the bottom line indicates the ratio of actual tax yield to the comprehensive tax base. The various shaded areas between the two lines depict the contribution of the major structural features of the 1996 tax law to tax erosion. 9 The latest year for which necessary data were available to me at the time of writing the third draft of this book was 1996. The results of the six selected years are slightly varied, and the extent of erosion has tended to decrese as the time has progressed. Detailed analysis for the years 1972 and 1975 appeared in Ishi (1979) and that for the year 1984 in Ishi (1989a).

Table 6.1 Comparison of actual and comprehensive income tax base and yield, 1996 (¥bn.) Income class ¥(m.)

under 0.7 0.7-1.0 1.0-1.5 1.5-2.0 2-3 3-4 4-5 5-10 10-15 15-20 20-30 30-50 50 and over All classes

Comprehensive tax base

Tax yield Actual

Increase in yield from:

(1)

(2)

(3)

(4)

(5)

Separate taxation (6)

1 140.6 2 594.4 6 754.7 11280.4 35 962.6 43 206.5 43 056.3 126 862.0 34799.5 11 365.9 6 002.9 5475.6 8 714.3 337215.5

12.6 29.8 121.1 183.7 851.7 1253.6 1 338.5 5425.2 3 185.3 1 600.9 1 145.5 1119.3 1889.8 18157.1

4.2 9.1 17.6 23.1 92.4 148.0 189.0 714.4 250.1 87.3 46.9 44.5 42.6 1 669.2

170.6 298.6 552.3 754.3 2261.5 2779.5 3221.6 10544.2 2582.9 689.4 331.1 248.6 153.6 24588.1

0.0 0.0 0.1 0.3 2.3 5.0 6.6 23.4 8.6 4.9 3.8 4.4 9.2 68.6

2.6 6.3 8.5 10.0 54.1 108.7 158.4 657.3 287.5 126.7 194.6 398.5 1335.5 3348.6

Note: (7) = (2) + (3) + (4) + (5) + (6).

Exclusions

Deductions and exemptions

Tax credit

Comprehensive

190.0 343.7 699.6 971.5 3262.0 4294.7 4914.1 17364.5 6314.5 2509.1 1721.9 1815.4 3430.7 47831.7

(7)

Table 6.2

Comparison between actual and comprehensive income tax rates, 1996

Income class (¥m.)

under 0.7 0.7-1.0 1.0-1.5 1.5-2.0 2-3 3-4 4-5 5-10 10-15 15-20 20-30 30-50 50 and over All classes

Tax erosion (%)

Actual income tax base (%)

Exclusions

(1)

1.10 1.15 1.79 1.63 2.37 2.90 3.11 4.27 9.14

14.06 19.07 20.45 21.93 5.32

Tax credit

Separate taxation

(2)

Deductions and exemptions (3)

(4)

(5)

Comprehensive income tax rate (6)

0.37 0.35 0.26 0.21 0.26 0.34 0.44 0.56 0.72 0.77 0.79 0.83 0.56 0.50

14.96 11.51 8.18 6.69 6.29 6.43 7.48 8.31 7.43 6.09 5.58 4.63 2.02 7.32

0.00 0.00 0.00 0.00 0.01 0.01 0.02 0.02 0.02 0.04 0.06 0.08 0.12 0.02

0.23 0.24 0.13 0.09 0.15 0.25 0.37 0.52 0.83 1.11 3.19 7.13 14.64 0.93

16.66 13.25 10.36 8.61 9.07 9.94 11.41 13.69 18.14 22.07 28.69 33.12 39.28 14.09

Note: Percentages in each column are computed as the ratio of tax yield to the comprehensive tax base in Table 6.1.

102


Overall, effective tax rates under the comprehensive income tax are considerably higher at every income level than actual effective tax rates. The main reason for the disparity, and largest single contributor to the gap, is the elimination of provisions for exemptions and deductions. Indeed, these items comprise a substantial proportion of the total erosion over all income classes. After deductions and exemptions, the next-largest addition to actual effective tax rates results from the elimination of separate taxation. In contrast, the elimination of exclusions and tax credits under the comprehensive tax system adds very little to effective tax rates. As expected, the impact of the comprehensive tax differs for taxpayers at different income levels. As Figure 6.1 shows, under the comprehensive income tax, the effective tax rate rises much more sharply with income level than does the actual effective tax rate, revealing the extent to which tax erosion destroys the vertical equity of the system and reduces the effective progressivity of statutory tax rates. The greatest differences between effective tax rates occur at the two ends of the income scale. An elimination of deductions and exemptions would result in a larger increase in effective rates for taxpayers with incomes under about ¥5 million, while the curtailment of separate taxation would increase tax rates substantially for taxpayers with incomes over ¥20 million. In other words, higher-income taxpayers benefit from the separate taxation provision, while lower-income taxpayers gain from the provisions for exemptions and deductions. On the other hand, the influence of erosion arising from exclusions from the tax base is equivalent over all income classes, partially reflecting the effect of special tax

Fig. 6.1

Income tax erosion by income class, 1996

Note: Figures in parentheses on the income scale are in US dollars, converted at $1 =¥125. The horizontal axis is drawn to a logarithmic scale.


103

measures for the exclusion of certain interest on small deposits and postal savings, which are not heavily concentrated at the higher end of the income distribution. Tax credits, mainly for dividend payments, benefit relatively higher-income taxpayers (over ¥10 million), but these credits have only a small effect on the overall picture. In proposing more equitable income taxation, it would be too controversial to suggest the elimination of all deductions and exemptions. Because the major beneficiaries of an across-the-board elimination of deductions and exemptions would be those in the lower tax brackets, we do not need to place so much importance on this source of tax erosion in order to promote tax equity. Instead, greater attention should be paid to the second-largest area of tax erosion, the separate taxation provisions. It is the separate taxation of certain items of income that is responsible for the largest income tax erosion among the higher-income brackets. This phenomenon is most outstanding among taxpayers with incomes of more than ¥30 million. Our findings on the extent of tax erosion in lapan are similar in a couple of respects to studies carried out in the USA (see Pechman 1986, ch. 5; Pechman and Okner 1972, 28). In both countries, tax erosion is greater at higher-income levels. In the Japanese case, major causes for tax erosion are the provisions for separate taxation of capital gains and certain other types of income. These correspond to the preferential treatment of capital gains and income-splitting provisions which were found to be the greatest sources of tax erosion benefits to upper-income taxpayers in the USA.

IMPROVING THE EQUITY OF INCOME

TAXATION

One Measure of Tax Equity So far, we have devoted our attention to estimates of tax erosion by income class. As is often pointed out, the extent of tax erosion reveals how inequitable the tax burden has been. The gap between the two effective tax rates as depicted in Figure 6.1 is very important in illustrating this fact. Using these tax rates, we can compute the ratio of tax erosion, defined as follows:

where t c =the effective rate of the comprehensive income tax, and f a = the effective rate of actual tax revenues to the comprehensive tax base. As is evident from the formula, the larger the value of 8, the greater the tax erosion is and, in turn, the more inequitable the income tax burden (see Ishi 1989). In Figure 6.2, the values of 8 are depicted for five years, 1972, 1975, 1984, 1991, and 1996. Looking at the movements between each year, we can clearly see that the provisions allowing for tax erosion tend to impair the vertical equity of the tax system, particularly in the 1970s. As we have indicated, high-income taxpayers receive more of the benefit from tax erosion, in terms of greater overall reduction in effective tax rates, than do low-income taxpayers.

104


Fig. 6.2 The ratio of tax erosion, selected years The ratios in 1983 and 1987 are omitted because they are almost the same as those in 1984 and 1991, respectively.

Most conspicuous, however, is the fall in the ratio for the upper-income brackets between 1972 and 1996. This implies that vertical equity has been recovering to a certain degree during these years. What has contributed most to this phenomenon? When we draw the same graphs as in Figure 6.1 for 1972, 1975, 1984 and 1991, the answer becomes clear. We can see that the elimination of separate taxation has played a major role in reducing the tax erosion for higher-income classes. Changes in the income tax law have reduced the scope of special provisions such as capital gains from the level reached in the 1970s. Special Treatment of Capital Gains on Land Of the types of income eligible for separate taxation at reduced rates, short- and long-term capital gains on land are the most important in estimating the phenomenon of tax erosion for selected years. Approximately 60-70 per cent of tax erosion of the top two income brackets in 1972 and 1975 is due to the separate taxation of capital gains on land through special measures in the income tax law, although these ratios have tended to decrease steadily. Special provisions were included in the 1969 tax reform to promote the supply of land and to discourage speculation on land sales. Until 1975, the new tax rates were applied to capital gains on land separately from other incomes: 10 per cent in 1969—71, 15 per cent in 1972-73, and 20 per cent in 1974—75 on long-term capital


105

gains, and 40 per cent on short-term gains.10 This provision of the 1969 tax reform had far-reaching effects, but also, it substantially sacrificed tax equity. The special treatment of capital gains on land was thus a crucial factor contributing to the income tax erosion in the first half of the 1970s. In accordance with the critical view of inequitable taxation, the government in 1976 began to make some modifications towards heavier tax burdens in the income tax law for long-term capital gains on land. In 1983-84, taxable long-term capital gains of ¥40 million or less were taxed at a 20 per cent tax rate as before. In contrast, taxable amounts in excess of ¥40 million (X) were computed as follows: ~(X—¥40m.) X regular tax rate. The regular tax rate is the rate applicable to the taxpayer from the statutory tax schedule, which depends upon the taxpayer's income bracket. Thereafter until recent years, capital gains taxes on land have continued to be levied more heavily to stop land price hikes. Such modifications of the special tax provisions on land sale have contributed significantly to a slowing of the erosion of income taxation. Their success is a major factor in explaining the decrease in tax erosion observed in Figure 6.2. In addition, in 1991 and 1996 a large scale reformation of land tax was enforced, in which capital gains taxes on land were further increased (see Chapter 13 for an expanded argument). This treatment is likely to be of great use in weakening the tendency to erosion of income taxation in higher-income classes. On the other hand, investment income, such as interest and capital gains from the sales of securities, was taxed more heavily in 1988-89, as noted above. In fact, instead of eliminating non-taxable measures, flat tax rates were adopted for such investment income which is separate from other incomes. Therefore, the scope for eroding the tax base by applying separate taxation must have expanded in 1991 and 1996 due to the new treatment of the taxation of investment income in 1988-89. However, as far as the results in Figure 6.2 are concerned, the extent of tax erosion in both years shrunk to a considerable extent, as compared with that in the previous three years. The effects of increasing the burden on capital gains on land is likely perhaps to surpass those of new taxes on investment income. More detailed attention should be paid to the variation of tax erosion in 1996, as compared with 1991; i.e. the increased erosion in lower-income classes while the eroding phenomenon almost disappears in the upper-income classes. We must stress two facts. One is the large scale of income tax reduction in raising deductions and exemptions in 1996 which has contributed to increased erosion among the lower classes. The other is that capital gains tax on land was further strengthened in 1996, reducing the effect of tax erosion in the upper classes. In addition, it is important to note that capital gains on land themselves greatly decreased in 1996, reflecting the sharp drop of land prices after the collapse of the bubble phenomenon. 10 For more detailed information, see MOF Tax Bureau (1977a). The discussion between long-term and short-term capital gains is based on capital gains from property acquired before or after 1 Jan. 1969. More precisely, this type of capital gains, short- or long-term, covers not only land but also building and the right to use land.

106


The preceding analysis has assumed that, in the change from separate taxation to full or aggregate taxation, all special measures for capital gains were eliminated. However, because capital gains on land differ markedly from ordinary income in that they cannot be realized every year, this premiss seems unrealistic. If all capital gains were subject to taxation only in the year realized, the taxpayer would be hit too heavily all at once. Instead, a modified treatment should be introduced for assessing taxes on such irregular income. For this reason, we should consider another case in which the averaging of such income for tax purposes is employed. If I make such an alternative assumption concerning the averaging method of timber income, I can develop a more realistic picture of the possibilities in the real world. (The relevant figure is omitted; for the 1975 estimate, see Ishi 1979.) Certainly, the decrease in the extent of tax erosion at higher-income levels is marked, but the remaining provisions for separate taxation still contribute considerably to eroding the effective tax rates for higher-income-earners. The estimates clearly show that provisions allowing for tax erosion violate the vertical equity of the tax system, although the extent of the erosion has gradually been reduced in the past two decades. High-income taxpayers derive more of the benefit from tax erosion, in terms of greater overall reduction in effective tax rates, than do low-income taxpayers. Moreover, inequity would be even greater if provisions for deductions and exemptions were not classified as tax erosion. Policy Implications Obviously, the elimination of tax erosion in higher-income classes is desirable, both in terms of equity and on fiscal grounds. For the same reasons, a broadly based income tax is more appropriate than the present income tax with its eroded tax base. In order to achieve such a broadly based income tax, the following three tax reforms were suggested in relation to the 1984 income tax. 1. Separate taxation should be reduced as much as possible by aggregating incomes currently taxed separately with other income and taxing the total accordingly. 2. Unnecessary deductions, exemptions, and credits should be eliminated where possible. In addition to such exemptions included in the special tax measures (i.e. the deduction for fire and other casualty insurance premiums, the deduction for life insurance premiums, the credit for acquisition of a dwelling, etc.), other provisions in the law for miscellaneous exemptions also should be eliminated. 3. Income that is currently non-taxable should be fully or partially taxed. This reform would entail the removal of all exclusions for interest income and dividends, and the full or partial taxation of capital gains from the sale of securities. In the 1991 estimate, item 3 listed above was partially enforced, although far from satisfactory. In addition to restoring the vertical equity of the income tax system, two other merits of a broadly based income tax that were suggested earlier should be recapitulated. The first is the increase in tax revenues that would be achieved


107

under a more broadly based tax structure,11 and the second is the benefit derived from tax rates, which could be cut to maintain a constant yield. If the restoration of a comprehensive tax base were co-ordinated with a reduction in tax rates, moving towards the flat-rate tax, the equity, fiscal posture, and compliance rate of the tax system would all be greatly improved. Lastly, the probable effect of one omission from my calculation must be mentioned. I was unable to incorporate the phenomenon of non-reporting and under-reporting into my calculations as a source of tax erosion. If I had been able to take account of this factor, tax erosion in all income classes would have been even more pronounced.

APPENDIX Data Sources

Generally speaking, the quality and quantity of basic data for Japan are much less adequate than those of the MERGE data used by Pechman and Okner (1972) in their estimations.12 In Japan, it is administratively impossible to obtain the necessary information directly from the tax return files. We are forced to employ official tax statistics published by the government and to adjust them to obtain estimates of the desired aggregates. Methods of collecting data and making estimates by income class are even more complicated. Two basic sources for income tax data used in estimating a comprehensive income tax were obtained from the National Tax Administration (NTA). Statistics on the Self-assessed Income Tax (Shinkoku Shotokuzei no Jittai) contains data for self-employed taxpayers, and Statistics on Private Wages and Salaries (Minkan Kyuyo no Jittai) covers data for workers in the private sector falling under the withholding system. In addition, the Annual Report of National Tax Administration (Kokuzeicho Tokei Nenposho) was an indispensable source of data. However, these official statistics alone were not sufficient to complete my estimations, and I was forced to seek unpublished data at the worksheet level from the NTA. Even though I had access to unpublished data from the NTA, my data are still unsatisfactory in several ways: for instance, the basic data have not been corrected 1 ' Needles to say, all estimates in this study are confined to the 'first-order' effects of the revenue loss arising from erosion under the income tax law. No attempt has been made to take into account possible induced changes in before-tax income sources or income received by tax-payers with 'zero-elasticity assumption'. Since I assume that taxpayers try to minimize their tax liability, in practice, these 'second-order' effects would reduce the additional yield below what might be expected on the basis of my estimates. 12 For their basic data, Pechman and Okner (1972) used the 'MERGE file', which combines information on 30000 families and single persons from the Survey of Economic Opportunity (SFO) with the data from 90 000 federal individual income tax returns. Since it contains both data for low-income SEO families who are not in the tax-filing population and income tax information for higher-income individuals, the MERGE file is a much more satisfactory source for a study such as this than the data available for Japan. Furthermore, it is corrected for non-reporting and under-reporting.

108


for non-reporting and under-reporting because no relevant information was available. Moreover, I did not have data on low-income individuals who do not file taxes. In spite of these significant deficiencies, I carried out estimations based on the income tax information available to me, as this was the only possible way of attempting to analyse the extent of tax erosion. Creating a Comprehensive Tax Base The steps in deriving the comprehensive tax base are summarized in Figure A6.1. (Note that I disregard the steps necessary to calculate the tax base by income class.) Since basic tax statistics provide data on taxable income only before deductions and exemptions, the calculations begin with these data (step .1 in the figure). However, not even these data are published directly in the statistics on self-assessed and wage/salary taxpayers. Substantial additional data collection and estimation are

Fig. A6.1

Steps in creating a comprehensive tax base, 1996.

Figures in parentheses show actual amounts in ¥000bn. for 1996.


109

required to separate out the amount of deductions and exemptions (step 2 in the figure). Although deductions and exemptions are partially available from the original data on those falling under the self-assessed income tax, the data from the withholding system contain no such information. Conversely, for wage and salary workers it is necessary to calculate separately the 17 deductions and exemptions provided in the tax law. Finally, the actual tax base (step 3) is estimated by subtracting step 2 from step 1. The amount of exclusions (step 4) has been calculated by estimating from the data collected for the five items in this category. The addition of these exclusions to the tax base in step 1 gives the estimate of the comprehensive tax base (step 5). The difference between the actual and comprehensive tax bases, at the top and bottom of the figure, is the erosion from the tax base. Results of these calculations are provided by income class in each column of Table 6.1. The estimate of the actual total taxable income for 1996 is ¥148000 billion. Under the comprehensive income definition, taxable income would rise to ¥337 200 billion, an increase of 2.27 times. Erosion from the Tax Base and Yield

Estimates of the comprehensive tax yield and tax erosion can be obtained from the comprehensive tax base. Estimates of two types of tax erosion are also presented in Table 6.1. First, erosion of the tax base is computed as the sum of exclusions, deductions, and exemptions. If taxed at 1996 rates, these additions to the tax base would increase tax collection by ¥1669.2 billion and ¥24588.1 billion, respectively (see Table 6.1). Second, increased liability under a comprehensive income tax does not result solely from increases in the taxable income base. Tax credits and separate taxation represent erosion of the tax yield, and an elimination of these provisions would have a substantial impact on the total amount of taxes paid. While elimination of tax credits alone would add only ¥68.6 billion in total to the tax yield, the termination of special provisions for separate taxation would produce a much larger increase of ¥3 348.6 billion.

7

Inflation Adjustments Inflation has probably had the greatest impact on the structure of the Japanese individual income tax system during the postwar period. At least, it was so until the late 1980s. It is widely acknowledged that the major goal of income tax reductions was to adjust for the heavier tax burdens arising from inflation. Inflation causes significant increases in the effective rate of the individual income tax on real income if there are no adjustments for inflation. The Japanese government attempted to reduce the extra income tax liability arising from inflation nearly every year during the postwar period until the late 1970s, but it has ceased such adjustments since then. Although the rate of inflation declined in the 1990s, it produced distorting effects on income tax burdens until then. It is therefore important to identify how effective annual tax reductions were in adjusting the income tax for inflation, and to observe what has happened after inflation adjustment devices were dismantled. The following discussion is divided into three parts. In the first, an attempt will be made to clarify the mechanics of inflation adjustment (i.e. tax indexing), using the 1975 income tax system as a base. The second section examines the impact of the government's decision deliberately to pursue annual tax reductions rather than introduce tax indexing. Finally, a rough estimate is presented of the fiscal dividend that would have been produced by tax cut measures, and this is followed by a brief consideration of policy issues. Needless to say, focus is upon the estimation during the inflation period before 1990. INFLATION AND THE TAX SYSTEM

Distortion of the Individual Income Tax Inflation affects tax liabilities in various ways. Above all, it has a tremendous impact on an individual income tax with a progressive structure of tax rates, exemptions, deductions, and credits. The main aim of this chapter is to explore the effect of inflation on the individual income tax in Japan.1 This chapter is primarily based on Ishi (1981). 1 Of course, it is also important to pay attention to the effect of inflation on the corporate income tax. The difficulty of measuring particular items for tax purposes—e.g. capital gains, business income, interest—complicates the calculation of the tax base of corporate income tax during periods of inflation. Also, indirect taxes require some adjustment for inflation in the case of specific duties; see e.g. Maxwell (1955), Aaron (1976), Kay and King (1986).

Inflation Adjustments

111

In general, inflation induces distorting effects on both the tax base and the rate structure of income tax. As a result, tax authorities attempt to take all steps necessary to adjust for inflation. Income tax laws are generally written for a world of stable prices. Many of their provisions are expressed in fixed nominal amounts, irrespective of variation in price levels. Under the 1992 tax code in Japan, for instance, taxpayers pay 10.0 per cent of the first ¥3 million of taxable income; basic exemptions and deductions for dependants are ¥350 000 each. When prices rise, the real value of those and other quantities is reduced. At the same time, if wage incomes increase proportionally and thus remain unchanged in real terms, income tax liability automatically increases because the rate structures are progressive. Past increases in inflation have raised the question of whether and how the income tax system should be adjusted to cope with the rising trend in prices. Many countries have adopted the practice of indexing their tax structures automatically as price levels change (see Tanzi 1976). Other countries, including Japan, still pursue ad hoc remedies by periodic legislation to deal with the distortion of inflation on tax burdens. It would be interesting, in the case of Japan, to investigate how effective the deliberate measures for tax reductions have been in compensating for inflation. Inflation causes significant increases in income tax liability in two distinctive ways. First, it raises tax liability, because rate bracket boundaries are expressed in terms of money income. When prices rise, say, 10 per cent, a family's money income declines by 10 per cent in real value. Money income is usually elevated just enough to offset inflation (i.e. by 10 per cent) through wage hikes, in order to avoid a decline in real income. So, although real income remains unchanged, the family is thrown into a higher tax bracket, reflecting the illusory increase of money income. As a result, tax liabilities increase faster than inflation and reduce the increase in the family's real income. Second, an increase in tax liabilities is also caused by exemptions and deductions of fixed nominal amounts. The increment of a family's money income is fully included in progressive taxation with the levels of exemption and deduction fixed. Similarly, lower-income families who were below the minimum taxable level before inflation, and therefore held tax-exempt status, are often thrown into the tax rolls. Clearly, the structure of the individual income tax system tends to impose a higher tax burden on taxpayers whose real income stays the same in an inflationary world. Thus, when prices rise sharply over a long period of time, some means of price adjustment becomes necessary in order to reduce the increase of tax liabilities. Mechanism for Indexing Schemes How can the distorting effects of inflation on income tax liabilities be eliminated? One practical way to achieve this is to index the income tax system in order to offset or moderate the distortion of inflation on the calculation of taxation. There are

112


Table 7.1 Effect of 10 per cent inflation on the tax liability of a family of four with an annual income of ¥5 million, 1975 1975 level plus 10% inflation (¥000)

Income before exemptions and deductions (a) less: Basic exemption Exemption for spouse Exemption for dependants Employment income deduction Taxable income Income tax liability (b) Effective tax rate (b/a)%

Case 1: 1975 level

Case 2: No adjustment for inflation

Case 3: After adjustment for inflation

(1)

(2)

(3)

5000

5500

5500

260 260

260 260

286 286

520

520

572

1450 2510 332 6.64

1450 3010 410 7.45

1595 2761 365 6.64

Note: These exemptions and deductions are based on 1975 levels. Deduction for employment income and income tax liability are calculated using Table 7.2. several kinds of indexing schemes.2 Using one of them, we shall move to the question of how and whether the Japanese income tax system can be adjusted for the effects of inflation. Generally speaking, a 10 per cent increase in prices lowers by 10 per cent the real value of the rate brackets, exemptions, and deductions under the income tax law. Thus, a price escalator is seen as a useful scheme for eliminating the distortions caused by fixed nominal amounts in the tax law. Under such an indexing scheme, fixed monetary values are increased annually at a rate equal to that of inflation. For example, assuming 10 per cent inflation now, the basic exemption and the exemption for spouse would be increased from ¥350 000 to ¥385 000 each.3 Table 7.1 shows the results of this type of indexing scheme, applied to a family of four with a ¥5 million wage and salary income in 1975 level.4 It is assumed that 2 Tanzi (1976) argues that there are four different schemes for indexing a tax system, based upon his study of such schemes in a number of countries: (1) all statutory tax rates are lowered proportionally to eliminate the increase in revenue arising from inflation; (2) the increase in income attributable to inflation is deducted from the taxpayer's gross income; (3) taxable income is deflated to a base year; (4) price escalators are introduced into the income tax structure such that income tax rates apply to conslant real incomes rather than constant nominal incomes. As Tanzi correctly pointed out, only the last two methods represent well designed schemes. We shall use these two schemes for our analysis. 3 These figures are based on the 1989-92 tax law. They were ¥290000 in 1980, ¥300000 in 1983, and were further increased up to the previous level of Y330 000 in 1984.1 do not admit the necessity of updating the data for 1975 because it was in the midst of inflation adjustments for tax policy. 4 In constructing Table 7.1,1 am indebted to Sunlcy and Pechman (1976).

113


Table 7.2 Ten per cent inflation, statutory tax rate, and the deduction for employment income, 1975 Deduction for employment income3

Statutory progressive rate 1975 level (¥m.)

Tax rate (%) (3)

1975 level (¥m.)

(D

After indexation (¥m.) (2)

under 0.6 0.6-1.2 1.2-1.8 1.8-2.4 2.4-3.0 3.0-4.0 4.0 and over

under 0.66 0.66-1.32 1.32-1.98 1.98-2.64 2.64-3.30 3.30-4.40 4.40 and over

10 12 14 16 18 21 24

Deduction rate (%)

(4)

After indexation (¥m.) (5)

under 1.5 1.5-3.0 3.0-6.0 6.0 and over

under 1.65 1.65-3.30 3.30-6.60 6.60 and over

40 30 20 10

(6)

a

Minimum amount deductible, ¥0.5 million. Note: In regard to the statutory tax rate, only the necessary income brackets for our calculation are listed above.

this family claims only four exemptions and deductions; the basic exemption, the exemption for spouse, the exemption for dependants, and the deduction for employment income.5 Before inflation, the family would have to pay a tax of ¥332 000 on ¥5 million, and thus would register an effective rate of 6.64 per cent. Assuming that an inflation rate of 10 per cent has been accompanied by a 10 per cent wage rise, the family income would increase to ¥5.5 million. With no adjustment for inflation in column (2), taxes would rise to ¥410000 with an effective tax rate of 7.45 per cent. This amount represents the inflation-induced tax burden. In contrast, let us assume that the principal fixed monetary parameters are indexed, as shown in column (3). Table 7.2 includes indexing for both the statutory schedules of progressive tax rates and the deduction for employment income. The deduction for employment income allows a specific percentage of deduction for each income bracket. Thus, the boundaries of each income bracket, as well as those of statutory tax rates, must be elevated by 10 per cent. Under indexing mechanics, both income and taxes would increase by the same percentages, and so there would be no rise in the effective tax rate. Undoubtedly, such a method is successful in removing the distorting effects of inflation on income tax liability. If the same family earns a different amount of income, how would the situation be altered? Table 7.3 summarizes the same calculations applied to a family of four at different income levels. Columns (1) and (2) in case 1 of the table show the actual tax amount and the effective tax rate at the 1975 level, while columns (3) and (4) in case 2 present the effect of 10 per cent inflation with no indexing. Unless there is an adjustment for inflation, income tax liability clearly increases at each income level. 5 The minimum taxable level (i.e. tax threshold) is calculated by adding these four to the deduction for social insurance premiums.

Table 7.3 Ten per cent inflationary effects on income tax liability for a family of four, selected income levels, 1975 Income level3

After 10% inflation Case 1: 1975 level

3000 5000 7000 10000 15000 20000 30000 50000 100000

(3 300) (5 500) (7 700) (11000) (16500) (22 000) (33 000) (55 000) (110000)

Case 2: No indexing

Effect of not indexing Case 3: With indexing

% increase in tax

% point increase in effective rate

Amount (¥000) (1)

Effective rate (%) (2)

Amount (¥000) (3)


Amount (¥000) (5)


(7)

(8)

98 332 678 1411 3 130 5213 9959 20650 51 691

3.27 6.64 9.69 14.11 20.86 26.07 33.20 41.30 51.69

132 410 815 1655 3596 5941 11237 23 178 57769

4.00 7.45 10.58 15.05 21.79 27.00 34.05 42.14 52.34

107 365 748 1 554 3445 5737 10957 22717 56862

3.24 6.64 9.71 14.13 20.87 26.08 33.20 41.30 51.69

23.36 12.33 8.96 6.50 4.38 3.56 2.56 2.03 1.60

0.76 0.81 0.87 0.92 0.92 0.92 0.85 0.84 0.65

Notes: Calculation in Table 7.1 is applied to selected income levels. (7) = (3)-(5)-l. (8) = (4)-(6). ' Figures in parentheses are adjusted for inflation.


115

In contrast, the results for indexing major monetary variables after a 10 per cent inflation are observed in case 3. It is obvious from comparing column (4) with column (6) that indexation eliminates the distorting effects of inflation on income tax liability. The effects of inflation can be judged from the two sets of indicators in columns (7) and (8). In terms of percentage increases in (7), inflation has the greatest effect on tax liabilities at the lowest end of the income scale. If there is no indexing, lower-income classes encounter greater damage owing to the increase in nominal tax burdens caused by inflation. For instance, the tax payment for a ¥3 million income increases by 23 per cent when prices rise by 10 per cent, while that for a ¥30-50 million income rises only 2.0-2.5 per cent under the same conditions. However, as far as the change in tax burdens is concerned, it is more accurate to observe it in terms of effective tax rates. The effect of inflation on effective rates is much more uniform by income classes, as is seen in column (8). On the basis of these figures, the impact of inflation increases somewhat from the lowest income levels to about the ¥10—20 million income level, and declines thereafter. An explanation for this movement seems to be related to the progressive nature of the statutory tax rate. The effect of inflation is most markedly observed at the income level having the maximum progression of statutory effective rates.6

AN OFFSETTING EFFECT OF TAX REDUCTIONS

Estimates and Data The postwar Japanese economy has undergone very rapid economic growth, which has brought about a continuous rise in nominal income. Any rise in nominal income from economic growth increases income tax liability, and inflation accentuates this pattern. Unless there is an adjustment for the rise in income, tax burdens automatically increase, as we saw from the results of Tables 7.1 and 7.2. Consequently, every nation must take steps to arrest the automatic increase in the income tax burden. In Japan, the government has not provided automatic adjustments through indexing mechanics to offset the impact of inflation on income tax liabilities. Instead, it reduced taxes nearly every postwar year by undertaking deliberate tax cut measures until the mid-1970s, and only in 1960 and 1976 were there no reductions 6 The extent of rate progression is calculated by using the concept of the income elasticity of tax yield ET in the statutory tax schedule. As in Figure 5.2, ET is defined as the quotient of marginal rate tm by average rate ta: E 7 -=f m /t 0 . At the lowest income level, ET is equal to 1 since tm = (0, following the statutory schedule. As income rises, however, the value of Er increases, reflecting the fact that tm increases faster than la. These values reach a maximum near the middle range of income and decline thereafter. At the highest income level, ET returns to 1 because ta becomes 100 per cent and tm cannot exceed 100 per cent. An income of ¥10-20 million shows the greatest disparity between marginal and average rates under the progressive pattern of the statutory tax schedule.

116


in income taxes. From 1977 there were no major income tax cuts until the late 1980s, because the huge fiscal deficit prevented the government from reducing income tax liabilities (see Table 3.3 above). Once again in the 1990s, substantial amounts of cuts in income tax were enforced, a policy carried out to encourage the depressed state of the Japanese economy, not to adjust for inflationary bias. Thus, our major concern is with the effects of inflation adjustments for the inflationary period of 1960-90. Hence, we must question whether these annual tax cut policies have successfully corrected for the distorting effects of inflation on the individual income tax system and if so, in what way. In the past, some argued that the tax reductions proved insufficient to offset the impact of inflation on income tax liabilities.7 However, I doubt this. It seems to me that the reductions in taxes have more than offset the increases in taxes caused by inflation. It is necessary to investigate the offsetting effect of deliberate tax cuts enacted by periodic legislation. For this purpose, creating an indexing scheme with 'deflated taxes' would be of great value.8 According to such a scheme, taxable income is first deflated to a base year. Next, after the tax system is also fixed to a base year, it is assumed that tax revenues increase with the growth of real income each year. The result is deflated taxes keyed to a base year, which are thought of as tax revenues after offsetting the inflation-induced tax burdens. The concept of deflated taxes is of central importance. Deflated taxes Tdn in year n are defined according to the Sunley-Pechman formula (Sunley and Pechman 1976, 165):

where T*, Y* = taxes on taxable income, fi = the elasticity of tax liabilities to real taxable income, in a base year, respectively, and Y%= deflected taxable income in year n (« = 0,1, 2, ...), assuming automatic inflation adjustment. Necessary data can be obtained from two kinds of tax statistics collected by the National Tax Administration (NTA).9 In order to investigate the effect of inflationary adjustments, actual reductions in tax liabilities are compared with the tax liabilities that would have applied under an indexed system for two sub-periods (i.e. 1960-75, and 1975-90). We need to set out two base years, i.e. 1960 and 1975, when 7

For example, see Wada (1974, ch. 2), Nakagiri (1978); for another view, see Ihori (1984). This scheme is the third one proposed by Vito Tarm: see n. 3. Two basic sources for income tax data are obtained from the NTA data which have already been employed in Chapter 6: Statistics on the Self-Assessed Income Tax (Shinkoku Shotokuzei no Jittai), and Statistics on Private Wages and Salaries (Minkan Kyuyo no Jiltai). 8

9


117

both the tax system and the price level are fixed. Values of (3 are calculated by using cross-sectional data in 1960 and 1975 compiled on the basis of the two tax data sources. We have two values of (3, 1.477 and 1.819 in 1960, and 1.503 and 2.040 in 1975, for the self-assessed income tax and the income tax on wages and salaries, respectively.10 Tables 7.4 and 7.5 summarize the estimates of income tax liabilities and effective rates that would have applied to taxable income if the individual income tax had been indexed for inflation from 1960 (case 1) and 1975 (case 2). Columns (1) and (4) show the actual income and tax yields in current prices, and column (7) presents the percentage ratio of (4) to (1). In column (5) we have deflated taxes to the 1960 and 1975 price levels, calculated by the above formula. Furthermore, in column (-6) they are converted into current prices multiplied by the GNP deflator.11 The most interesting observations can be derived from the comparison between actual taxes and those assessing automatic inflation adjustment (i.e. deflated taxes) in current prices. Deflated taxes are considered as theoretical standards fully adjusted for inflation, since they are calculated according to the growth of real income and the income elasticity of tax yields, having fixed the price level and tax system to base years 1960 and 1975. In contrast, actual taxes have varied with the periodic changes in tax law imposed for the purpose of inflation adjustment.12 If the government succeeded in achieving its initial objective of inflation adjustment through deliberate tax reductions, actual taxes must be nearly equal to deflated taxes. Let us compare column (4) with column (6) in Tables 7.4 and 7.5. In case 1 of self-assessed income taxes (Table 7.4), actual tax liabilities rose from ¥95 billion in 1960 to ¥1413 billion, a factor of 15 times over the 16-year period. On the other 10 Values of (3 in 1960 were each estimated, according to the following regression equations. Selfassessed income tax:

logl 1 = 0.066+1.477 logy (20.191) R 2 = 0.926,

n=ll

Withheld income tax on wage and salary: logll = 0.049+1.819 logy (12.670) R 2 = 0.925, n=14 where 11 = per capita income tax liability, y = per capita income, R 2 = coefficient of determination adjusted for degree of freedom, n = number of observations, and figures in parentheses are t-values. The results for 1975 were calculated in a similar manner. 11 The consumer price index or the cost-of-living index might be a better measure than the GNP deflator because the tax burden of an individual is closely related to his ability to purchase consumption goods. However, because there are no data that cover the entire period using the same base year (i.e. I960), I am obliged to use the GNP deflator. There is little difference between the two indices when comparing them for overlapping years (i.e. 1967-73). 12 Tax reductions were not undertaken only for the purpose of inflation adjustment. Nevertheless, the main object of the periodic tax changes was to adjust for price hikes. Thus, it is reasonable to treat all tax reductions as if they were intended for inflation adjustment.

Table 7.4 Inflation and the self-assessed income tax liability, 1960-1990 Income (¥bn.) Current price

Base-year price

Deflator (Base year = 100)

Tax liabilities (¥bn.)

(2)

Case 1: under the 1960 tax system 1282 1282 1960 1375 1484 1961 1641 1835 1962 2252 1930 1963 2644 2167 1964 2818 2198 1965 3257 2423 1966 3943 2800 1967 3210 4718 1968 6386 4158 1969 4902 8044 1970 5912 10127 1971 11658 6487 1972 17258 8616 1973 5380 13010 1974 14339 5521 1975

(3)

(4)

100.0 107.9 111.8 116.7 122.0 128.2 134.4 140.8 147.0 153.6 164.1 171.3 179.7 200.0 241.8 259.7

95 123 144 189 216 230 265 336 420 540 664 812 I 120 1819 1165 1413

Actual

Gap: (8)-(7)

Current price (6)

Assuming automatic inflation adjustment

(7)

(8)

(9)

95 113 150 194 234 251 296 368 450 630 807 1033 1 197 1801 1318 1454

7.41 8.29 7.85 8.39 8.17 8.16 8.14 8.52 8.90 8.46 8.25 8.02 9.61 10.54 8.95 9.85

7.41 7.64 8.17 8.60 8.86 8.92 9.08 9.32 9.53 9.86 10.04 10.20 10.27 10.43 10.13 10.14

0 -0.65 0.32 0.21 0.69 0.76 0.94 0.80 0.63 1.40 1.79 2.18 0.66 -0.11 1.18 0.29


Actual current price

Base-year price (1)

Effective tax rates (%)

(5)

95 105 134 166 192 196 220 261 306 410 492 603 666 899 545 560

Case 2: under the 1975 tax 1975 14339 1976 14387 1977 16107 1978 17910 1979 20625 1980 22652 23904 1981 1982 25816 1983 27535 1984 28712 30440 1985 1986 33302 37750 1987 40887 1988 48993 1989 57550 1990

system 14339 13534 14333 15229 17084 18234 18734 19889 21 100 21065 21994 23635 25803 27852 32749 37787

100.0 106.3 112.4 117.6 120.7 124.2 127.6 129.8 130.5 136.3 138.4 140.9 146.3 146.8 149.6 152.3

1413 1307 1497 1733 2192 2364 2409 2549 2678 2741 2881 3367 4187 4324 5363 6602

1414 1293 1412 1544 1819 1989 2066 2237 2417 2411 2549 2789 3111 3415 4143 4890

1413 1375 1586 1816 2196 2472 2636 2904 3154 3286 3528 3930 4551 5013 6197 7447

9.85 9.08 9.29 9.67 10.62 10.44 10.08 9.87 9.73 9.55 9.46 10.10 11.08 10.58 10.95 11.47

9.85 9.56 9.85 10.14 10.65 10.91 11.03 11.25 11.45 11.46 11.59 11.80 12.06 12.26 12.65 12.90

Note: Income, before exemptions and deductions are applied, is used for this calculation. Effective tax rates are expressed in percentages after dividing taxes by income; i.e. (7) = (4) ^ (1) X 100, (8) = (6H (1)X 100 or (5)^ (2)X100, (6) = (5)X (3)-^ 100. Base year is 1960 in case 1 or 1975 in case 2. Source: Calculated from National Tax Administration (NTA), Statistics on the Self-assessed Income Tax (Shinkoku Shotokuzei no jittai); Economic Planning Agency, Annual Report of National Income Statistics.

0.00 0.48 0.56 0.47 0.03 0.47 0.95 1.38 1.72 1.91 2.13 1.70 0.98 1.68 1.70 1.43

Table 7.5 Inflation and withheld income tax on wages and salaries, 1960-1990 Income (¥bn.) Current price

Base-year price

Deflator Base year = 100

Tax liabilities (¥bn.)

Effective tax rates (%)

Actual current price

Actual


Gap: (8)-(7)

(7)

(8)

(9)

Assuming automatic inflation adjustment Base-year price

(1)

(2)

Case 1: under the 1960 tax system 3516 3516 4094 4417 5362 4796 6425 5506 7523 6166 8704 6789 10625 7459 12264 8710 14604 9935 17865 11631 13887 22788 27992 16341 32858 18285 20584 41229 22512 54435 61559 23704

1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975

(3)

100.0 107.9 111.8 116.7 122.0 128.2 134.4 140.8 147.0 153.6 164.1 171.3 179.7 200.3 241.8 259.7

(4)

174 215 252

332 392 420 441 510 613 794 1020 1 201 1 536 2184 2310 2240

(5)

174 226 289 353 413 469 529 642 752 905 1108 1329 1504

1711 1885 1992

Current price (6) 174 244 323 412 504 601 711 904 1 105 1390 1 818 2277 2703 3427 4558 5173

4.95 4.86 4.70 5.17 5.21 4.83 4.40 4.16 4.20 4.44 4.48 4.29 4.67 5.29 4.24 3.64

4.95 5.52 6.02 6.41 6.70 6.91 7.09 7.37 7.57 7.78 7.98 8.13 8.23 8.31 8.37 8.40

0 0.66 1.32 1.24 1.49 2.08 2.69 3.21 3.37 3.34 3.50 3.84 3.56 3.02 4.13 4.76

Case 2: under the 1975 tax system 1975 61559 61559 1976 71125 66909 1977 76547 68117 83555 1978 71050 1979 90777 75193 1980 98359 79178 104023 81523 1981 1982 108693 83739 1983 114980 88107 120064 1984 88088 129907 98863 1985 135202 1986 95956 1987 140 068 95736 1988 145875 99370 103473 1989 154796 1990 167128 109736

100.0 106.3 112.4 117.6 120.7 124.2 127.6 129.8 130.5 136.3 138.4 140.9 146.3 146.8 149.6 152.3

2240 2961 3139 3747 4493 5250 5780 6245 6630 6838 7661 8158 8107 8196 8442 9735

2240 2637 2726 2944 3252 3547 3716 3880 4204 4204 4634 4789 4749 5015 5318 5779

Note: See Table 7.4. Source: Calculated from NTA, Statistics on Private Wages and Salaries (Minkan Kyuyo no Jittai).

2240 2803 3064 3462 3962 4407 4742 5036 5486 5730 6413 6748 6948 7362 7956 8801

3.64 4.16 4.10 4.48 4.95 5.34 5.55 5.75 5.77 5.70 5.90 6.03 5.78 5.62 5.45 5.82

3.64 3.94 4.00 4.14 4.33 4.48 4.56 4.63 4.77 4.77 4.94 4.99 4.96 5.05 5.14 5.27

0.00 -0.22 -0.10 -0.34 -0.62 -0.86 -0.99 -1.12 -1.00 -0.93 -0.96 -1.04 -0.82 -0.57 -0.31 -0.55

122


hand, deflated tax liabilities began at the same level in I960, and would have reached ¥1 454 billion in 1975, a level ¥41 billion higher than the actual liabilities for that year. Actual taxes showed smaller amounts every year except in 1961 and 1973. This implies that most of the past tax reductions have more than offset the impact of inflation on income tax liabilities. By and large, the same holds for case 2 using the base year of 1975. Next, we shall move to the case of withheld income tax on wages and salaries (Table 7.5). The gaps between actual and deflated revenues in current prices are even more conspicuous than in case 1 of Table 7.4. Actual taxes are smaller than deflated taxes in all years, and they are short by approximately ¥3 000 billion in 1975. Thus, it is conjectured that actual income tax liabilities on wages and salaries were much smaller than they would have been if tax laws had been left unchanged and no inflation occurred. By contrast, in case 2 the opposite phenomenon occurred for the withheld income tax on wages and salaries from 1975 to 1990. Actual taxes exceeded the level of deflated taxes in current prices, hence the amounts necessary to adjust the income tax for inflation were insufficient. Obviously, this is due to the lack of periodic tax cuts and adjustments for inflation since 1977. Tax revenues have automatically risen under a fixed tax system with a growing money income. Let us turn to the changes in effective tax rates in columns (7) and (8). Figure 7.1 illustrates the variations in effective rates for the two types of income taxes under the 1960 and 1975 tax systems, in accordance with Tables 7.4 and 7.5. It is apparent that the dotted lines, which indicate the effective tax rates assuming automatic inflation adjustment, indicate a steady upward increase. This reflects the fact that tax revenues tend to grow in real terms even though the tax system is completely indexed for inflation because of the interaction between the progressive rate structure and growing real incomes. Evident from values of (3 greater than 1, the growth of the tax revenues has been faster than that of real income. In Figure 7.1 (a), under the 1960 tax system, actual effective tax rates vary considerably in a pattern of ups and downs, reflecting the periodic changes in tax law. They tend to decline in the long run in the case of the withheld income tax on wages and salaries, while in the case of the self-assessed tax a slight increasing pattern is indicated, although it is not nearly so apparent. The deviation between the two lines is conspicuously large for the income tax on wages and salaries.13 Again, this strengthens our contention that tax reductions had a very strong offsetting effect on the tax burdens of wage and salary recipients. It appears that the degree of the offsetting effect has a close bearing on the rate of inflation. In the case of the self-assessed income tax, no effective adjustment was made for inflation in 1961 and 1973 under the sharp rise of price levels. Indeed, inflation rates were 7.9 per cent in 1961 and 11.6 per cent in 1973 in terms of 13 The chief reason for the greater reduction in the income tax on wages and salaries is that the deduction for employment income has frequently been raised on a large scale. In fact, the elevation of the deduction for employment income in 1974 reached an all-time high.


123

Fig. 7.1 Changes in effective tax rates, actual and after inflation adjustment, (a) 1960-1975, under the 1960 tax system; (b) 1975-1990, under the 1975 tax system (i) Self-assessed income tax; (ii) income tax withheld on wages and salaries Source: Tables 7.4 and 7.5.

the GNP deflator, which are enormously high, excluding the unusual year of 1974 just after the occurrence of the oil crisis. In both years, actual effective rates were higher than the effective tax rates after inflation adjustment, and therefore ad hoc remedies by annual tax reductions failed to correct the increased tax burden caused by inflation. The higher the rate of inflation, the more difficult it is to correct the inflation-induced tax burden in the income tax system.

124


How, then, can we explain the case of 1974, when prices rose by over 20 per cent and yet the reduction of the tax burden more than offset the tendency of inflation to push income-earners into higher-rate brackets? We must recall that the largest amount of income tax reductions (i.e. ¥1 783 billion) in the postwar period were enacted in 1974 (see Table 3.3). As far as my estimates are concerned, this tax cut measure not only perfectly eliminated all the distortions caused by the abnormal level of inflation, but in fact offset them. Figure 7.1 (b) depicts the effective tax rates under the 1975 tax system in a similar fashion. As regards the case of the self-assessed income tax, the major findings are the same as those in the previous sub-period. However, in the case of withheld income tax, the actual effective rates outstrip the effective rates assuming automatic inflation adjustment, reflecting the end to annual tax reduction policies. Tax distortions caused by inflation were left uncorrected in the income tax system, and this resulted in inflation-induced tax burdens. The reason why the self-assessed income tax remained the same during the two sub-periods is not clear, but tax avoidance and evasion would probably explain to a considerable extent the fact that tax liabilities in the actual effective rates were lower than those yielded after inflation adjustments.

CERTAIN

ASPECTS

OF

POLICY

ISSUES

Fiscal Drag and Dividend As we have seen, income tax burdens are expected to increase automatically as a result of a progressive rate structure, coupled with an increasing growth in nominal income. Inflation tends to accelerate the tempo of such growth. This function of the individual income tax system is often regarded as a stabilizing effect which compensates counter-cyclically for the fluctuations in national income over changing business conditions. This is referred to as the built-in flexibility of taxation. However, this flexibility does not always have a beneficial effect on the economy. In the long-run process of economic growth, it tends to depress the level of aggregate demand especially during recoveries, and to slow down the potential path of economic growth. Thus, it is generally acknowledged that the depressing effect of progressive income taxes is not desirable from the standpoint of long-term policy objectives. Even if the distortion of inflation on income taxes can fully be offset, tax revenues grow faster than real incomes, and thus effective tax rates increase constantly. Let us return to Figure 7.1. The upward movement of effective tax rates, assuming automatic inflation adjustment (drawn with broken lines), indicates a clear-cut observation of this process. In the USA the phenomenon has been called the 'fiscal drag', drawing attention to the fact that it is one of the shortcomings in the mechanism of


125

built-in tax stabilizers in a growing economy.14 There are two policy measures that can cure the fiscal drag: (1) tax reductions and (2) increases in government expenditures. These remedies together are frequently called the 'fiscal dividend', in contrast to the concept of fiscal drag. As stressed repeatedly, the most conspicuous characteristic of postwar tax policy in Japan is the successive rounds of annual tax reductions which occurred until the late 1970s, focusing primarily on the individual income tax. This policy seems closely related to the fiscal dividend. The Japanese government selected tax reductions, rather than increases in government expenditures, as the means of effecting the fiscal dividend, mainly because it considered the tendency towards increasing tax burdens undesirable. As a result, this policy option has prevented the level of government expenditures from expanding rapidly in the past years, and has contributed to the construction of a comparatively small government. Now let us examine whether reductions in income taxes were able to function efficiently as the fiscal dividend. It has already been argued that tax reductions in the past have more than offset the increase in tax liabilities arising from inflation during most of the postwar period. This implies that some portion of the tax reductions was appropriated to reduce tax burdens in real terms which more than offset the tax burden caused by inflation. This portion must have some bearing upon the concept of fiscal dividend. It is difficult, however, to estimate what proportion of past tax cuts were meant to eliminate the fiscal drag. In crude terms, I shall attempt to estimate the size of the fiscal dividend. The first step is to calculate tax revenues that are assumed to cause fiscal drag. They must be calculated on the basis that taxes increase after adjusting for inflation because of the interaction between the progressive rate structure and growing real incomes. Until about 1970, the Japanese economy traced a growth path close to its potential output (see Patrick and Rosovsky 1976; Ackley 1976). Therefore, it is assumed that the growth of real income required for calculating hypothetical tax revenues has also been close to potential output. Given such an assumption, our hypothetical taxes can be regarded as those that induce the fiscal drag according to the formulation used in the USA. In this fashion, it is possible to consider certain aspects of the fiscal dividend in the first sub-period. The tax revenues in question have been calculated in columns (5) of Tables 7.4 and 7.5, i.e. as tax revenues assuming automatic inflation adjustment in terms of 1960 prices. In these calculations, it is assumed that the growth rate of real income is

14 The definition of fiscal drag was first developed by the Council of Economic Advisers. They stale: 'As the economy moves along the potential oulput path with reasonably stable prices, the federal tax system generates an increase in revenues of about 6 per cent a year. Unless this revenue growth is offset by reductions in taxes or by increases in expenditures, it acts as a "fiscal drag" by siphoning off income' (see Council of Economic Advisers 1962, 72-3). For reference, see also Heller (1965), Blinder and Solow (1974).

126


14.8 per cent, and that the elasticities of tax liability to real taxable income in two types of income tax are the same as in the preceding analysis.13 Let us further assume that the hypothetical tax revenue based on the 1960 tax law has been realized since 1960. The growth rates of income taxes have become rather high; the average rate of growth is 13.3 per cent for the self-assessed tax and 17.9 per cent for the income tax on wages and salaries. If the government had left tax increases unchanged, the phenomenon of fiscal drag would certainly have occurred in the Japanese economy, as it did in the USA. It was the deliberate tax cut measures that prevented tax increases from resulting in fiscal drag. In Table 7.6, the changes in both actual and hypothetical revenues are indicated in 1960 prices, making partial use of the previous results. The gap between the two can be defined as a sort of fiscal dividend, incurred by annual tax reductions. In this context, tax reductions are assumed to remove the fiscal drag that would occur and depress the economy if the hypothetical tax revenues continued to persist in the long run. Table 7.6 Fiscal year

1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975

Income taxes and fiscal dividend, 1960-1975 (¥bn.) Self-assessed tax (1960 price)

Income tax on wages and salaries (1960 price)

Tax adjusted for inflation (1)

Actual tax (2)

Gap (D-(2) (3)

Tax adjusted for inflation (4)

Actual tax (5)

Gap (4)-(5) (6)

95 105 134 166 192 196 220 261 306 410 492 603 666 899 545 560

95 114 129 162 193 179 197 239 286 352 405 474 623 908 482 544

0 -9 5 4 -1 17 23 22 20 58 87 129 43 -9 63 16

174 226 289 353 413 469 529 642 752 905 1 108 1329 1 504 1 711 1 885 1 992

174 199 225 284 321 328 328 362 417 517 622 701 855 1090 955 863

0 27 64 69 92 141 201 280 335 388 486 628 649 621 930 1 129

Note: Columns (1) and (4) are the same as (5) in Tables 7.4 and 7.5, respectively. (2) and (5) indicate (4) in those tables after converting from current prices to 1960 prices. 15 The CEA assumes a reasonably stable price, say 2-3 per cent, but in my calculation the price level is fixed at the 1960 level.


127

The amount of fiscal dividend is clearly different for the two types of income taxes. The income tax on wages and salaries produces a steadily larger amount of fiscal dividend than the self-assessed income tax. In particular, it began to increase sharply after 1974 and accounted for over ¥1 000 billion in 1975. The scale of tax reductions seems to be too large around these periods compared with previous trends. In contrast, annual tax reductions do not always result in the fiscal dividend for the self-assessed income tax. As is seen in column (3) of Table 7.6, the gap shows negative values in 1961, 1964, and 1973, when tax cut measures did not perfectly offset automatic increases in tax revenues caused by inflation. In real terms, tax increases were incurred. In other years, the self-assessed income taxes incurred a smaller amount of fiscal dividend, and therefore the annual tax reductions played a minor role in adjusting for inflation. Were Deliberate Measures Successful?

Inflation adjustment for the income tax system poses an important problem, particularly in Japan. The Japanese economy has grown with great speed, and this expansion was accompanied by inflation and a sharp rise in nominal income in the 1960s and 1970s. It is easy to conjecture that inflation caused significant increases in income tax liability with distorting effects on the rate structure and the tax base of income taxes. The government selected one remedy from two possible alternatives to cure the increase in income tax burden caused by inflation. Its policy was to administer annual tax reductions through periodic amendments of income tax law. In contrast with tax indexation, deliberate tax cut measures adjusted for inflationary distortion in crude terms. As my statistical evidence reveals, such measures are far from satisfactory in eliminating the effects of inflation on income tax liability. Since ad hoc remedies were not linked precisely with the variations in both rate structure and tax exemption, only rough corrections could be made. Although tax indexation would be a better scheme to remove inflationary distortions, the Japanese government has so far resisted the adoption of such a system, and it will surely be reluctant to accept it in the future. It seems that Japan's experience has not been as successful as that of other foreign countries that adopted tax indexing systems. For instance, discretionary changes in income tax laws enacted by the Diet have more than offset the automatic increases in tax revenues and have induced tax reductions in real terms. Also, lacking a tax indexing system, it would be impossible to adopt more detailed corrections matched to each type of taxable income, e.g. interest, capital gains, earned income, etc. The inability to undertake such corrections is due to the fact that the whole tax system is adjusted for inflation in crude terms (see Kay and King 1986, ch. 13). After the late 1970s, the Japanese government stopped adjusting the individual income tax for inflation with annual tax reductions. As a consequence, traditional

128


remedies for inflation adjustment were abandoned, and the increase of the income tax burden must have been induced substantially by inflation, despite the smaller increases in the price level. Looking back at this way of reducing income tax liability, major portions of the tax cuts were due to a rise in the minimum taxable level (i.e. tax threshold), rather than to a lowering of the progressive rate structure. In Figure 7.2 the trend of the tax threshold since 1970 is illustrated, in addition to the rise in consumer prices in terms of percentage increases in Figure 7.3. It is apparent that the level of the tax threshold remained almost unchanged between 1977 and 1986 with the exception of 1984. As a result, more income-earners were thrust above the minimal level of income subject to tax as their nominal incomes rose. Since the rate of increase of consumer prices exceeded that of the tax threshold during the decade, the real value of the minimal taxable level was lowered substantially in accordance with the rate of inflation. This level should have been raised to avoid increasing the number of taxpayers. In fact, the ratio of taxpayers among wage and salary workers to the total number of taxpayers increased steadily, from 86.1 per cent in 1977 to 91.7 per cent in 1982 and 89.1 per cent in 1985. During the same period, progressive tax rates had no major changes either, although a small reduction was undertaken in 1982, to correct the distortion caused by inflation. Since measures of inflationary correction were not adopted to enable the government make the proper adjustments in the 1980s, what has happened to the income tax system? There are two points worth noting in this context. First, real tax burdens rise because of bracket creep as nominal income increases. Thus, the

Fig. 7.2 Tax thresholds for a working family of four, 1970-1992 Source: MOH Tax Bureau, Primary Statistics of Taxation (Zeisci Shuyo Sanko Shiryoshu), Feb. 1992.


129

Fig. 7.3 Tax thresholds and inflation, 1970-1992 The rate of inflation is calculated based on the rate of increase in the consumer price index. Source: See Fig. 7.2.

'fiscal drag' occurs, demanding a higher proportion of the taxpayers' rising money income as tax each year. Second, the government automatically receives increasing tax revenues as a result of inflation. The impact of inflation on the tax system is very clear: individual income tax revenues rise relatively rapidly, while revenues from indirect taxes fall throughout the whole tax system. This represents a switch from indirect to direct taxation, as was seen in Table 1.3, and it should be regarded as an unintended by-product of inflation in the absence of proper adjustment. However, this trend was reversed in the 1990s with income tax reductions in the context of fiscal stimuli. In recent years, action has been taken to prevent further unplanned shifts and to restore the relative proportion of indirect taxes to total tax revenue. These efforts have proven successful to some extent in accordance with introducing the consumption tax (Japan's VAT) in the tax system. Minimum tax thresholds have been raised considerably since 1992. Indeed, the tax threshold for a working family of four was ¥3 277m. in 1993-94, ¥3 539m. in 1995-97, and ¥3 616m. since 1998. Such an upward trend has been induced by income tax reductions as fiscal stimuli, rather than inflationary adjustments.

8 The Taxation of Investment Income and Savings

An important issue relating to the Japanese tax system is the impact of income taxes on saving behaviour. Japan's high savings rate during the postwar period has received world-wide attention in connection with the country's notable economic performance. There has been a general belief that the high rate of personal savings has been stimulated by tax policy adopted by the government. In fact, the most salient feature of these special tax measures is the substantial portion of revenue losses that have been sustained in order to stimulate personal savings (see Table 3.2). Moreover, the country's relatively light tax burden may have helped to increase the savings rate. Despite great interest in such issues, past studies have reached no conclusion on the influence of Japan's tax policy in promoting saving behaviour. Since the subject is a complex one, I do not hope to achieve definitive results. We shall begin with a brief explanation of the way in which the existing income tax system tends to favour income from investment and savings, with special reference to special tax measures prior to fiscal year 1988. Next, some policy issues are discussed relating to Japan's high personal savings rate, mainly related to the late 1980s. Lastly, an attempt is made to obtain some empirical relationships between personal savings and income taxation.

MAIN FEATURES OF TAX P R E F E R E N C E S

Background It is often pointed out that special tax measures were initiated to implement specific policy goals—in particular, capital accumulation, i.e. the promotion of savings and equity investments, during the process of rapid economic growth. Capital accumulation was promoted in two ways: 1. by the full or partial exclusion of certain investment income from the tax base; and 2. by the separate application of special reduced tax rates to certain investment income.

This chapter is partially based on Ishi (1983a).

Investment Income and Savings

131

These measures were designed to give preferential treatment to interest, dividends, capital gains, and other such incomes in order to stimulate personal savings. Although some special provisions were phased out, tax preferences for savings were maintained intact until fiscal year 1987. Needless to say, such special measures applicable to specific sources of income are a clear deviation from a global tax system proposed by the Shoup Mission. There were several forms of savings that received exemption or highly favourable tax treatment. Thus, a major effect of the Japanese tax system has been its impact on saving behaviour, although the nature of its impact is difficult to pin down. The Privileged Savings System until Fiscal Year 1987 Our first concern should be with the tax preferences for specific personal savings.1 According to this system, until March 1987, income earned from interest, up to a specific limit of total principal (or total face value) of personal savings (or public bonds), was exempt from income taxation. These privileged savings consisted of five sources, including the well-known 'Maruyu' system (see note to Table 8.1). In April 1988, these special treatments were replaced by a flat rate of 20 per cent at source. These tax preferences developed gradually over a long time as a variety of government activities were designed to promote savings. For instance, the origin of tax-free postal savings dates back to 1920, and the limit on the principal was altered frequently, although the postal savings system was founded in 1875. Similarly, the idea of not taxing the interest earned on small savings was instituted in 1920 to encourage national saving for the war effort. In contrast, a provision for savings enabling

Table 8.1 Personal saving with non-taxable interest income, 1987 Sources ( 1 ) 'Maruyu' system (2) Postal savings (3) National and local bonds (4) Savings for the formation of employees' assets (5) Postal instalment savings for housing Total

Limit (¥m.) 3 3 3 5 0.5

14.5

Note: the 'Maruyu' system includes deposits at banks, securities companies and other private financial institutions. Savings under this system can exceed ¥3m., the interest of which is subject to taxation. In contrast, the amount of postal saving cannot exceed the threshold by law. Only salaried workers arc eligible for savings for the formation of employees' assets.

1 For an expanded discussion, see e.g. MOF Tax Bureau (1986), and Gomi (1985); see also Suzuki (1987) for the general discussion of financial assets in Japan.

132


the formation of employees' assets was enacted in 1972, following the West German attempt to stimulate employees to creat their own assets. Accordingly, the individual income tax law in Japan permitted each individual to earn tax-exempt interest on personal savings of up to ¥14.5 million (US$96666, where $1 =¥150). In spite of the strict requirements necessary to take advantage of these measures, the amount of savings that was tax-favoured was abnormally high by international standards. Table 8.2 shows the significant role of personal savings under the privileged savings system in accumulating total savings. In March 1986, the total principal outstanding of personal savings, including those with both taxable and non-taxable interest, amounted to ¥503 000 billion. The largest components were deposits and savings held by banks, other financial institutions, and post offices, followed by insurance. Savings, the interest on which was exempted from taxation by the income tax law, made up approximately 60 per cent of total personal savings. Most notable is the large share of the 'Maruyu' and postal savings system, which together account for more than 90 per cent of all tax-free savings.

Table 8.2

Total principal outstanding of personal saving, by source, as of March 1986

Sources

Outstanding (¥000bn.)

Total principal outstanding Deposits and savings Private financing institutions Postal offices Trusts Insurance Security investment Bonds and debentures Investment trusts Total

318 215 103 26 98 61 45 16 503

Principal outstanding of small savings with non-taxable interest 163 Maruyu 153 Private financial institutions 10 Security companies 103 Postal savings 11 National and local bonds Savings for the formation of employees' assets 10 Total Note: Postal Instalment Savings for housing is omitted. Source: Data from the MOF.

287

(%)

(63.2)

(5.2) (19.5) (12.1)

(100.0) (56.7)

(35.9) (3.8) (3.5) (100.0)


133

Reduced Tax Rates on Interest In addition to the full or partial tax exemption of earned interest under the 'Maruyu' system and the other measures mentioned above, the Japanese tax system developed another type of tax concession on interest. In the case of taxpayers who exceeded the limit on privileged savings, excess interest income was taxed either separately or aggregately in conjunction with other incomes, at the option of the taxpayer. If it was to be aggregated, interest earned on taxable savings was subject to a withholding tax of 20 per cent at source and then was combined with other incomes on the tax return. As an exceptional treatment, interest on ordinary deposits and other similar deposits was taxed up to a 20 per cent flat rate at source, but this income could be excluded as taxable income on the tax return. This meant that, in practice, some interest was taxed only at the preferential rate of 20 per cent. It was also possible to apply reduced tax rates under the separation method to interest exceeding the threshold of tax-favoured savings. Taxpayers who obtained such interest could choose the option of having additional income taxed separately at lower rates. Moreover, they did not need to count such income as taxable income for tax purposes. The past record of reduced tax rates since 1951 is summarized in Table 8.3. Following relatively higher tax rates of 50 per cent in 1951—52, tax rates on interest were lowered to 10 per cent in 1953 when the Shoup proposals were substantially Table 8.3 Special tax measures for interest income: reduced tax rates in the case of separate taxation at source Reduced tax rates (%)

Period

50 10 5 10 0 0 10 10 5 10 15 20 25 30 35

1951-2 1953 (long-term saving) (short-term saving) (short-term saving) (long-term saving) (short-term saving) (long-term saving) (both) (both) (both) (both) (both) (both) (both)

1954

1955-6 1955-8 1957-62 1959-62 1963-4 1965-6 1967-70 1971-2 1973-5 1976-7 1978-87

Source: MOF, Primary Statistics of Taxation (Zeisei Sanko Shiryoshu), February 1988.

134


modified for the first time. Furthermore, no tax was levied on interest income earned on both short- and long-term savings in 1955 and 1956. Since the 1970s, these special rates have begun to be raised in response to criticisms of inequitable income taxation. In addition, it should be noted that gains from discount bonds were subject, at the time of issuance, to a withholding tax rate of 16 per cent at source. Special Treatment of Dividends Dividends as well as interest income are treated under special tax measures. Although they have no counterpart to the 'Maruyu' system, two types of separate reduced tax rates, 20 and 35 per cent, are applied to dividends in a similar fashion. For one thing, dividends are taxed at source at the rate of 20 per cent, but are included in taxable income for filing a return. The other is that they may be levied at the reduced rate of 35 per cent, separate from other incomes, and removed from taxable income when filing a tax return. This is an optional system for taxpayers. Furthermore, recipients of dividends are now allowed to credit 10 per cent of their dividends against their income tax as compensation for the double taxation of dividends. The tax credit is reduced, however, to 5 per cent if the individual's total income is more than ¥10 million. A tax credit of 5 per cent is thus applicable to such fraction of dividends as is equal to the amount of total income minus ¥10 million, and the rest is credited against total income at 10 per cent. Capital Gains Of great interest is the favourable tax treatment given to capital gains on the sale of stocks, compared with taxes on other investment incomes. Non-taxable treatment of capital gains on the sale of stocks had become a symbol of unfairness of the existing income tax. Although capital gains had once been fully taxed (losses fully exempted) in accordance with the Shoup tax proposals, they were exempted from taxation in 1953 because the administration of the tax was too difficult to enforce. Since then, capital gains had in principle been tax-exempt, but those who continuously dealt with stocks in large volumes (thirty transactions a year, involving more than 120000 shares in 1988) had been required to include capital gains in their tax base, which was then subject to aggregate taxation. In response to criticism of the non-taxable status of capital gains, the Takeshita tax reform in 1988 proposed two alternative capital gains taxes, irrespective of the introduction of the Tax Identification Number (TIN). The first was to apply realized capital gains to the self-assessed declaration method, separate from other incomes, at a rate of 20 per cent (plus 6 per cent for the local inhabitants' tax). The second was to impose a 20 per cent tax on capital gains at the sources. Capital gains were deemed as 5 per cent of gross proceeds derived from the stock sales price. Consequently, the taxpayer would be required to pay 1 per cent (0.2 X 0.05) of the


135

stock sales price as a deemed capital gains tax under the withholding system. This is exactly the same as the existing securities transaction tax. The taxpayer would have the choice of self-assessed or withholding methods. Taxation of capital gains from the sale of stocks became effective from April 1989, but additional requirements were introduced. When capital gains are obtained from the initial public offering (IPO) of stocks in the market, taxpayers cannot opt for separate taxation at source. Such gains must be filed in a return if they are acquired before the IPO and sold within one year, being subject to national and local income taxes in the following formula: Short-term gains (stocks held for 3 years or less prior to the IPO)—the entire capital gains are included in taxable income. Long-term gains (stocks held for more than 3 years)—only half of such gains are included in taxable income. In addition, the taxation of capital gains from the sales of stocks was to be reviewed in the future, the aim being to move towards a unified tax from the current method of separate taxation, on condition that the system of using the Tax Identification Number could be introduced into the tax system. In relation to this shift, the possibility of treating interest income taxation in a similar way was also to be reconsidered. Moreover, capital gains on land, buildings, or the right to use land are taxed separately at low rates, depending on the length of ownership. This category of capital gains taxation has been strengthened substantially in the past decade, as mentioned in Chapter 6. Other Favourable Treatments In addition to the special tax measures applied to interest, dividends, and capital gains, there are certain types of smaller tax concessions allowed on other privileged savings. Particular attention should be paid to the following three types. Insurance is the first kind of savings that attracted tax preferences in this context. Premiums for life insurance and fire and other casualty insurance are non-taxable, with special deductions permitted up to a limit. Second, housing investments as a form of personal savings are given some favourable tax treatment. For practical administrative considerations, imputed rent on owner-occupied housing is not taxed as is normally done in other overseas countries. The interest paid on loans for house purchase and construction is eligible for a tax credit, up to an amount equal to 1 per cent of the total balance of the loan payment. Of course, there are strict requirements for using this credit in terms of the extent of floor space and the level of the borrower's income. Furthermore, the Postal Instalment Savings for housing mentioned above was created as a means for saving for the deposit to be used towards the purchase of a house. Lastly, capital gains obtained on the sale of the taxpayer's residence are treated very generously.

136


The third type of privileged savings is public or private pension funds.2 Pension contributions made by the employer are excluded from taxable income. An employee's pension contributions are also eligible for a special deduction on social insurance premiums in the individual income tax law. On the other hand, payments made out of a pension fund are taxed fully as employment income, but an exemption for employment income and a special exemption for pension contributions for the aged have been adopted to calculate taxable income. This treatment is considered very lenient. In addition, taxes are generally not levied on the accumulated pension funds themselves, with the exception of some private pension plans. Considering the favourable tax treatment on both contributions and payments, savings made through pension funds are regarded as a means of tax-free accumulation of personal savings.

PERSONAL SAVINGS IN JAPAN: SOME POLICY

ISSUES

Significance of the High Savings Rate It is widely acknowledged that Japan's savings rate has been one of the highest among advanced countries. Figure 8.1 illustrates the movement of six major countries including Japan during the period 1965-90. Clearly, Japan maintained the highest level over two decades,3 followed by Germany and France. Indeed, the personal savings rate in Japan reached an historic high of 23.2 per cent in 1974, and then declined steadily to as little as about 16 per cent in 1985. In 1990, Japan still occupied the top rank while the USA and the UK showed personal savings rates of only 6-7 per cent. The lowest savings rate was in Sweden, where it became negative after 1985. It is well known that the high personal savings rate in Japan played a vital role in the achievement of a remarkably high-growth economy between the 1950s and the first oil shock in 1973. It contributed greatly to the financing of private investments in plants and equipment and to the rapid expansion of the productive capacity of the economy. Moreover, since the late 1970s high personal savings have enabled the country to finance the huge fiscal deficit without causing a crowding-out effect in financial markets. This situation should be compared with that in the USA and other countries with lower rates of personal savings. Traditionally, however, Japan's high savings rates have been greatly criticized overseas. The main criticisms are directed towards the fact that the excess of domestic savings over domestic investment induces a massive volume of trade and current

2 For a detailed discussion of the tax treatment of the pension system, see the Tax Advisory Commission (1986), Noguchi (1983). 3 Italy sometimes showed a slightly higher rate of personal saving than Japan.


Fig. 8.1

137

Movements of personal savings rates in six major countries, 1965-1995

The personal savings rate is defined as the ratio of net saving (= total current receipts - total current disbursements) to final consumption expenditure plus net saving in the account for households and private unincorporated enterprises. Source: Calculated from OECD, National Accounts, vol. II, 1983-95,1973-85 and 1962-79.

account surpluses in Japan, strengthening the country's international competitiveness in the world market. In turn, capital has flowed outward from Japan to the USA and other countries, and hence Japan's high rate of personal savings is now regarded as a primary cause of trade friction and economic disorder in the world. In addition to the macroeconomic criticisms of Japan's high savings rate, many questioned the value of maintaining favourable tax treatment for personal savings. In the late 1980s, this view attained particular prominence with the movement to eliminate inequities from income taxation. As a part of the movement towards tax reforms, there was a growing feeling that the privileged saving system should be revised. Originally this system was established to aid small savers, but it has been more widely used by those with higher incomes. This is evident from Table 8.4: the higher a family's income level, the greater the proportion that it is benefiting from the tax preferences on tax-exempt savings. The percentage of total non-taxable savings used by taxpayers tends to increase with income. Furthermore, it is generally believed that higher-income persons took advantage of this system by opening numerous small savings accounts and using fictitious or false names to evade taxation. The limits on the privileged savings system were applied to individuals, and so families as a whole were eligible for much higher limits.4 Thus, abuses were believed to be common. To the extent that higher-income 4 In general, individuals can open tax-exempt accounts in the name of their spouses, children, and relatives when they have used up their personal tax-free savings. Gift taxes should be paid on the transferred income if the income exceeds their exemption level of ¥600 000 a year. In practice, however, gift taxes were not imposed on such transfers. Thus, interest income derived from virtually unlimited amounts of personal savings is non-taxable.


138 Table 8.4 1986(%)

Proportion of families using the privileged saving system, by income class, Maruyu

% of families using the system Annual income (¥m.) less than 2 2-3 3-4 4-5 5-7 over 7 % of families not using the system Unknown (%) Total (%)

Postal savings

78.9

56.4 71.9 76.9 83.0 88.2 93.0

(19.6) (18.9) (17.3) (19.8) (24.9) (43.1)

54.0

(13.1) (13.9) (11.5) (8.9) (12.0) (29.3)

37.8 49.9 54.0 59.3 59.3 61.8

National or local bonds

13.0

6.4 (9.1) 5.7 (20.5) 10.6 (21.7) 13.6 (16.5) 16.5 (17.4) 28.2 (30.7)

16.3

36.5

69.7

4.8

9.5

17.3

100.0

100.0

100.0

Note: Percentages in parentheses are the proportions of the total number of families that use the full amount of privileged savings. Source: Council for Savings Promotion (Chochiku Zokyo Chuo linkai).

families tend to benefit more from this system, it should be considered unfair or inequitable. Are Tax Benefits

Effective?

Next, we shall consider the effects of large tax benefits on personal savings. There is still controversy about the relation between the high rate of savings and the tax-exempt status afforded by many types of savings. Foreigners in particular often claim that the high savings rate in Japan has been created by these tax concessions. In Japan, opinion is divided among academics, businessmen, and politicians. Some believe that the tax concessions have substantially contributed to the high rate of Japanese savings, while others think they have contributed little or nothing. It seems to me that the special treatment of privileged savings has played no significant role in encouraging greater saving. It is almost impossible to estimate empirically the impact of tax incentives on high personal savings (see Komiya 1966a; Pechman and Kaizuka 1976). A great number of studies have sought possible explanations for Japan's high savings rate,3 5 See e.g. Mizoguchi (1970), Shinohara (1982), Kosai and Ogino (1984), Horioka (1986), Ishikawa (1988). In most of the literature, the relation between personal savings and taxation alone has not been studied.


139

but it seems that no attempt has so far been successful in determining how much of the high savings rate is attributable to tax policy and how much to other factors. The high rate of personal savings has also been explained by various other factors, such as rapid income growth, the bonus effect, the life cycle hypothesis, the socio-cultural background of savers, the under-developed social security system, etc. (For an excellent survey of the literature, see Horioka 1985.) Many, if not all, of these possible factors help explain Japan's high savings rate, and tax incentives should be regarded merely as one of a number of minor factors. A lot of tax experts, including some members of the Tax Advisory Commission, are sceptical about the effectiveness of special tax measures. They point out that the desirable effects, if any, of the special tax measures have been achieved at the expense of simplicity, neutrality, and equity in the individual income tax system. As already noted in Chapter 6, many of these measures impair the progressiveness of the nominal tax rates as they apply to specific individuals. While the intended effects of tax incentives are ambiguous, their distortion of the structure of individual income taxation is apparently clear-cut. Some reference should also be made to the interest elasticity of savings, because this plays an important role in determining whether the special tax measures can significantly influence the level of personal savings. If the interest elasticity of saving is substantially greater than zero, tax incentives for saving would have a substantial impact on the level of personal savings through the effects on the after-tax real rate of return. As is well known, the sign or magnitude of the interest elasticity of savings is theoretically indeterminable because of two offsetting effects: substitution and income effects. Thus, our concern is with the empirical results of using econometric models. In the USA recent evidence suggests that the interest rate elasticity of saving is positive and large in magnitude (see Boskin 1978; Summers 1981; Kotlikoff 1984; etc.). If the same evidence could be found in Japan, this would imply that various incentive measures for saving have had a significant positive impact on the level of personal savings, because they lowered the effective tax rate on interest and in turn increased the after-tax real rate of return. Unfortunately, I have obtained no convincing evidence in Japan which would suggest a significantly positive value for the estimated elasticities in question. Perhaps such poor results can be attributed to the past regulation of interest rates. Much of the affected savings were thus earned below market-regulated rates because of government intervention. A typical example was seen in the interest earned on postal savings. On this point, the regulation of interest rates can be regarded as a rationale for tax-exempt saving because interest income had already been 'taxed' through the regulatory limits on interest rates. However, these savings accounts have been deregulated, and their interest rates have been adjusted to market levels. Consequently, as time passes, we may see better results for the interest elasticity of saving under deregulation.

140

Individual Income Tax EMPIRICAL RESULTS OF PERSONAL SAVINGS AND TAXATION

The Effect on Portfolios Although scepticism is widespread about the effect of tax incentives on the level of personal savings, it appears that the special tax measures have some effect on saving behaviour. In fact, empirical evidence suggests that tax incentives may have greater influence on the portfolio decision of households (i.e. on the form in which households wish to hold their personal savings) than on the total level of savings. The first three rows of Table 8.5 show the proportion of typical forms of financial investments (e.g. (1) savings deposits, (2) bonds and equity, and (3) life insurance and pensions) to the net acquisition of financial assets in Japan, the USA and the UK for selected years. In comparison with other countries, Japan's portfolio holdings are characterized as follows. First, and most importantly, the proportion of personal holdings of savings deposits occupies more than 60 per cent of total financial assets. It accounted for 70.2 per cent in 1995, and the comparative figures were 2.6 per cent in the USA and 30.1 per cent in UK. This means that Japanese households have given priority to savings deposits as a vehicle for personal savings. Second, life insurance and pension Table 8.5

Proportions of portfolio investments in the household sector,'1 selected years (%) Japan

Proportions of financial investments (1) Saving deposits'1 (2) Bond and equityc (3) Life insurance and pension Financial investments (4) The ratio of netd lending to gross capital accumulation

USA

UK

1975

1985

1995

1975

1985

1992

1975

1985

1987

66.2 12.0 10.8

63.5 2.4 28.3

70.2 -10.3 39.5

49.8 5.5 32.0

23.8 14.6 45.1

2.6 42.8 48.7

44.1 -3.6 43.9

33.5 2.5 44.0

30.1 15.0 37.3

57.4

64.8

59.7

44.7

21.5

11.8C

56.3

36.7

38.5f

Note: Percentages in (1 )-(3) are calculated as the ratio of each portfolio investment to total acquisition of financial assets. Since other items are excluded, totals are not equal to 100%. " Includes private unincorporated enterprises. b Deposits other than transferable deposits. ' Short-term and long-term bonds and corporate equity security. Negative figures in Japan and the UK imply larger values of capital loss in equity. '' Net lending means financial investments after subtracting real investment (i.e. increases in stocks, gross fixed capital formation, purchases of land, etc.) from total capital accumulation. c 1994 figure. 1 1994 figure. Source: OECD, National Accounts, vol. II, 1973-85 and 1983-95.


141

funds are not so attractive to households in Japan, although the relative shares have accelerated rapidly. By contrast, households in the USA hold over 40 per cent of their financial assets in the form of life insurance and pension plans, and those in the UK hold even higher proportions. Third, portfolio investment, which takes the form of bond and equity securities, was present in a relatively lower proportion in Japan than in the USA and the UK in 1990. The next question concerns the proportions of total assets that have been allotted to financial and to real investments. It is difficult to determine this because of the lack of reliable data, but the ratio of net lending to gross capital accumulation might be of great help here. This ratio roughly indicates the relative share of financial investments in total capital accumulation.6 Accordingly, the higher the value of this ratio, the greater the number of households holding financial assets. In turn, these investments help to finance plant and equipment in the private sector and to make up for fiscal deficits in the government sector. The ratios of net lending to gross accumulation in Japan are much higher than those in the other two countries (see row (4) of Table 8.5). It is of great interest to observe the 1985 figures: Japan showed 64.8 per cent in 1985, as compared with 11.8 per cent in the USA and 36.7 per cent in the UK. Purchases of financial assets in Japan have been maintained at a high level. Given these data, it is often conjectured that households in Japan tended to exhaust the tax-exempt portions of the privileged savings accounts first (the 'Maruyu', postal savings, and national bonds, in that order), and then to direct their income towards the purchase of real assets (e.g. housing and land) and equity investments. In general, however, individuals have limited access to both of these. The Japanese tax system must have had its impact on saving behaviour, mainly reflecting a hybrid of income and expenditure taxes in favour of special treatments on investment income or savings. Three features of a hybrid income tax affect households' portfolio decisions. First, investment income arising from saving deposits, bonds, and equity has greatly benefited from highly favourable tax treatment. As stressed earlier, we have long enjoyed the privileged savings system and the tax exemptions for capital gains on the sale of stock. This has resulted in a relatively larger proportion of savings deposits, bonds, and equity in Japan. Second, special tax measures on the other forms of personal savings, such as insurance and pension funds, have not developed to the same extent as in other countries, say the UK. For instance, premiums for life insurance are given only a small deduction, and private pension funds themselves are specially taxed. Overall, income tax has not played a vital role in establishing forms of personal wealth other than the privileged savings system. 6 Net lending is the difference of real investments (i.e. the sum of increases in stock, gross fixed capital formation, net purchase of land and intangible assets, and capital transfers) and total gross accumulation. It indicates the money amounts appropriated for the purchase of financial assets.

142


Third, the interest paid on loans is usually not deductible under the income tax law, unlike in the USA where households are allowed to deduct 100 per cent of their interest payments. In addition, current laws do not allow taxpayers to deduct their local property tax liability from their national income taxes. Capital gains on the sale of land and buildings are taxed more heavily, although ad hoc tax relief was given during a specific period. It is permitted under the income tax laws to deduct the interest paid on owner-occupied mortgages, but the extent to which home owners are allowed to deduct this interest is much limited. Less favourable tax treatment for the purchase of real assets has resulted in a smaller share of total capital accumulation going towards real estate investment. Determinants of Personal Savings: An Estimated Result In spite of some recent results in explaining Japan's high savings rate,7 it is still very difficult to determine the direct impact of tax incentives on savings from empirical data. The subject is a complex one, and the present discussion is intended more to provide a sense of the argument than to develop it in detail. The light tax burden may be a significant factor in the explanation of the high rate of personal savings in Japan (see Pechman and Kaizuka 1976, 369). Indeed, it has often been argued that the high personal savings rate may be due to the low level of taxes and social security contributions. In particular, this argument seems to be widely accepted within the government.8 We can identify this relation in international comparison since the personal savings rate and the ratio of tax burden (including both direct and indirect taxes) plus social security contributions to personal income tend to show a negative correlation (see Ishi 1983a).9 Although some correlation between savings and taxation can be obtained from an international comparison, we need to investigate Japan's case in greater detail. What factors determine the variation of personal saving in Japan? In this regard, the aggregate savings function seems to be of use to obtain a rough sketch of major determinants. An attempt is made to apply Taylor's type of saving function to the Japanese case. As formulated by Taylor (1971), the reduced-form estimation equation is

where S = personal saving, S__i = personal saving in the preceding period, Yw= labour income, Tr= transfer income (i.e. social security benefits, social assistance

7 In recent studies, the relative importance of new explanatory factors (e.g. the low ratio of the aged to the working-age population, or higher land prices) is added to make the discussion more convincing (see e.g. Horioka 1986; Ishikawa 1988; Yoshikawa and Ohtake 1988). 8 This view was a 'Kasumigaseki' (the name of the district in Tokyo where the main governmental office buildings are located) theory. Tor example, an official statement in support of this theory was published in EPA (1982). 9 Personal income consists of (1) wages and salaries of employees, (2) the self-employed income, and (3) property income in the national account concept.


143

grants, and unfounded employee welfare benefits), Yp = property income (i.e. interest, dividends, and rent and entrepreneurial income), S/= social security contributions, Tp = personal tax (including non-tax revenues), r=the interest rate on one-year time deposits, and u = an error term. Although the derivation of (8.1) is complicated, the equation itself is rather simple. It requires only the regression of personal savings on its value in the preceding period and the first differences of the components of disposable income. In addition, the estimated model contains one additional variable, the interest rate on one-year time deposits. With the exception of the interest rate, data are obtained from 'Income and Outlay Accounts in Households' in the Annual Report on National Accounts published by the Economic Planning Agency. The quarterly data and nominal values are employed. Observations cover the period 1965 (III)-1991 (I), and the sample period is divided into three sub-periods mainly because tax policy changed substantially in the late 1970s and mid-1980s. (See Chapter 3.) Empirical results are tabulated in Table 8.6, in which four equations have been estimated for each period. Equation (1), based upon the disaggregation of disposable income and the interest rate as an additional variable, is the full model. Equation (2) is estimated with social security contributions and personal tax combined, since the coefficient of personal tax is not statistically significant. Equation (3) is added to (2) with labour and transfer income combined. Finally, equation (4) is intended as a benchmark for comparison, and differs from (2) in that disposable income is not disaggregated. Major findings are summarized as follows. 1. The most important explanatory factor is personal saving itself during the preceding period in all equations. 2. The coefficient of transfer income in the first period is not statistically significant. If it is combined with labour income in equation (3), the result becomes meaningful with a positive sign. 3. Property income has a significant relationship to the variation of personal savings, as was expected, with minor exceptions. 4. Personal contributions to social security have negative coefficients in all equations. Likewise, personal tax has the same result except during the first period. This suggests that in the short run the bulk of the adjustment to a change in social security contributions and personal tax falls on savings rather than consumption. Thus, it seems that households consider social security contributions to be a form of tax. In addition, tax incentives on savings seem to be obtained from our estimate since the late 1970s. 5. The interest rate has no significant impact on the variation of personal savings. This is quite natural, given the fixed interest structure and savers' behaviour during the past years. With regards to tax incentives, a number of reservations are required before reaching more conclusive results. For example, the aggregate savings function

Table 8.6 Equations estimating personal savings, 1965-1998 Equations

Personal saving

St-i

Changesa Labour income

Transfer income

Property income

AY"W

ATr

WP

Personal contributions to social security AS7

R2

DW

0.984

2.547

Personal Interest tax on 1 -year time deposits AT p Ar

(a) Quarterly 1965-7 6b 0.821 (1) (15.266)

1.283 (5.121)

1.668 (0.873)

0.813 (2.015)

(2)

0.857 (19.674)

1.166 (4.955)

0.799 (0.576)

1.025 (2.850)

-1.877 (-2.961)

0.983

2.608

(3)

0.853 (21.519)

1.024 (2.880)

-1.974 (-3.959)

0.984

2.646

-1.993 (-4.085)

0.984

2.670

0.998

2.887

0.998

2.817

(4)

1. 147 (5. 198)

0.856 (22.768)

1.106 (7.336)

(b) Quarterly 1977-86C 0.949 (1) (100.654)

0.561 (10.386)

1.457 (5.656)

0.682 (4.353)

0.949 (104.918)

0.552 (10.674)

1.620 (9.480)

0.682 (4.450)

(2)

-2.572 (-2.411)

-0.131 (-0.089)

-1.034 (-5.010)

-1.211 (-14.623) -1.198

( -14.857)

-400.755 (-1.151)

-97.299 (-0.541)

(3)

0.955 (82.044)

(4)

0.956 (75.661)

0.767 (19.942)

0.224 (1.395) 0.666 (56.451)

(c) Quarterly 1987-1998d 0.949 (1) (55.803)

0.560 (14.293)

1.516 (17.704)

0.715 (3.920)

(2)

0.952 (60.060)

0.557 (14.595)

1.521 (19.549)

0.720 (4.177)

(3)

0.934 (33.197)

(4)

0.942 (31.764)

0.806 (15.791)

-0.007 (-0.025) 0.699 (23.069)

-0.999 (-10.923)

0.997

2.648

-1.213 (-23.306)

0.996

2.596

0.987

3.051

-1.204 (-13.783)

0.988

3.047

-0.973 (-6.471)

0.961

3.124

-1.291 (-14.481)

0.956

3.072

-1.182 (-8.588)

-1.196 (-12.347)

-244.935 (-0.632)

Note: Data are quarterly before seasonal adjustments. R2 is the coefficient of determination adjusted for degrees of freedom. DW is the Durbin - Watson statistic and the values in parentheses are t-statistics. a b c d

First differences. From 1965 (III) to 1976 (IV). From 1977 (I) to 1986(1). From 1986 (II) to 1998 (I).

146


formulated by Taylor (1971) seems to be too crude to obtain any meaningful estimates of tax incentives on savings. Most salient is the fact that the coefficient of A T is an amalgam of two effects: (1) variations arising from changes in tax rates, exemption, and deductions, and (2) changes in the tax base, reflecting those in the general level of economic activity. Unfortunately, it is very difficult to isolate each change into separate components.10 If we could distinguish one factor from another, more conclusive results might be obtained in this estimation. According to my present calculations, tax incentives cannot be interpreted as having been a relevant factor contributing to the high savings rates in Japan for one decade after 1965. More attention should be paid to other possible factors, such as socio-cultural factors, the country's rapid economic growth, etc. Even if no special tax measures had been devised as incentives for saving, the savings rates would still have remained at a high level. While the effects of tax incentives on savings are ambiguous, their deleterious effect in producing tax erosion is very clear. As tax erosion in higher-income classes is considered evidence of an unfair tax system, its elimination would be desirable in terms of equity. At present it would no longer be necessary to use the tax system to stimulate savings while sacrificing tax equity. This will remain the case while personal savings stay at a high level. Greater importance should be attached to eradicating inequities in the present tax system in order to prepare for possible tax increases in the future. On the other hand, however, it is anticipated that the rate of personal savings will fall persistently towards the twenty-first century, as has been observed in Figure 8.1. This would be mainly explained by the coming of an ageing society in Japan. This being the case, it might be possible to revive tax incentives on savings again to promote the higher rate of personal savings. 10 It might be possible to avoid some of the shortcomings of the crude model at the macro level if we estimate the equations of the saving function at the micro level, using the household budget data. In so doing, it is necessary to devise some tax parameters including changes in tax law for savings-induced measures.

9 Effects of Taxation on the Distribution of Income It is widely acknowledged that taxation can reduce the inequality of income distribution or at least slow its growth. In particular, a progressive tax system is expected to play an influential role in determining the size distribution of income. When considering the distributional effects of taxation, emphasis should be placed on the individual income tax, because it is typically equipped with steeply progressive rates. Since a greater tax burden is thereby placed on higher-income classes than on lower-income-earners, it is commonly believed that income tax substantially narrows income inequality. The purpose of this chapter is to correlate information about the redistributional effects of taxation with the distribution of income, and to interpret them as one aspect of policy behaviour in Japan. Given the current state of available data, it is not easy to determine the influence of taxation on the distribution of income. Frequently, sceptics question the quality of the statistics that are obtainable and the results derived from them. The following discussion will be treated in two steps. First, I shall examine the influence of the individual income tax on income inequality, using tax statistics collected by the National Tax Administration (NTA). Based on these data, we can clarify the redistributional effects of taxation on income-earners (e.g. on wages and salaries, self-employment income, property income). Second, I shall use the same procedure on other data sources to ascertain the equalizing effects of income tax on income distribution. Similar calculations are made to assess income distribution more comprehensively by occupation, using non-tax statistics released by the Ministry of Health and Welfare (MHW).

MAIN FEATURES OF THE STUDY

Data Sources

Ideally, the impact of income tax on income equality should be measured by reliable statistics based on a comprehensive definition of income, including accrued or realized income (see Chapter 6). The nature of the data available in Japan, however, leads us to depart from this ideal level to a considerable extent. Before beginning major calculations, let us examine the data sources and their adequacy.

148


Broadly speaking, two kinds of statistics are available for the purpose of our estimations: (1) tax statistics, and (2) other statistics, based mainly on the Annual Household Survey. Whichever type is employed, data on the distribution of income suffer from a lack of comprehensiveness with regard to both population coverage and the definition of income. In this study the tax statistics data are utilized more, for the following reasons. 1. 2. 3. 4.

Information relating to tax revenues is more accurate. Time-series data on individual categories of income-earners are available. Data on higher-income classes are more abundant and reliable. The movement of certain capital gains is captured to some extent.

For our purposes, the accuracy of the information on taxes is most important, because the main aim of this analysis is to focus on the measurement of the redistributional effects of income taxes, rather than on the income distribution itself, as in many other studies.1 Hence, tax statistics provide a better data source than other statistics in that the latter have no specific intention to compile information relating to tax payments. Also, with regard to reason 2, tax statistics permit a more detailed inspection of income-earners for tax purposes, while statistics from the Annual Household Survey are confined to wage and salary-earners. Furthermore, it is extremely important that data on higher-income brackets can be obtained copiously from tax statistics. Under the income tax law pertaining between 1984-86, progressive tax rates began at 10.5 per cent of taxable income up to ¥500 000 and rose to a maximum of 70 per cent above ¥80 million. Since the redistributive effects of tax depend greatly on this progressive rate structure, basic data should cover as many income brackets as possible, especially at the higherincome levels. The highest income bracket found in tax statistics is that above ¥50 million, while the one used in other statistics is usually defined at much lower income levels, at best ¥10 million. Thus, we can get only very limited information for our purposes from non-tax statistics. Since capital gains are not included in the concept of national income, the usual approach for analysing income distribution tends to omit capital gains from the definition of income. They are, however, a very significant item in taxable income, and they play an important role in influencing income distribution. Therefore, capital gains cannot be neglected in this study, although the data on capital gains are very restricted even in tax statistics. On the other hand, tax statistics are not free of problems in collecting the necessary information. For instance, a basic criticism is that the income recorded on tax statistics may be subject to under-reporting and may be less accurate. However, the major proportion of income tax revenues (i.e. more than 80 per cent) is derived from taxes withheld at source from wages and salaries. In contrast, less than 20 per cent of total income is taxed under the self-assessed income tax provisions, whereby 1 There are a number of estimates in Japan that are not directly related to taxation. For example, see Mizoguchi and Takayama (1984) and their reference list.

Effects on Distribution of Income

149

individuals file their own tax returns and are therefore liable to under-report their incomes. Another criticism is based on the argument that tax statistics do not cover those lower income-earners who do not pay taxes. If this study were aiming to consider the redistribution of income through government expenditures (e.g. transfer payments), the lack of data on income-earners below the minimum taxable level would be significant in view of the redistributional effects of budgetary policies. In this study, however, the main focus is on the tax side of the budget, with the assumption that government expenditures will be taken as given so as to isolate the influence of income tax. Therefore, non-reporting persons, who do not need to pay taxes owing to exemptions and deductions, can naturally be omitted.2 A Simple Measure of Income Inequality Next, we shall examine the measurement of income inequality. Several alternative indicators are available, and reliable methods to measure income distribution have developed in recent years.3 Chiefly for the simplicity of calculation, I have utilized the familiar approach of the Lorenz curve and the Gini coefficient. As seen in Figure 9.1, the percentage of incomes cumulated from lowest to highest is plotted on the vertical axis, and the percentage of households or persons cumulated from lowest to highest is plotted on the horizontal axis. Obviously, the more uneven the income distribution, the greater the curvature in the Lorenz curve. Thus, the ratio of the area between the Lorenz curve and the line of equal distribution to the entire area below the line of equal distribution is useful in clarifying questions of income equality. This ratio is called the 'Gini coefficient'. By using before-tax and after-tax data, we can draw two curvature lines on the map of the Lorenz curve and compute two Gini coefficients; one is for the before-tax ratio, Rb, and the other for the after-tax ratio Ra. The effects of taxation on income distribution can be defined a follows:

tp is called the 'equalization coefficient'. The larger
25 years old) First son ( > 25 years old) Second son ( > 15 years old) Third son ( < 15 years old) This family composition is assumed to be the same in all income classes in both cases. (3) Salaries—entrepreneur, wife, and first son are assumed to earn the following salaries from the firm (in thousands of yen): Net income 500 800 1000 1500 2000 3000 5000 Entrepreneur 240 300 420 540 600 720 840 Wife and first son (each) 96 120 120 150 180 210 240 (4) Unincorporated businesses are permitted to deduct the salaries of the wife and the first son as a special tax benefit for proprietors and their family employees. Source: Tax Advisory Commission (1961).

182

Corporate Taxation and Taxes on Capital

eliminate double taxation. In practice, the problem of double taxation continues under the current corporate tax. The emergence of'quasi-corporations', however, has declined since the 1970s, and their importance as a group of taxpayers has been greatly reduced. One reason is that special tax treatment of 'deemed corporate income' has been established in favour of unincorporated, self-employed taxpayers since 1974. Taxpayers filing a blue return can elect to be treated as 'corporations' for tax purposes. Under this 'deemed' method, business proprietors can deduct their own salaries, as well as those of family employees, as business expenses. Furthermore, the deduction for employment income is also applicable to their own salaries. The net income computed for business proprietors is considered 'deemed corporate income' and is taxed at the lower rate of dividend income.3 Thus, business proprietors no longer needed to become 'quasi-corporations' to mitigate their tax burdens. As noted before, this 'deemed' system was finally abolished in 1992 to make the income tax burden among taxpayers more equitable. Debt and Equity Financing Taxation on dividends has been discussed in relation to the choice between debt and equity in business financing. During the postwar period, the Japanese government has continually stressed the improvement of self-financing through the stimulation of capital markets. For this purpose, the government has given special treatment to taxes on dividends. However, the extent of self-financing in business enterprises has gradually declined until 1980 (see Figure 10.1). This was generally thought to be an undesirable feature of the Japanese economy. Many have argued that the large amount of debt financing by corporations should be discouraged. Debt certainly makes good business sense if firms have a sufficient margin for paying fixed interest charges. However, firms may be squeezed financially when the economy enters into depression, and at such times their margin would disappear swiftly. Bankruptcies may occur because of defaults on interest and principal payments. It has been maintained that corporations should try to finance a substantial share of their capital requirements through equity capital (mainly retained earnings) to avoid these risks. Obviously, of relevance is the fact that equity financing can be unattractive from a tax point of view compared with borrowing. Corporations are allowed to deduct interest payments on borrowed capital from taxable income, but there is no corresponding deduction for dividends paid out to shareholders. At the 37.5 per cent tax rate, a corporation must earn ¥16000 before tax to pay ¥10000 dividends, but it need earn only ¥10000 to pay ¥10000 interest. Needless to say, this asymmetry 3 Business proprietors electing to be 'deemed corporations' for tax purposes amounted to 135597 in 1992. They represent 3.1 per cent of total taxpayers filing blue returns, which has been reduced from 7.3 per cent in 1986.

Principles of Corporate Taxation

Fig. 10.1

183

Trends in self-financing by corporate firms, 1968-1998

Source: Data presented to the Tax Advisory Commission.

makes the cost of equity finance more expensive for corporations than an equal amount of borrowed capital. Were there no tax considerations, there would be little to choose between alternative methods of business financing. However, when the tax system favours one method rather than another, a corporation, being very sensitive to tax considerations, will look for the cheapest source of finance. This is true in Japan, as it is in many other countries. Throughout the postwar period, the banking community has exerted strong pressure for the reduction or elimination of the tax on interest income. As a consequence of this pressure, complete or partial tax exemption for interest income has been put into effect, and this, in turn, has resulted in excessive reliance upon financing through borrowing in the corporate sector. Once the interest provisions were adopted, securities dealers and representatives of investment companies demanded favourable treatment for dividend incomes in order to maintain a competitive position in relation to the banking industry. Attempts were made to distribute tax benefits more or less equally among the banking and securities industries on the plea of encouraging capital accumulation, in particular, or increasing the ratio of equity capital. For example, a reduced tax rate on dividends and the separate taxation of dividend income as well as interest income at a specific rate were used to mitigate the tax on distributed profits. In addition, the Securities Transaction Council frequently proposed that dividends paid at the company level be made deductible as business costs (Securities Transaction Council 1960). In 1961, when the split rate system was first introduced, the Tax Advisory Commission distributed a questionnaire to investigate the mitigating effects of

184


dividend taxation on the equity financing of businesses (Tax Advisory Commission 1962, 38f). The results indicated that this tax device did not stimulate an increase in the relative share of equity in the supply of corporate funds to any substantial degree. Many Japanese tax experts outside of the government, as well as some members of the Tax Advisory Commission, were sceptical of the effectiveness of special tax measures in correcting the imbalance between debt and equity in corporate finance (see Tax Advisory Commission 1964a, 12). Figure 10.1 depicts the movement of the self-financing ratio and corporate tax rates since 1968. The self-financing ratio shows a long-run decline in the 1960s and the 1970s but turns upward slightly in the 1980s and the 1990s. Unfortunately, it is impossible to observe any significant relationships among the variations of the selffinancing ratio, corporate tax rates, and tax credit ratio. Apart from the issue of business financing, we should pay brief attention to the distribution of corporate dividends. Recently there has been a marked reduction in the percentage of corporate dividends received by individual shareholders. This is in sharp contrast with the enormous increase in the number of individual shareholders during the 1950s. The share of total dividends paid out to individual shareholders was 61.3 per cent in 1950, 33.5 per cent in 1975, 23.2 per cent in 1985 and, 22.5 per cent in 1995, although it rose slightly to 23.6 per cent in 1997. It is widely believed, however, that this phenomenon is not related to the current working of the corporate tax system. The Incidence of the Tax The question of 'who actually pays the corporate tax' is the most controversial issue in tax theory and practice. Although the corporate tax is normally levied on corporate profits, no one can identify who bears the ultimate burden. There are many divergent views among economists and businessmen which hypothesize the shortand long-run incidence of corporate taxation, but no scientific basis has yet been provided by economists to accept or reject the various arguments. Tax specialists have indicated two main groups which probably bear part of the corporate tax burden (see Goode 1951; Harberger 1962; Krzyzaniak and Musgrave 1963; Pechman 1986; etc.). One is the group that supplies capital and/or entrepreneurial skills to corporations. If the corporate tax is not shifted in the short run, the after-tax rates of return on invested capital decrease, and investment, in turn, may be discouraged. Given the fixed supply of capital, the burden of the corporate tax would then fall on the owners of capital. Another group of people who may share the burden of the corporate tax are the consumers who purchase the goods and services produced by corporations. In imperfect markets, business firms set their prices using the full-cost pricing method (i.e. the price is determined at the level that covers the full cost plus a profit margin).

Principles of Corporate Taxation

185

In this case, firms regard tax as a factor of cost and lift their prices to recover the cost. Alternatively, firms might aim at a certain after-tax rate of return on their corporate equity, and so the imposition of a corporate tax would influence the determination of output and prices. To secure the target level of return, the tax must be passed on to consumers. If this forward-shifting takes place, the corporate tax acts as a general sales tax on goods and services that firms produce. Also, part of the burden of the corporate tax may also fall on the workers, but such backward-shifting is not likely to occur so smoothly. We should note that the incidence of the corporate tax is a complicated issue and depends on the answers to empirical questions. However, empirical analyses do not provide a clear determination of the factors affecting corporate profits, prices, and wages. The evidence is generally inconclusive. Empirical analyses of corporate tax-shifting have been carried out less frequently in Japan than in such countries as the USA. Despite the limited research in this area by Japanese economists, there have been some attempts to study quantitative aspects of corporate tax-shifting in Japan. In particular, two articles by Furuta (1965, 1970) are worth noting, and the following is a brief survey using his estimates. In the first place, let us focus on two primitive approaches that have been used in the past: (a) the profit share approach, and (b) the corporate rate of return on capital approach. In the USA, the long-term change in corporate profit share before tax has remained fairly constant, in spite of a rise in the rate of corporation income tax. However, the corporate profit share after tax seems to have decreased with the rise in the tax rate. This fact has been cited as evidence that further tax-shifting has not occurred. On the other hand, looking over the long-term movement in the rate of return on capital in the USA, one notices that the rate of return on capital before tax rose along with the increase of the tax burden, while the rate of return after tax has tended to remain constant. This fact is cited as evidence that the long-term shifting of the corporate tax is nearly complete. After empirically investigating these figures pertinent to pre- and post-war Japan, Furuta concluded that three aspects of the corporate tax—the effective tax rate, the rate of return on capital, and the corporate profit share—show a surprising similarity between Japan and the USA with respect to the shifting of the corporate tax. Next, Furuta tried to apply three models developed for the American situation by (1) Krzyzaniak and Musgrave (1963) (K-M), (2) Kilpatrick (1965) (K), and (3) Gordon (1964) (G), to some industries in Japan. The well-known conclusion of the K-M model is that more than the full amount of the corporate tax burden is shifted forward to the consumer in the short run by US manufacturing firms. In the Japanese manufacturing sector, Furuta also demonstrated an 'over-shifting of the corporate tax'. The degree of shifting estimated by the K model for US manufacturing ranges from 62 to 94 per cent, but this conclusion differs from the K-M and G models. According to Furuta, this result is not appropriate in the case of Japan. In contrast to the 'over-shifting' of the K-M model,

186


Gordon asserts that tax-shifting by US manufacturing is not significantly greater than zero. Furuta finds that the Japanese results obtained by using the G model give a description of tax-shifting in Japan contrary to that of the USA. In sum, the empirical evidence by Furuta suggests that the corporate tax is shifted forward into prices to a considerable extent, in accordance with the K-M model. Furuta's conclusions seem to be accepted by many, although not all, scholars in Japan.

11

Corporate Tax Levels and Tax Incentives

In addition to raising revenues from corporate firms, the Japanese government used the tax system to promote specific policy objectives, such as capital accumulation, export activity, pollution control, and so on. This function persisted at least until the oil shocks of the 1970s. Such tax incentives took the form of numerous special tax measures. Since the late 1970s, however, the basic thrust of tax policy has shifted towards placing greater importance on tax equity and neutrality than on tax incentives. As a consequence, during the past three decades the number of special tax incentives has been curtailed substantially, a move especially affecting the corporate sector. This has tended to increase the tax burden on business firms, which often stimulates debates as to whether or not the corporate tax burden in Japan is too heavy by international standards. The aims of this chapter are twofold. One goal is to give an overview of tax incentive policies, paying much attention to the role played by the special tax measures for corporations. The other is to make clear the tax burden of the corporate sector, with special reference to tax incentive policies, referring to the past debates on the corporate tax burden and its policy implications. GENERAL

FEATURES OF TAX I N C E N T I V E POLICIES

Tax Incentives as an Industrial Policy Japan's industrial policy has attracted world-wide attention, and as a result has become an important target of study for scholars. It is frequently argued that during the postwar period the policies of the Japanese government to stimulate private saving and investment, export, and innovation were effected mainly through tax incentives (see Hollerman 1984). Although it is almost impossible to assess how effective these policies were as an industrial policy in quantitative terms, many people admit the significant role of tax incentives in the policy-making process. There were several steps in the development of tax incentives for the promotion of specific policy goals. First, earlier in the postwar period, both the promotion of exports and the encouragement of certain key industries (i.e. petrochemicals) This chapter draws heavily upon Ishi (1988c), and the empirical estimates of E. Tajika and Y. Yui.

188


formed a major consideration in tax policy. In the main, tax exemptions were used to achieve these objectives. Today, however, these provisions are of very minor significance. Second, a more important goal was to stimulate business saving and investment among targeted industries. Specific targeted industries were authorized to use various types of tax devices, such as special depreciations, tax-free reserves, and tax credits. In particular, we should note the interaction between investment and exports. Tax incentives encouraged the formation of capital in export-related industries as well as in those with export potential. It is argued that this interaction contributed to Japan's rapid economic growth. Third, tax incentives were directed towards developing technological innovation. Tax devices were used to stimulate such innovation indirectly by generating the growth of capital formation in important, selected industries. Major beneficiaries of such special measures were the steel and machinery industries, which may have thereby developed international competitiveness to a great degree. Lastly, as time went by, the benefits of special tax measures were extended from key industries to cover a wide variety of policy objectives such as pollution control and the enhancement of social welfare. The use of the tax incentives was diversified to advance various economic and social policy goals. However, since the late 1970s tax benefits for business investment, corporate savings, and industrial production have begun to be phased out because of the large fiscal deficit. The scope of tax incentive policies has thus become narrower.

Development of Tax Incentives As described in Chapter 3, changes in revenue losses resulting from the special tax measures are a good indicator of the past development of tax incentives. The data relevant to the corporate tax are available from 1965 onwards. Figure 11.1 illustrates the trend of revenue losses from special tax measures as a percentage of corporate tax revenues. In 1966 and 1972, revenue losses reach the top level of 9.0 per cent, and during the intervening years they maintain a relatively high level. Since 1972, however, they declined steadily to as low as 2.2 per cent in 1982, thereafter turning upwards to 3.8 per cent in 1987, and again moving around 3 per cent until 1991. Thereafter, they continued to fall to as low as 2.7 per cent in 1997. The low level of the ratio since the late 1970s mainly reflects the fact that a number of special tax measures have been abolished and have diminished in size. In general, special measures included in the corporate tax system are classified into two types: 1. tax exemptions and credits 2. tax deferrals (a) accelerated depreciation (b) tax-free reserves


189

Fig. 11.1 Corporate tax revenue losses from special tax measures as a percentage of corporate tax revenue, 1965-1997 Source: Data presented to the Tax Advisory Commission.

Broadly speaking, greater importance was placed on type 1 measures in the 1950s, while type 2 measures became more significant in the 1960s and 1970s (see Ikemoto etal. 1984). The origin of the special tax measures can be traced to the 1938 enactment of the Temporal Tax Measures Law which provided tax incentives for the war effort. Such measures proliferated during the war and immediate postwar period, but were drastically reduced when the Shoup Mission recommended tax reform in 1949. Since then, the list of special measures was considerably expanded again until the late 1970s. During the earlier period (i.e. in the 1950s), the exemption type of tax incentive measures was most prevalent. To give two examples, income from the production of certain products (mostly petrochemical products) was exempted from taxation between 1953 and 1966; and to promote export-oriented industries, export income was exempted from 1953 to 1963. Furthermore, to stimulate corporate savings, income spent on dividends for increased shares went untaxed from 1954 to 1957. Most of the exemption type of measures, however, were phased out in the early 1960s. It is important to notice that these exemptions were not applied to all industries across the board, but only to specific ones, such as infant or exportoriented industries. Instead of tax exemptions, the tax credit device has been used more often in recent years. In particular, the tax credit for experimental and research expenses, which started in 1967, is of special importance. Since 1981, an investment tax credit for energy-saving equipment has been allowed.

190


Tax deferral measures include accelerated depreciation and increased initial depreciation. Both promote business investment. These devices were used often in the 1950s, and some of them have remained in effect during much of the postwar period. However, many accelerated depreciation and increased initial depreciation plans have been replaced. For instance, accelerated depreciation for important industrial equipment was made an option in the form of a three-year, 50 per cent more than normal, rate of depreciation from 1951 to 1961. Also, additional initial depreciation for important industries (i.e. 50 per cent write-off during the first year) was introduced from 1952 to 1961. In the 1960s, depreciation plans expanded to new areas. For instance, accelerated depreciation was granted to small and medium-size enterprises in 1963. The initial depreciation on machinery and equipment used for the prevention of environmental pollution was increased in 1967. As time progressed, the trend towards expanding accelerated depreciation provisions began to slow. In 1972 accelerated depreciation for export industries was abolished, and in 1973 increased initial depreciation for important industries was reduced. Overall, the recent trend has been to reduce the range of tax devices, especially the accelerated depreciation provisions. Another type of tax deferral measure is that of tax-free reserves. Although the term 'tax-free reserves' is used in English, in Japanese two types of tax-free reserves are usually distinguished. The first represents reserves that are set aside for purposes of business accounting (hikiatekin), while the second type is to stimulate involvement in risky activities (junbi-kin). Because thejunbi-kin are those reserves that may not be duly justified by generally accepted accounting principles, these are officially listed among the special tax measures. Both types of tax-free reserves have been allowed since the early 1950s. Those set aside for tax-free accounting permitted reserves for bad debts and special repairs in 1950, reserves for retirement allowance in 1952, and reserves for bonus payments and losses on returned goods in 1965. Special inducements included reserves for price fluctuations in 1952 and reserves for drought in 1952. Numerous other reserves were added in the 1960s and 1970s. Of course, the size of the tax-free reserves and the amount allowed annually are subject to specific limits, but in all cases, the limits are generous in relation to the potential risks. Thus, the tax-free reserves are criticized on the grounds that they conceal profits retained from the amount subject to income taxation. More detailed information on recent special tax measures for corporations in comparison with those in 1977, 1987, and 1998 is given in Table 11.1. Total amount of revenue losses fell from ¥457 bn. in 1987 to ¥356 bn. in 1998 although they doubled from 1977 to 1987. We see three broad types of tax incentives: (1) special depreciation, (2) reserves, and (3) tax credits and allowances. In 1998, the third type occupied the largest share (47.8 per cent), having greatly expanded from 16.2 per cent in 1977. Besides a big increase in credit for research and experimental expenditures, three new items (3(b), 3(c), and 3(e)) to promote specific investments have been added.

Corporate Tax Levels and Tax Incentives Table 11.1 1998

191

Revenue losses of special tax measures of corporations, by type, 1977, 1987, and 1977

¥bn.

1. Special measures for depreciation (a) Special depreciation on qualified equipment (e.g. machinery and equipment for reducing pollution, saving energy, recycling, etc.) (b) Special depreciation on manufacturing machinery used in underdeveloped regions (c) Special depreciation on machinery and equipment acquired by small- and medium-sized firms (d) Others

1998

1987

¥bn.

%

%.

¥bn.

%

128 34

56.1 14.9

129 33

28.2 7.2

99 23

27.8 6.5

14

6.1

23

5.0

11

3.1

43

18.9

43

9.4

24

6.7

37

16.2

30

6.6

41

11.2

2. Reserves (a) Reserve for overseas investment loss (b) Reserve for losses caused by repurchase of computer (c) Others

63 12

27.6 5.3

76 2

16.5 0.4

87

24.4

1

0.4

24

5.2

50

21.9

50

10.9

87

24.4

3. Tax credits and allowances (a) Credit for research and experimental expenditures (b) Tax measures to promote investments for efficient use of energy (c) Tax measures to promote investments by small- and medium-sized firms on equipment utilizing electronic devices (d) Special deduction for income derived from overseas technical service transactions (e) Tax measures to improve managerial fundamentals of small- and medium-sized enterprises (/) Others

37 17

16.2 7.5

252 90

55.1 19.7

170 44

47.8 12.4

0

0.0

55

12.0

0

0.0

52

11.4

41

11.5

12

5.2

27

5.9

5

1.4

0

0.0

18

3.9

8

2.2

8

3.5

10

2.2

72

20.2

228

100.0

457

100.0

356

100.0

Total

Source: iMOF, Tax Bureau (1977, 1987) and NTA 1998.

192


On the other hand, the relative share of the first and second types in 1987 is reduced considerably. Particular attention should be paid to the sharp decline of the tax-free reserves system. Reserves for price fluctuation, for instance, were abolished in 1986, and this helped to recover revenue losses. Use of Tax Incentives What size corporation actually benefits most from the special tax measures? In order to estimate the effect of special tax measures on corporate tax burdens, we need to explore what is meant by the effective rate of tax. Obviously, statutory or nominal tax rates are not representative of relative tax burdens on corporate incomes because the government has adopted numerous devices, such as tax-free reserves, accelerated depreciation allowances, investment credits, and others, which reduce that burden. Tax rates on a nominal basis only imply a weighted average of the multiple-rate structure with split and reduced tax rates, before the adjustment of special tax measures. By contrast, the effective rate of tax paid by corporations on gross profits is estimated after adjusting for special tax measures. Table 11.2 summarizes the two types of tax rates, the nominal and the actually paid, at three different levels of paid-in capital. Of major importance is the fact that we observe a discrepancy in the corporate tax burden by size of corporation. On an actually-paid basis, corporate incomes are broadened to include the major items eroded by the special tax measures (i.e. tax-free reserves, accelerated depreciation, and special deductions), and the effective tax rates are calculated in column (8). These are to be compared with nominal tax rates in column (3). Figure 11.2 depicts the movement of the two tax rates listed in Table 11.2 in a more clearly visible fashion. There are a couple of points worth noting. To begin with, nominal tax rates consistently showed an upward tendency for corporations of all sizes in the first 12 years until 1985. Afterwards, they began to fall up to 1996. This suggests that the corporate tax burden was raised by the government on a statutory basis, but it began to decline substantially after 1986, reflecting the more recent movement towards lowering corporate tax rates (see Table 10.3). Particular attention ought to be paid to the sharp declining trend in the 1990s. In addition, medium-size corporations with ¥0.1-10 billion of paid-in capital incur a heavier tax burden than the other two types of corporations. Small corporations benefit from reduced rates of tax on a larger portion of their corporate income, while large corporations in general could pay out more dividends, taxed at a lower level. By contrast, medium-size corporations seem to have relatively fewer opportunities to make use of such preferential tax rates. To turn our attention to another indicator, the movement of effective tax rates on an actually-paid basis has shown ups and downs over the long run, reflecting the availability of special tax concessions during each period. It is often pointed out that large corporations are given more opportunities to make use of tax-free reserves

Table 11.2 Fiscal year

Effective corporate tax rates before and after adjusting for special tax measures, by size of corporation, 1973-1996 Corporate income (¥bn.)

Corporate tax revenue (¥bn.)

(2)/(D

(1)

(2)

(3)

Taxfree reserve (¥bn.)

Accelerated depreciation (¥bn.)

(4)

(5)

Special deduction for overseas technical service' (¥bn.) (6)

Tax credit for experimental and research expenses (¥bn.) (7)

(8)

1.1 0.5 1.1 0.8 1.7 1.9 0.8 1.3 0.7 1.9 3.3 1.1 5.1 0.6 0.5 0.7 0.5 0.2 1.8 0.4 0.2 0.7 0.0 0.1

1.3 1.8 1.6 1.6 1.4 11.0 23.1 18.1 13.5 6.2 5.9 8.0 15.9

32.4 34.1 35.7 35.0 35.7 35.6 35.6 35.9 37.9 38.2 39.3 38.8 39.8 40.0 39.1 38.5 38.6 37.1 35.0 34.9 35.4 35.5 35.2 35.4

(2)-(7) (l) + (4) + (5) + (6)

Case 1: less than ¥0.1 billion of paid-in capital

1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996

5 994.2 7331.1 5 626.5 5 700.0 5 652.6

6 395.2 8 168.7 8 960.2 9098.2 9 026.2 9 009.5 9 584.3 10436.7 10412.6 12 624.4 14490.0 17831.2 19 176.5 19 347.4 17208.0 14359.8 13071.1 12253.2 14050.5

2036.9 2578.7 1 992.2 2021.9 2 007.0 2 274.8 2951.1 3 254.4 3 406.4 3 458.8 3 472.0 3 768.4 4 176.9 4 146.2 4959.1 5637.5 6 957.6 7 188.4 6 868.2 6 064.8 5 057.4 4603.1 4 303.2 4970.5

34.0 35.2 35.4 35.5 35.5 35.6 36.2 36.3 37.4 38.2 38.5 39.3 40.0 39.8 39.3 38.9 39.0 37.5 35.5 35.2 35.2 35.2 35.1 35.4

162.6 58.3 -71.1 64.6 -44.5 -7.6 81.6 5.2 -154.3 4.4 -204.8 59.9 -24.9 -80.5 42.4 -10.1 -3.8 -14.0 16.6 3.0 -34.5 -30.7 8.7 11.0

116.2 177.9 14.5 3.0 15.6 -26.1 -11.7 42.9 8.0

2.0 3.7 35.9 37.0 -5.2 — 11.8 76.6 128.2 112.5 172.8 68.1 -62.0 -75.8 -47.0 -57.5

15.2

16.7 27.0 22.0 33.2 27.1 28.6 4.4 6.4 9.1 7 3

Table 11.2 Fiscal year

Contd. Corporate income (¥bn.)

Corporate tax revenue (¥bn.)

(2)/(l)

(1)

(2)

(3)

Case 2: ¥0.1-10 billion of paid-in capital 1973 4072.1 1.445.9 1974 4483.7 1 675.2 1975 3 055.2 1 182.8 1976 3 766.8 1 461.4 1977 4480.2 1 742.9 1978 5 092.8 1 984.5 1979 6 183.2 2416.1 1980 7 280.7 2 844.8 1981 7330.6 2 935.6 1982 7413.5 3 048.2 1983 7583.6 3 115.0 1984 8 250.5 3 440.5 1985 8 390.0 3 562.8 1986 8 309.5 3519.8 1987 9 191.7 3 854.2 1988 LO 726.4 4 424.6 1989 11 770.0 4 865.4 1990 11393.6 4527.1 1991 11 728.8 4 409.0 1992 10436.0 3912.9 1993 9 070.7 3401.5 1994 8 439.5 3 164.8

35.5 37.4 38.7 38.8 38.9 39.0 39.1 39.1 40.0 41.1 41.1 41.7 42.5 42.4 41.9 41.2 41.3 39.7 37.6 37.5 37.5 37.5

Taxfree reserve (¥bn.)

Accelerated depreciation (¥bn.)

(4)

(5)

133.4 117.3 -55.8 80.5 50.7 -46.7 -11.1 39.9 -76.3 -108.3 -107.3 -26.3 -29.5 -29.5 -23.7 -1.2 27.6 -4.5 29.7 11.2 10.5 4.3

60.3 91.9 -0.3 -26.2 27.1 -42.9 -40.0 -41.

-35.3 -32.4 -23.5 -30.1 -9.8 -33.6 -16.2 13.8 -7.7 -3.7 -5.4 2.5 -17.1 -1.4

Special deduction for overseas technical service' (¥bn.) (6)

3.5 7.6

6.0 9.5 8.8 10.9 13.0 17.8 16.8 14.6 17.7 17.3 13.3 8.1 6.0 6.5 7.7 8.3 3.8 4.9 8.3 4.3

Tax credit for experimental and research expenses (¥bn.)

(2)-(7) (l) + (4) + (5) + (6)

(7)

(8)

12.0 19.0 9.0 11.8 10.1 11.6 15.1 18.1 21.4 24.8 24.0 21.9 24.6 23.3 19.3 19.9 20.2 21.3 22.5 18.5 11.2 9.4

33.6 35.2 39.1 37.8 37.9 39.3 39.1 38.7 40.3 41.5 41.4 41.6 42.3 42.4 41.9 41.0 41.1 39.5 37.3 37.3 37.4 37.4

1995 1996

8 629.9 9010.5

3 236.9 3378.9

Case 3: ¥10 billion and over paid-in capital 1973 1 078.4 3 074.8 1974 1 151.3 3 108.9 1975 2 650.2 1 014.1 1976 3 434.5 1317.5 1977 4296.1 1 647.1 1978 1 683.7 4359.7 1979 1 958.1 5 080.9 1980 2359.0 6089.1 1981 2 740.4 6 934.9 1982 7395.2 3 007.0 1983 7901.8 3 207.3 1984 3 882.7 9412.0 1985 11459.2 4 829.6 1986 4 866.0 11 572.7 1987 5400.1 12929.9 1988 6409.8 15646.1 1989 17811.1 7 300.0 1990 19811.1 7819.1 1991 6 926.4 18445.3 1992 15274.8 5 727.9 1993 4 866.4 12977.2 1994 4079.1 10877.6 1995 4031.4 10750.4 1996 5 806.0 15482.8

37.5 37.5

18.2 -21.4

6.3 -13.6

4.0 2.9

9.4 14.0

37.3 37.5

35.1 114 38.3 38.4 38.3 38.6 38.5 38.7 39.5 40.7 40.6 41.3 42.1 42.0 41.8 41.0 41.0 39.5 37.6 35.7 37.5 37.5 37.5 37.5

165.7 151.9 — 6.7 117.3 47.9 -61.1 10.7 32.4 124.2 -32.8 -80.9 61.1 -2.8 16.1 -14.8 88.7 244.3 153.7 268.2 -24.1 111.2 189.6 7.7 105.1

106.0 103.5 -22.2 -41.4 39.6 -59.8 -60.7 -80.6 6.5 -59.0 -72.0 -69.4 -26.8 -34.5 -7.7 33.1 15.0 -3.3 52.3 4.2 -19.8 -18.5 91.7 3.5

19.0 24.6 12.9 19.9 31.1 25.7 25.8 30.6 30.2 38.5 35.8 37.7 48.8 37.2 33.1 39.5 47.7 53.4 37.5 38.0 29.6 34.1 29.9 13.1

14.9 20.4 4.7 10.5 16.6 20.6 30.8 38.3 54.9 63.6 53.2 57.6 89.0 68.0 49.3 74.8 111.6 133.8 162.2 92.1 28.9 23.1 24.4 63.4

31.6 33.4 38.3 37.0 36.9 39.0 38.3 38.2 37.8 40.1 40.5 40.5 41.3 41.4 41.3 40.1 39.7 38.4 36.0 36.9 36.9 36.6 36.8 36.8

Source: Unpublished data presented to the Diet by the Tax Bureau, MOF.

Fig. 11.2 Effective corporate tax rates before and after adjustment for special tax measures, by size of corporation, 1973-1996: (a) nominal basis; (b) actually paid basis Source: Table 11.2.

and accelerated depreciation than smaller corporations. For instance, tax-free reserves such as retirement allowances or bonus payments are used mostly by large corporations. As a result, the regular trends in the tax rates of the three types of corporations on a statutory basis are thrown into disorder when we turn to the actually-paid tax rates. Note should be taken of the lowest tax rate paid by the largest corporation in the early 1970s.


197

ARE CORPORATIONS OVERTAXED?

The Corporate Tax Burden The level of corporate tax burden is important for a number of reasons.1 It may affect the savings rate and the cost of capital, and in turn the economic growth rate. Thus, different corporate tax burdens may lead to differential growth rates by different countries. In addition, in a world economy where international capital can move relatively freely, corporate tax burdens may contribute to determining the inflow and outflow of capital (see Gravelle 1983). Despite the immense importance of the subject, no attempts have been made to clarify the relationship between the corporate tax burden and other economic phenomena. I am not so ambitious as to propose any conclusive result, but simply approach the subject by estimating the average rate of corporate tax burdens.2 It is possible to gain a simple estimate of the corporate tax burden by using the statutory structure of the corporate income tax. The MOP has traditionally used this formula to clarify the tax burden in the corporate sector. The corporate tax had a somewhat complicated structure until 1989, consisting of multiple rates as shown in Chapter 10. For the corporate tax burden, the MOF usually computes what we shall call a nominal rate of tax burden (II) on a statutory basis. If we take the split tax system before 1989, the formula of the corporate tax burden is as follows:

where f r = tax rate on retained profits, td=tax rate on distributed profits (both of these are national taxes), t/ = tax rate of local corporate income tax, and f;, = tax rate of local enterprise tax. If we use the statutory tax rate data, II is 52.92 per cent in 1984-6 (i.e. tr= 0.433, fj = 0.333, fe = 0.173, ffc = 0.12). In this calculation, it is assumed that corporations pay out dividends of 30 per cent of their before-tax earnings. Simply calculated on the statutory rate basis, this is the nominal figure of corporate tax burden presented by the MOF. Thus, nominal tax rates are readily obtainable by using the MOF formula, but these are not representative of the actual tax burdens paid by corporations. The MOF estimate merely covers the sum of national and local taxes on corporate income on a statutory basis. Put another way, the effects of tax incentives offered

1 In discussing corporate tax burdens, I imply that corporations bear the ultimate burden of the corporate tax without any forward or backward shifting. 2 Most popular discussion has so far focused on the effect of income taxation on the cost of capital. There are several papers worth noting pertinent to the Japanese experience: see e.g. Shoven and Tachibanaki (1988), Tajika andYui (1987), and King (1986).

198


through the special tax measures are not taken into consideration. If such incentive effects are included in corporate tax burdens, the results obtained are very different. Keidanren (the Federation of Business Organizations) has criticized the MOF's concept of nominal tax rates, showing that Japan has the highest level of corporate tax burden of any major industrial country (Kubouchi 1984). This is due mainly to the past curtailment of special measures for tax incentives. The Keidanren estimate has induced a controversial debate among many economists, including the MOF staffs.3 Keidanren insisted that the corporate tax burden should be estimated to take account of the effects caused by tax incentive policies. Thus, it calculates the effective rate of tax burden in the corporate sector based upon the tax amounts actually paid by corporations. Specifically, the MOF estimate II is changed into the Keidanren's effective rate Ile, by applying certain coefficient to the former:

where Ts = revenue loss from special tax measures, and Tf= final tax payments. In Table 11.3, the results of the different estimates of corporate tax burden are indicated in an international comparison. The most conspicuous result is that the smallest gap between the nominal and effective rates of the corporate tax burden appears in Japan. The statutory-base figure presented by the MOF does not provide a realistic picture of the actual tax burden. If greater importance is placed on the Keidanren estimate, the Japanese tax burden in the corporate sector is higher than in any other industrialized country. This reflects the fact that the curtailment of special tax measures in the 1980s has been accelerated especially in the corporate sector. In contrast, governments in other major countries have taken the initiative to introduce bold measures for tax relief to

Table 11.3 An international comparison of corporate tax rates, 1984

MOF Keidanren a b c

Japan

USA

UK

W. Germany

France

52.92 51.57

51.18 32.28a

52.00 18.06b

56.52 49.84

50.00 45.70C

1985 figure. 1982 figure. 1980 figure.

Source: Kubouchi (1984). 3 On this point, Hollerman (1984) states: 'At the present time, a controversy is being waged between Keidanren (Japanese Federation of Economic Organizations) and the Ministry of Finance concerning the level of corporation taxes, in which Keidanren maintains that Japanese corporations are more heavily taxed than those in the USA. This controversy symbolizes the new state of events in Japan.'


199

promote economic revitalization, such as the ACRS (accelerated cost recovery system) in the USA, or the initial allowance in the UK. It is evident that the relative level of effective tax rates in Japan and the other major countries (e.g. the USA) has been reversed in the past decade (see Gravelle 1983). Now we shall compare the corporate tax burden in Japan with that in the USA. Since Japan and the USA are closely linked in business transactions, the level of tax burden in the corporate sector must affect substantially the flow of capital and the movement of business activities. Therefore, it is important to explore whether Japan's corporate tax burden is heavier or lighter than that of the USA in a meaningful way. Figure 11.3 compares nominal rates of national and local corporate taxes in Japan with those in the USA for the years 1970-93. These are considered to be ex post rates based upon actual tax payments and reported corporate income for tax purposes. Clearly, Japan shows an upward movement of tax rates until 1989 while the USA shows a declining trend until recently. This result reflects the fact that Japan has continued to increase basic rates of corporate taxation since 1970, whereas the USA has lowered its corporate tax rates and expanded investments tax credits. However, nominal corporate tax rates have been declining since the late 1980s in both Japan and the USA, although Japan's rate has fallen more swiftly. However, such a comparison using nominal tax rates to contrast the relative tax burdens of two countries is of little significance economically. In the first place, the two countries have adopted a wide variety of special tax measures, such as tax-free reserves, accelerated depreciation, and tax credits. As a consequence, the types of

Fig. 11.3

Nominal tax rates of corporate taxes, Japan and USA, 1970-1993

Nominal tax rates arc defined as the ratio of actual tax payment (including both national and local corporate taxes) to reported taxable income. Source: Tajika and Yui (1999), calculated from the following statistical data: NTA, Actual Performance of Corporate Firms in View of '[ax Statistics; Ministry of Home Affairs, Actual Reports of Local Public Finance; US Internal Revenue Service, Statistics of Income: Corporation Income Tax Return; and US Economic Reports of the President, yearly.

200


taxable income that are reported differ to a considerable extent. Second, different increases in the price level in each country influence the real tax burden of the corporate sector. We have already examined the first of these from several standpoints, but have not yet touched on the relation between the corporate tax burden and inflation. Since Japan and the USA have experienced different levels of inflation in the past decade or so, we need to include inflationary effects as well as incentive provisions when comparing the corporate tax burdens of the two countries. For the computation of the tax base, corporate net income is calculated by subtracting the expenses incurred in a firm's receipts from the gross receipts or sales. If all expenses are paid when sales are made, the difference between receipts and expenditures poses no problem in providing a correct concept of net income. There are, however, two complications (see Pechman 1987, 173): (1) business expenses are often paid long before sales are realized, and (2) expenses and sales are generally assessed at different price levels. Thus, a variance in accounting gains and losses appears because of inflation when calculating the tax base of corporate taxation. There are mainly four components included in accounting methods that are affected by inflation: tax-free reserves, depreciation allowances, net financial assets and liabilities, and inventory valuation. In order to define a concept of income adequate for taxing corporate profits, we need to have economic income that is not embodied in current tax law but is derived from 'accretion' or 'economic power'. In the case of business firms,4 this is equivalent to the variation in the real market value of corporations before profits are distributed to shareholders for a specific period. When we try to compare the level of corporate tax burden between Japan and the USA, it is necessary to calculate the effective tax rate, that is, the ratio of actual tax paid to economic income, not reported taxable income. The concept of economic income provides a common base to measure the corporate tax burden under different tax systems. An interesting attempt is made by Tajika and Yui (1999) to estimate effective tax rates for both countries on the basis of the economic power concept, as a result of making all adjustments to correct for inflation. We see first the relation between economic income and reported taxable income, and then the deviation between these two incomes. The incomes are related as follows: Economic income = taxable income + tax-free reserves + depreciation + net gains of financial liabilities + inventory valuation + others.3 Table 11.4 indicates the ratio of each item relative to economic income, neglecting the last item. To begin with, gains resulting from tax-free reserves are added to 4 If this concept is applied to individuals, it is equivalent to the sum of personal consumption plus the net increase in the value of the individual's assets during the taxable period, according to the Haig-Simons formula (see Ch. 5). ' 'Others' include both effects of income adjustment and local taxes.


201

Table 11.4 Taxable and economic incomes of non-financial corporations, Japan and USA, 1970-1993 (%)

Year

Taxable income -^ economic income (1)

Tax-free reserves (2)

Depreciation (3)

Net gains of financial liabilities (4)

Inventory valuation (5)

Japan 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993

70.0 70.5 67.9 58.1 60.5 87.4 70.1 79.5 83.7 75.7 75.7 90.5 88.1 85.6 84.3 84.0 84.4 81.0 84.6 81.0 84.3 82.4 85.0 79.5

4.2 3.1 4.3 4.3 5.3 2.2 3.3 5.0 2.3 2.4 1.5 0.9 -1.8 -0.3 1.8 1.8 0.3 1.1 1.6 2.3 0.4 1.6 0.5 1.9

14.7 18.6 15.0 8.6 6.2 -0.9 6.9 6.4 7.3 5.5 4.1 5.5 6.2 8.6 7.5 8.9 10.6 10.5 9.8 6.9 7.6 8.3 11.7 5.1

5.7 -2.3 10.7 40.4 37.3 8.9 20.2 5.4 -3.0 24.0 27.2 0.6 0.5 -2.4 0.5 -3.8 -13.7 -3.9 -0.8 5.2 2.2 1.0 -8.4 0.3

-1.1 0.6 -4.7 -15.0 -14.7 -12.9 -6.6 -2.7 1.9 -12.6 -13.8 -5.4 -1.6 1.6 0.6 3.6 10.0 5.3 0.1 -0.9 -1.8 -0.4 1.5 2.8

USA 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982

100.4 97.1 95.4 101.3 107.3 101.2 100.5 98.7 104.6 107.7 104.4 104.1 93.1

0.8 0.6 0.4 0.6 1.6 -0.5 0.2 1.2 -0.1 0.5 1.6 0.2 1.2

-1.3 -0.5 1.4 1.1 -3.6 -7.6 -5.8 -3.7 -6.5 -7.6 -9.4 -8.2 -3.4

-0.5 -0.3 -0.3 0.4 6.3 4.5 1.7 1.1 1.3 3.7 4.4 -0.4 -1.1

-8.5 -4.4 -5.4 -14.9 -27.3 -6.1 -6.5 -6.9 -8.3 -14.1 -12.5 -7.2 -2.9


202

Table 11.4

contd. Taxable income -reconomic income

Year

1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993

(1)

87.2 82.1 76.3 73.4 86.2 88.8 87.3 90.0 95.5 88.6 87.4

Tax-free reserves (2)

Depreciation

-0.5 0.8 -0.1 0.4 -0.1 0.4 0.4 1.2 0.6 0.6 -0.2

financial liabilities

Inventory valuation

(3)

(4)

(5)

4.5 9.1 15.4 16.3 9.2 7.1 6.2 2.2 -2.8 3.4 5.3

-0.5 -1.2 0.3 0.7 -0.8 -0.3 -0.3 -0.0 -0.2 -1.0 -0.9

Net gains of

-2.3 -1.6 -0.3 1.1 -3.3 -3.4 -1.8 -1.4 -1.5 0.8 0.2

Note: Columns (2)-(5) are computed as ratios of economic income. The difference between (1) and (2) + (3) + (4) + (5) implies other effects, such as income adjustment and local taxes. Source: As Fig. 11.3.

taxable income in column (2). Next, since depreciation is based on bistorical costs, gains and losses resulting from depreciation allowances should be recognized in column (3) as inadequate depreciation in the period in which they accrue. Also, real gains and losses on net financial assets and liabilities in column (4) should be included in the calculations of income. Lastly, the effects of inventory valuation due to price changes are added to other items in column (5). Using the Tajika-Yui estimates, we find a number of interesting facts which explain some of the differences between economic and taxable incomes in Japan and the USA. First, economic income has always been larger than taxable income in Japan, while in the USA the former is often smaller than the latter, especially during periods of high inflation from the mid-1970s to the early 1980s. Second, tax-free reserves have constantly got taxable income narrower than economic income in Japan. This was true in the 1970s. By contrast, they have not had so strong an impact in the USA. Third, taxable depreciation has been allowed more often than economic depreciation allowances for tax purposes, resulting in gaps between the two incomes. Japan generally allowed generous depreciation for tax purpose, generating a narrower tax base. On the other hand, the USA had a shortage of depreciation allowances until 1982, resulting in larger taxable income. However, as a result of introducing the ACRS in 1982 under the Reagan Administration, the use of accelerated depreciation contributed greatly to reducing the tax base after 1983. Fourth, inflation plays an important role in causing net gains of financial liabilities. Overall, the main factors that cause deviations between economic income and taxable income are net gains of financial liabilities in Japan while they are of little importance in the USA. Fifth, in Japan inventory evaluation also played an


203

Fig. 11.4 Average effective tax rates of corporate taxes, Japan and USA, 1970-1993 Average effective tax rates are equal to the ratio of actual tax payments (national and local) to economic income. Source: As Fig. 11.3.

important role in expanding taxable income due to gains from inventory caused by inflation in the 1970s, but thereafter its role turned to the reverse, reflecting inventory losses due to the slowdown of the wholesale price level after the mid-1980s. On the other hand, inflation in the USA has generally tended to make taxable income less by generating inventory gains. Turning now to consider the corporate tax burden calculated on the basis of economic income, Figure 11.4 shows the average effective tax rates in Japan and the USA during the period 1970-93. While in the USA tax rates continue to fall in the long run, those in Japan tend to move upward. Before 1980 Japan's tax rates were relatively lower than those in the USA (with the exception of 1975), but they began to exceed the US level after 1980. Particular attention should be paid to the widening gap between the effective tax rates of the two countries in the 1980s and the 1990s. In fact, the deviation in the tax rates expanded to as much as 20 per cent in 1984. Two major factors behind this phenomenon are the fact that gains on depreciable assets resulting from accelerated depreciations have sharply increased in the USA, and that net gains of financial liabilities have become minor in Japan. As is clear from the evidence, Japan's corporate tax burden is now much heavier than the US tax burden if one calculates the tax burden using the economic income concept as a common base. In this chapter, we have analysed the corporate tax burden in view of tax incentive policies. No attempt has been made to clarify the effectiveness of tax incentives on the promotion of private investment and savings. A further task is required to investigate the relationship between tax incentives and their policy goals, including the promotion of corporate investment, although such a study would be very difficult to achieve in quantitative terms (see, e.g. Takenaka 1986).

12

Inheritance and Gift Tax Estate transfers, bequests, and gifts are generally regarded as appropriate objects of taxation. There are two main objectives for taxing these items. First, such transfer taxes may check the accumulation of undue concentrations of wealth and thus may promote a more equitable distribution of economic resources. Second, taxes on property transfers raise revenue. They tend, however, to comprise only a small percentage of the total tax revenues in advanced countries, despite the imposition of high statutory marginal tax rates. Therefore, greater importance has been attached to the first objective in justifying the imposition of taxes on property transfers. In societies where property is privately owned, the government is responsible for protecting the property rights of individuals; but at the same time, it intervenes in the transfer of property from one generation to the next. Taxes on property transferred from individuals to their heirs have long been considered a desirable form of direct taxation, and constitute one of the oldest forms of taxes. In Japan, such taxes have been used exclusively at the national level from the inception of the tax system. We will consider two main topics in this chapter. In the first section, we explore the structural features of the Japanese inheritance and gift taxes in an historical perspective. In the second, we consider the role of property transfer taxes in affecting the tax burden and its redistributional effects. STRUCTURAL FEATURES

Accession Tax

The modern estate tax took effect in 1905 and was based on the value of the decedent's property. It was retained until 1950, when the government adopted the accession tax proposed by the Shoup Mission. The Mission recommended the accession tax as one of the best forms of transfer taxation for the two objectives noted above. The basic framework of the accession tax is as follows (see Shoup Mission 1949, 143-55). 1. It is a cumulative tax levied on the recipients of gifts and bequests. 2. It is graduated progressively according to the total amount of gifts and bequests received by a given individual. 3. The manner of its application is similar to that of the gift tax: when a gift or bequest is received, it is added to the total of taxable gifts and bequests previously received, and tax is computed on that total according to the current set of rates.

Inheritance and Gift Taxes

205

Tax is also computed at the current set of rates on the previous cumulative total, and the difference between the two taxes becomes the amount of tax currently due. In theory, there are two major forms of transfer taxation, depending upon how the tax base is determined and when the transfers are made (see Pechman 1987, ch. 8; Institute of Fiscal Affairs 1978, ch. 15). One is an estate tax on the privilege of transferring property at death, based upon the size of the entire estate. The other is an inheritance tax, in which tax is imposed on the privilege of receiving property from the decedent. This tax is based on the size of each individual share of the estate. In addition, the gift tax has been established to complement taxes on estates and inheritance. Without the gift tax, an individual would distribute his estate to his successors before he died in order to avoid or mitigate the burden of transfer taxes. When the accession tax was proposed, the Shoup Mission pointed out that it had several advantages over the separate estate and gift taxes. The main points of the Shoup Mission's argument are as follows. 1. The tax burden will be distributed more equitably among heirs. Under the accession tax, the total tax on the entire property is lower if the estate is divided among two or more heirs than if it is left all to one person. Under the estate tax, on the other hand, the tax is about the same, regardless of the number of beneficiaries. The former result is to be preferred, since a given sum of money spread thinly among a number of people does not give rise to as much taxpaying as the same sum concentrated in the hands of one person. Thus, the accession tax promotes a wider distribution of wealth than does the estate tax. The cumulative feature of the accession tax prevents circumvention of this intention through multiple gifts. It also encourages donors to make gifts and bequests to those who have not already received gifts and bequests from others. 2. Under the existing estate and gift tax law, in which each tax has its own separate rate scale, it pays the individual to split his giving carefully between gifts during life and bequests at death. If he achieves just the right combination, he minimizes his total gift and estate taxes. Under the accession tax, it makes no difference to the total tax whether the gift is made during life or at death. The accession tax is thus a neutral tax; the gift and death tax combination is not. The basic nature of the accession tax is evident from the following statement: The accession tax, so far as it applies to transfers at death, is simply an inheritance tax, and as such has a long record of experience in many countries. So far as it applies to gifts during life, it is similar to the well-known gift tax. The accession tax that we recommend for Japan is therefore a combination of two familiar taxes, and affords no hazard in the way of experimentation with new types of taxation. (Shoup Mission 1949, 145.)

Basically, the accession tax is closer to an inheritance tax than an estate tax and so is regarded as a transfer tax derived from a modification of the inheritance tax principle. If each receipt of the inheritance or gift is taxed separately in its unmodified

206


form, this will lead to serious defects. For instance, two individuals would pay different taxes on equal inheritances if one received them in a lump sum and the other received them from several decedents. Under the accession tax, this defect can be remedied. Total lifetime acquisitions of a given individual through inheritances and gifts are levied on the basis of a cumulative formula. One calculates the tax in any one year by subtracting the tax paid on earlier acquisitions from the tax on total acquisitions that one has received.

Present Methods of Taxing Property Transfers In spite of the Shoup Mission's strong support, the accession tax was enforced for only four years and was repealed in 1953. It was disintegrated into an inheritance tax and a gift tax, similar to the unmodified inheritance tax described above. One of the chief reasons for the change was the presence of serious deficiencies in administrative practices. The accessions tax was too sophisticated to be managed adequately with a poor level of tax administration. For instance, it was almost impossible for tax officials to determine total lifetime acquisitions, which was necessary for computing the tax base. Under a new system starting in 1953, heirs and donees paid their own taxes, based on the value of property left or transferred to them. As stated earlier, varying distributions among the same number of heirs could affect the tax payment to a substantial degree. This situation was considered undesirable, and stimulated the government to reform taxes on property transfers once more. In 1958 a hybrid system of an estate tax and an inheritance tax was instituted, and this system has remained basically intact up to the present time. This form of transfer taxation is unique, in comparison with the estate tax in the USA and the UK, or the inheritance tax in Germany and France. The main aim of combining the two taxes was to balance the tax burden on estates of the same size with the same number and types of heirs, even if the estate was distributed differently among the heirs. To begin with, let us explain briefly the present calculation method of the inheritance tax, separating the computation of the tax base from that of the tax amount in 1992 (see MOF Tax Bureau 1992; Gomi 1985). As indicated in Figure 12.1, the inheritance tax is levied on the gross estate acquired through inheritance or bequest minus non-taxable property,1 liability, and funeral expenses, that is, on the net estate. If an heir receives properties by gift from the decedent within three years of his death, the value of such property transfers is included in the tax base. Thereafter, a basic exemption of ¥48 million+ (¥9.5 million X number of statutory heirs) is deductible and the remaining property constitutes the amount of taxable inheritance. 1 In addition to the non-taxable treatment of life insurance, personal accident insurance, retirement, and similar allowances received by heirs, property transfers through inheritance or bequests to religious, charitable, educational, and scientific organizations are not subject to the inheritance tax.


Fig. 12.1

207

Computation of the Japanese tax base

Source: Data submitted to the Tax Advisory Commission.

Figure 12.2 summarizes the process of computing the tax amount. The amount of taxable inheritance (A in the figure) is divided among the statutory heirs (assuming wife, son, and daughter) in accordance with percentages prescribed in the Civil Code (i.e. -, -, and -) (B, C, D in the figure), irrespective of the actual shares of their inheritance.2 Then, tax rates are applied to each share computed in this way, result ing in the tax amounts (parts a, b, and c). Total inheritance tax liability is obtained by combining the shares of the individual heirs (part E). The total tax is then distributed among the heirs in proportion to the amounts they actually receive (parts d, e, and f). The inheritance tax is, however, increased additionally by 20 per cent for heirs other than spouses, children, and parents, although the total tax including this addition should not exceed 75 per cent of the heirs' actual share. Furthermore, there are several tax credits available to certain heirs. Of most importance is the credit for the spouse, computed as follows. If a surviving spouse receives properties, X can be credited against the inheritance tax (part d in the 2

Methods of dividing an estate among the statutory heirs are stipulated by the Civil Code as follows:

(1) spouse -, lineal descendants (children) —; (2) spouse -, lineal ascendants (parents) • ; (3) spouse4, brothers and sisters -; in the case of Figure 12.2, we assume two lineal descendants.

208


Fig. 12.2

Computation of taxes due

Source: see Figure 12.1.

figure) he/she actually pays, where

X=

total inheritance tax liability (part E)

(1) one-half of total value of taxable property (minimum ¥80m.) or (2) the amount of property actually received by the spouse x

Total value of taxable property


209

(With regard to (1) or (2), the smaller of these is applied.) The credit for the spouse enormously increases the amount of property that married couples may transfer free of tax. In fact, this method of credit usually provides complete exemption for transfers between husband and wife. (The appendix to this chapter contains a detailed description of how to calculate the relevant tax amount.) Likewise, three tax credits are available to other individuals under specific conditions. As mentioned above, properties received by gift from the decedent within three years of his death are included in the tax base, but the gift tax for such properties is credited against the inheritance tax to avoid double taxation. This is called the tax credit for gift tax paid. If an heir is under 20 years of age, his inheritance tax is reduced by ¥60000 for each of his years under 20. Similarly, if an heir is handicapped, his inheritance tax is decreased by ¥60000 (¥120000 in the case of the severely handicapped) for each year before he reaches 70. Next, the gift tax is levied on properties received in the calendar year with a basic annual exemption of ¥600 000. The value of the property is assessed on the basis of the current price or value at the time of acquisition, but several items are excluded from the tax base, e.g. properties acquired from relatives for living and educational expenses, and those used for religious, charitable, scientific, or other public purposes. In addition to the basic exemption, a lifetime allowance of ¥20 million is allowed for gifts of residential property to a spouse who has been married to the donor for over 20 years. Also, if a severely handicapped person becomes the beneficiary of money, securities, or other properties under the trust contract, a gift tax is not imposed on the beneficiary up to the value of ¥60 million. Rates, Exemption, and Credits Tax rates are applied to each share of the taxable inheritance (B, C, and D in Figure 12.2) or taxable gifts received by the donee. In Table 12.1, inheritance and gift tax rates are shown under the 1975-87 and 1998 tax law. Both rates begin at 10 per cent and rise to a maximum of 75 per cent in 1975-87 and 70 per cent in 1998, but the gift tax has much steeper progressivity than the inheritance tax. Indeed, the tax rate in 1998 reaches 70 per cent at ¥100 million under the gift tax, while it does so only at ¥2 000 million under the inheritance tax. The reason for the steeper progressivity of the tax on gifts is that it was substituted for the accession tax, which would have applied at a single set of rates to the cumulative gifts and bequests that a given individual receives during his life. These tax rates have not been adjusted frequently since 1958 when the new inheritance tax was adopted. At that time, the progressive rate scheduled started at 10 per cent on the first ¥0.3 million and rose to a maximum of 70 per cent above ¥100 million, with 13 brackets. Thereafter, there has been only a minor alteration in 1966. Similar adjustments have been made on the gift tax. In recent years the inheritance tax tends to be flatter in reducing the bracket number; that is to say, it was only 9 brackets in 1998 as against 13 in 1975-85.

210


Table 12.1 Tax rates (%)

10 15 20 25 30 35 40 45 50 55 60 65 70 75

Inheritance and gift tax rates, 1975-1987 and 1998 Taxable inheritance (¥m.)

Taxable gifts (¥m.)

1975-87

1998

1975-87

1998

under 2.0 2.0-5.0 5.0-9.0 9.0-15.0 15.0-23.0 23.0-33.0 33.0-48.0 48.0-70.0 70.0-100.0 100.0-140.0 140.0-180.0 180.0-250.0 250.0-500.0 500.0 and over

(under 8.0) (8.0-16.0) (16.0-30.0) (30.0-50.0) (50.0-100.0)

under 0.5 0.5-0.7

(under 1.5) (1.5-2.0) (2.0-2.5) (2.5-3.5) (3.5-4.5) (4.5-6.0) (6.0-8.0) (8.0-10.0) (10.0-15.0) (15.0-25.0) (25.0-40.0) (40.0-100.0) (100.0 and over)

(100.0-200.0) (200.0-400.0) (400.0-2 000.0) (2 000.0 and over)

0.7-1.0

1.0-1.4 1.4-2.0 2.0-2.8 2.8-4.0 4.0-5.5 5.5-8.0 8.0-13.0 13.0-20.0 20.0-35.0 35.0-70.0 70.0 and over

Source: National Tax Administration, Annual Report, 1987 and 1998.

In contrast, the basic exemption has been increased ten times since 19583 until now. A drastic change occurred in 1975 (the 1987 level), when the level of basic exemption was sharply increased from the modest level of 1973-74: i.e. from ¥6.0 million+ (¥1.2 million X number of statutory heirs) + a maximum of ¥6.0 million for the spouse to ¥20.0 million + (¥4.0 million X number of statutory heirs). Similarly, the basic exemption on the gift tax has been adjusted occasionally: ¥0.2 million in 1958, ¥0.4 million in 1964, and ¥0.6 million in 1975. Tax credits under the inheritance and gift taxes have also been increased on a few occasions. The basic exemption of inheritance tax was raised twice in 1992 and 1994, and at present it is ¥50.0 million + (¥10.0 million X statutory heirs). Until 1987, the number of decedents leaving estates with taxable value had steadily increased, mainly reflecting the absence of tax changes in more than a descade (see Table 12.2). The percentage of such decedents was only 3.4 per cent in 1970, but it had risen to 7.9 per cent in 1987, although there was a large drop between 1974 and 1975. In 1975, the basic exemption was raised drastically, causing a sharp decrease in the number of taxable decedents and in the amount of the inheritance tax relative to the value of all taxable property. Adjustments on the basic exemptions and tax rates under the inheritance and gift taxes are thus long overdue. The bubble boom clearly affected the value of taxable property and inheritance tax due. In fact, it is noteworthy that both substantially expanded around the period 1987—92. Since the collapse of the bubble, they have shrunk to a considerable extent. 3

In 1958, a basic exemption was allowed for ¥1.5m. + (¥0.3 X number of statutory heirs).

Table 12.2

Trends of tax payments in the inheritance tax, 1958-1996

Nos. deceased*

1958 1965 1970 1974 1975 1980 1982 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 a

The coverage of taxation

(B)/U) (%)

(A)

Objectives of taxation (= taxable decedents) (B)

684 189 700 438 712962 710510 702 275 722 801 711 883 740 247 752 283 750 620 751172 793 014 788 594 820 305 829 797 856 643 878 532 875 933 922 139 896211

5284 13407 24454 32898 14593 26797 35922 43012 48 111 51847 59008 36468 41655 48287 56554 54449 52877 45335 50729 48476

0.8 1.9 3.4 4.6 2.1 3.7 5.0 5.8 6.4 6.9 7.9 4.6 5.3 5.9 6.8 6.4 6.0 5.2 5.5 5.4

Total value of taxable property

Nos. of statutory heirs per decedent

Amount (¥b.)

Per decedent (¥000)

(C)

4.17 4.27 4.23 4.31 4.17 4.11 4.05 4.03 3.99 3.93 3.68 3.89 3.86 3.81 3.86 3.81 3.79 3.72 3.69

36.7 209.1 701.1 1 896.6 1 512.1 3021.5 4472.9 5428.7 6 246.3 6 763.7 8 250.9 9 638.0 11768.6 14105.8 17841.7 18820.1 16754.5 14545.4 15299.8 14077.4

Inheritance Tax amount (¥bn.)

Per decedent (¥000)

(D)/(Q (%)

884 3 111 5488 13303 13521 16417 17620 18620 19250 20 142 24307 42855 57448 61 149 70 112 62626 52514 46450 42835 39970

12.7 19.6 19.1 23.1 13.0 14.6 14.2 14.3 14.8 15.4 17.4 16.2 20.3 20.9 22.2 18.1 16.6 14.5 14.2 13.8

(D)

6345 15597 28670 57651 103618 112755 124518 126214 129831 130456 139826 264 286 282 526 292 124 315481 345 646 316858 320 843 301 599 290 399

4.7 41.0 134.2 437.7 197.3 439.9 633.0 776.9 926.1 1044.3 1434.3 1 562.9 2393.0 2952.7 3965.1 3409.9 2776.8 2105.8 2173.0 1937.6

Data based on Statistics of Population Census (Ministry of Health and Welfare). Other data are derived from Annual Report of the National Tax Administration. Source: Data submitted to the Tax Advisory Commission.

212


As is evident from the formula of calculating the inheritance tax, the greater the number of statutory heirs, the lighter the tax burden. In recent years, it has often been argued that some people try to avoid the inheritance tax by increasing the number of legally permissible statutory heirs before they die. Although the number of statutory heirs per decedent has declined in the long run (see Table 12.2), this type of tax avoidance is becoming more frequent. In fact, we find many cases in which individuals attempt to adopt grandchildren, nephews, and nieces as sons and daughters for tax purposes. The major part of the inheritance tax base consists of land, housing, business property, stocks and bonds, cash and deposits. Land, such as farming fields and residential sites, occupies the largest share—67.4 per cent of the total, followed by 14.4 per cent for bonds and stocks in 1990. Cash and deposits occupy 7.8 per cent of the tax base, housing 4.5 per cent, and business property 0.4 per cent.

THE ROLE OF I N H E R I T A N C E AND GIFT TAXES

Impact on the Tax System How significant have taxes on property transfers been in the tax system as a whole? The inheritance tax makes up only a small percentage of national tax revenues (see Table 12.3), but its relative share expanded from 0.5 per cent in 1950 to 3.9 per cent in 1998. Moreover, this share is the largest among major industrialized countries: in 1996-97 the corresponding figures were 2.0 per cent for the USA, 0.7 per cent for the UK, 0.6 per cent for Germany, and 2.4 per cent for France. The burden of the inheritance tax varies as the number of heirs differs. Two cases of the inheritance tax burden as measured by the value of taxable property are shown in Table 12.4. No tax is paid by the spouse because of the tax credit for spouses, and more children can make the tax burden lighter by increasing the amount of the basic exemption. For estates with taxable value of ¥3 billion, the inheritance tax amounts to 26.8 per cent of the taxable value when bequested to a spouse and one child, but to 23.0 per cent when turned over to a spouse and four children. It is difficult to judge whether the burden of the inheritance tax is actually heavier for a spouse with one child, as there are several other factors that affect the tax burden. The treatment of the gift tax seems to be generous, and it is hard to prevent tax avoidance on gifts from one living person to another. Likewise, land and closely held businesses are allowed to be greatly undervalued so as to lower the burden of the inheritance tax.4 4 Farm land was given special tax treatment under the inheritance tax until fiscal year 1991. An heir who receives agricultural land from a farmer through inheritance or bequest was allowed to exempt the difference between the market value of agricultural land and the present discounted value of agricultural income on land if he continues lo cultivate that land as a farmer for the next 20 years (see MOF Tax Bureau 1991, 132).


213

Table 12.3 Inheritance tax revenues, 1950-1998 Fiscal year

National tax revenues (¥bn.)

Inheritance tax" Amount (¥bn.)

1950 1955 1960 1965 1970 1975 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998

570.2 936.3 1801.0 3278.5 7773.2 14504.3 28368.8 30455.1 32003.1 34162.1 36774.8 39150.2 42851.0 42360.4 52193.8 57136.1 62779.8 62779.8 57396.4 57114.2 54000.7 54963.0 55226.1 57905.2 60345.1

2.7 5.6 12.3 44.0 139.1 310.4 440.5

552.1 664.5 786.1 877.3 1061.3 1369.6 1513.0 1830.9 2017.8 1918.0 2583.0 2746.2 2937.7 2669.9 2 690.3 2419.9 2395.0 2351.0

% of total 0.5 0.6 0.7 1.3 1.8 2.1 1.6 1.8 2.1 2.3 2.4 2.7 3.3 3.6 3.5 3.5 3.1 4.1 4.8 5.1 4.9 4.9 4.4 4.1 3.9

a

Including the gift tax. Source: National Tax Administration, Annual Reports.

As regards the gift tax, the tax burden is computed as the ratio of the tax paid to the amount of taxable gifts. For the past decade, the gift tax has averaged about 12 per cent of the taxable value. Redistributive Effects on Wealth In addition to the individual income tax, taxes on property transfers, such as inheritance and gift taxes, may influence the size distribution of wealth to some extent because of progressive tax rates. Despite their minor proportion in the total tax revenue (i.e. 3—4 per cent at the national government level every year), these taxes play an important role in redistributing economic power in connection with their goal of

Table 12.4 Value of

taxable property (¥m.)

(1)

The inheritance tax burden, 1998 Spouse's share (¥m.)

(2)

35 70 100 50 200 100 150 300 500 250 500 1000 1000 2000 1500 3000 2500 5000 Basic exemption

Spouse and 1 child

Spouse and 4 children

Tax paid by spouse (¥000) (3)

Tax paid by child (¥000) (4)

0 0 0 0 0 0 0 0 0

1850 14300 30800 72300 203 800 503 800 803 800 1 450 300

(3)+ (4) (1) (%) (5) 0 1.9 7.2 10.3 14.5 20.4 25.2 26.8 29.0

u

Notes: Columns (5), (8): ratio of total taxes to total taxable inheritance. Source: As Figure. 11.1.

¥70m.

(6)+ (7)

Tax paid by spouse (¥000) (6)

Taxes paid by children (¥000) (7)

(1) (%) (8)

0 0 0 0 0 0 0 0

0 0 7850 20000 52000 157000 414 500 689 500 1 304 500

0 0 3.9 6.7 10.4 15.7 20.7 23.0 26.1 ¥100m.


215

avoiding an excess concentration of wealth. Hence inheritance and gift taxes can be treated as a supplementary part of income tax in terms of redistributive taxation. Therefore, it is worth investigating the extent to which they affect wealth distribution. I shall use the same procedure on wealth distribution that I followed to determine the impact of the individual income tax, i.e. the Lorenz curve and Gini coefficients. For this estimation, it is essential that corresponding figures for income and tax yields be used. Necessary data are available from the Annual Report of the National Tax Administration. Let us consider the case of the inheritance tax. The available information on tax statistics poses some problems for the purpose of our study. For example, data on the wealth distribution of statutory heirs do not exist. Since taxes on property transfers take the form of an inheritance tax in Japan, it is necessary to investigate how the wealth of statutory heirs, obtained through inheritance and bequest, can be affected by taxation. Despite this requirement, no data are available on heirs, and only data on decedents are found in the tax statistics. Accordingly, we are compelled to use the figures concerning the wealth of decedents subject to the inheritance tax. The redistributive effect of the inheritance tax in this analysis is not related to how heirs inherit the estates in practice. If we ignore such an important limitation on the available statistics, we can obtain the number of taxpayers,3 the value of property, and the tax yields necessary for our estimations. The results are summarized in Table 12.5, in which the equalization coefficient

The Japanese Tax System, Third Edition

Japanese Maples, Third edition

Essential System Administration, Third Edition

An Introduction to Japanese Society, Third Edition

System Operation, Volume Volume L, Third Edition

Manipulative Therapy in Rehabilitation of the Locomotor System, Third Edition

Primer on the Autonomic Nervous System, Third Edition

Primer on the Autonomic Nervous System (Third Edition)

The Japanese Economy, 2nd Edition

The Rat Nervous System, Third Edition (Paxinos, The Rat Nervous System)

Fushigina Toshokan [Japanese Edition]

Japanese Culture, 4th Edition

The China Tax Guide, Fifth Edition

Japanese Culture, 4th Edition

Skin Immune System: Cutaneous Immunology and Clinical Immunodermatology, Third Edition

The Probabilistic Method (Third edition)

The Beekeeper's Handbook, Third Edition

The International Pharmacopoeia, Third Edition

The Universal Journalist, Third Edition

The debater's guide (Third Edition)

The Chef's Companion, Third Edition

Ancient Iraq: Third Edition

Molecular Medicine, Third Edition

Building Materials, Third Edition

Paradoxes, Third Edition

Chest Sonography, Third Edition

Gene Biotechnology, Third Edition

Classical Algebra, Third Edition

The Japanese Tax System, Third Edition

Japanese Maples, Third edition

Essential System Administration, Third Edition

An Introduction to Japanese Society, Third Edition

System Operation, Volume Volume L, Third Edition

Manipulative Therapy in Rehabilitation of the Locomotor System, Third Edition

Primer on the Autonomic Nervous System, Third Edition

Primer on the Autonomic Nervous System (Third Edition)

The Japanese Economy, 2nd Edition

The Rat Nervous System, Third Edition (Paxinos, The Rat Nervous System)

Fushigina Toshokan [Japanese Edition]

Japanese Culture, 4th Edition

The China Tax Guide, Fifth Edition

Japanese Culture, 4th Edition

Skin Immune System: Cutaneous Immunology and Clinical Immunodermatology, Third Edition

The Probabilistic Method (Third edition)

The Beekeeper's Handbook, Third Edition

The International Pharmacopoeia, Third Edition

The Universal Journalist, Third Edition

The debater's guide (Third Edition)

The Chef's Companion, Third Edition

Ancient Iraq: Third Edition

Molecular Medicine, Third Edition

Building Materials, Third Edition

Paradoxes, Third Edition

Chest Sonography, Third Edition

Gene Biotechnology, Third Edition

Classical Algebra, Third Edition

Recommend Documents