10647 A WORLD BANK COUNTRY STUDY Hungary Reform of Social Policy and Expenditures I l - ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~9 _ r A WORLD BANK COUNTRY STUDY Hungary Reform of Social Policy and Expenditures The World Bank Washington, D.C. Copyright e 1992 The International Bank for Reconstruction and Development/THE WORLD BANK 1818 H Street, N.W. Washington, D.C. 20433, U.S.A. All rights reserved Manufactured in the United States of America First printing April 1992 World Bank Country Studies are among the many reports originally prepared for internal use as part of the continuing analysis by the Bank of the economic and related conditions of its developing member countries and of its dialogues with the governments. Some of the reports are published in this series with the least possible delay for the use of governments and the academic, business and financial, and development communities. The typescript of this paper therefore has not been prepared in accordance with the procedures appropriate to formal printed texts, and the World Bank accepts no responsibility for errors. The World Bank does not guarantee the accuracy of the data included in this publication and accepts no responsibility whatsoever for any consequence of their use. Any maps that accompany the text have been prepared solely for the convenience of readers; the designations and presentation of material in them do not imply the expression of any opinion whatsoever on the part of the World Bank, its affiliates, or its Board or member countries concerning the legal status of any country, territory, city, or area or of the authorities thereof or concerning the delimitation of its boundaries or its national affiliation. The material in this publication is copyrighted. Requests for permission to reproduce portions of it should be sent to the Office of the Publisher at the address shown in the copyright notice above. The World Bank encourages dissemination of its work and will normally give permission promptly and, when the reproduction is for noncommercial purposes, without asking a fee. Permission to photocopy portions for dassroom use is not required, though notification of such use having been made will be appreciated. The complete backlist of publications from the World Bank is shown in the annual Index of Publications, which contains an alphabetical title list (with full ordering information) and indexes of subjects, authors, and countries and regions. The'latest edition is available free of charge from the Distribution Unit, Office of the Publisher, Department F, The World Bank, 1818 H Street, N.W., Washington, D.C. 20433, U.S.A., or from Publications, The World Bank, 66, avenue d'Iena, 75116 Paris, France. ISSN: 0253-2123 Library of Congress Cataloging-in-Publication Data Hungary: reform of social policy and expenditures. p. cm. - (A World Bank country study) Includes bibliographical references. ISBN 0-8213-2085-8 1. Social security-Hungary. 2. Hungary-Social policy. I. International Bank for Reconstruction and Development. II. Series. HD7172.5.H86 1992 92-6128 361.6'1'09439-dc2O CIP iii FISCAL YEAR January 1 to December 31 CURRENCY EQUIVALENTS (Forints per US$1.00) (period average) 1980 32.53 1981 34.31 1982 36.63 1983 42.67 1984 48.04 1985 50.12 1986 45.83 1987 46.97 1988 50.41 1989 59.10 1990 63.60 LIST OF ABBREVIATIONS CSO = Central Statistical Office GDP = Gross Domestic Product HBS = Household Budget Survey IBRD = International Bank for Reconstruction and Development ILO = International Labor Organization IMF = International Monetary Fund OECD = Organization for Economic Cooperation and Development PIT = Personal Income Tax SIF = Social Insurance Fund Ul = Unemployment Insurance v PREFACE The present Report was produced by a team of World Bank staff and consultants, in collaboration with a team of Hungarian experts, at the request of the Government of Hungary. The Hungarian researchers, under the leadership of Mihaly Kupa, produced the two empirical studies on the incidence of social benefits which underpin much of the analysis and recommendations of this Report. The World Bank team, led by Christine Kessides, visited Hungary on two missions, in April/May and October 1990. During these missions discussions were held with officials from the Ministry of Welfare (prior to May 1990, the Ministry of Health and Social Affairs), the Ministry of Finance (and the National Planning Office, which merged with it in May 1990), the National Social Insurance Administration, the Ministry of Labor (formerly the Wage and Labor Office), Central Statistical Office, Ministry of Interior, Ministry of Culture, and representatives of academia, local government, and social services. The Bank team wishes to express its appreciation to all of the Hungarian counterparts and collaborators for their assistance and insights shared during these discussions. The members of one or both of the missions who contributed to the present Report included Kenneth Davey (local government issues), Robert Holzmann (pensions and sick pay), John Micklewright (unemployment benefits, family support, maternity/child care benefits, social assistance), Andrew Smith (benefit administration), and Carlos Hinayon (comparative data and projections). Emmanuel Jimenez provided advice to both teams concerning methodology and interpretation of the incidence analysis. Frances Rosenthal was responsible for document production. The missions also benefited from analysis contained in an earlier study by the IMF Fiscal Affairs Department on social security reform in Hungary. Material in this report was updated during a mission in February 1991 to appraise the Second Structural Adjustment Loan. The report was discussed with Government officials in June 1991. I vii Table of Contents SUMMARY AND CONCLUSIONS xv PART ONE: BACKGROUND . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1. Introduction. .... .... 1 2. The Economic Context of Social Policy and Expenditures . . . . . 4 International Comparison of Hungarian Social Expenditures . . . . 4 Sources of Growth in Household Incomes - Implications for Income Distribution .... . . . . ...... . . . . . . . . . 16 The Profile of Poverty . ... . . . . . . . . . . . . . . . . . . 23 Demographic Prospects - the Demand for Social Policy . . . . . . 27 PART TWO: ASSESSMENT OF THE SOCIAL POLICY PROGRAMS AND RECOMMENDATIONS FOR REFORM . . . . . . . . . . . . . . . . . . . . . . . . 30 3. Overall Distributional Implications of Social Transfers . . . . . 30 Effectiveness of Social Transfers . . . . . . . . . . . . . 30 Efficiency of Social Transfers . . . . . . . . . . . . . . 34 4. Social Insurance Pensions and Sick Pay . . . . . . . . . . . . . 38 Introduction .38 Pension Expenditure and Eligibility Trends . . . . . . . . 39 Financial Scenarios and Retirement Reform Options .. . . . 41 Transitional Measures .46 Reform of Sick Pay . . . . . . . . . . . . . . . . . . . . 49 Conclusions and Recommendations . . . . . . . . . . . . . . 50 5. Support for the Unemployed .52 Introduction .......... 52 Unemployment in Hungary . . . . . . . . . . . . . . . . . . 53 The Former Scheme and Pressures for Reform . . . . . . . . 54 The New Unemployment Insurance . . . . . . . . . . . . . . 56 Evaluation of the New Unemployment Benefit System . . . . . 57 Administration, Costing, and Monitoring of Unemployment Support .62 Conclusions and Recommendations . . . . . . . . . . . . . . 63 6. Support for the Family .64 Introduction .64 The Incidence of Family Support and Child Poverty . . . . . 65 Three Possible Reforms of Family Allowance . . . . . . . . 70 Family Support and the Tax System . . . . . . . . . . . . . 74 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . 76 viii 7. Maternity and Child Care Support ..... . . . . . . . . . . . 78 Introduction ......... ... .. ... ... .. . 78 The Rationale for Child Care Payments in Hungary . . . . . 80 Suggestions for Reform ................. . 81 Reform of Nursery Care ...... . . .. . . .. . . . . 84 Conclusion and Recommendations . . . . . . . . . . . . . . 85 8. Social Assistance and Social Services . . . . . . . . . . . . . . 87 Introduction ......... ... .. ... ... .. . 87 The Problems of the Current Social Assistance Scheme . . . 88 Issues Relating to Reform . . . ..... . 91 Conclusions and Recommendations . . . . . . . . . . . . . . 95 Social Services and Other Social Income in Kind . . . . . . 96 9. Scholarships for Students . . . ...... . 99 10. Medium-Term Projections of Social Policy and Expenditures . . . . 102 The Baseline Projection . . . . . . . . . . . . . . . . . . . . . 104 An Alternative Scenario .... ................. . 107 ANNEXES I. Bibliography II. The Economic Context of Social Policy and Expenditures III. A Methodological Note on the Analysis of Incidence of Social Income IV. Social Insurance Pensions V. Unemployment Support VI. Support for the Family VII. Maternity and Child Care Support VIII. Social Assistance and Social Services IX. The Financing of Local Government X. Derivation of Medium-Term Projections GLOSSARY OF TERMS ix List of Text Tables Projected Macro Implications of Social Policy Reforms vii 2.1 Social Expenditure Trends in OECD Countries and Hungary, 1980, 85,86 7 2.2 Hungary: Social Expenditures, 1975-90 10 2.3 Net Social Insurance Tax and Transfers 13 2.4 Gini Coefficients for Components of Household Income 20 2.5 Age Distribution of Poor and Nonpoor Households 25 3.1 The Relative Importance of Different Kinds of Incomes, Taxes and Transfers in the Lower and Upper Deciles, 1989 31 3.2 The Ratio of Social Incomes in Kind to Net Personal Income in the Lower and in the Upper Deciles, 1989 32 3.3 The Ratio of Subsidies to Personal Net Income in the Lower and Upper Deciles, 1989 33 3.4 Gini Coefficients for Social Income in Cash 35 3.5 Gini Coefficients for Social Income in Kind and Subsidies 37 5.1 Number of Unemployed and Beneficiaries of Support Programs 55 6.1 Family Allowance Expenditure by Income Decile Shares, 1989 66 6.2 Proportion of All Households with Three or More Children in each Per Capita Income Decile, 1989 68 6.3 Poverty and Family Allowance, 1989 69 7.1 Child Care Benefits and the Distribution of Income, 1989 79 10.1 Macroeconomic Parameters for Medium-Term Projections 110 10.2 Summary of Medium-Term Projections 111 10.3 Summary of Medium-Term Projections - Alternative Scenario 112 List of Figures 2.1 Current Government Expenditure as Percentage of GDP 2.2 Total Tax Revenue as Percentage-of GDP 2.3 Government Health Expenditure as Percentage of GDP 2.4 Education Expenditure as Percentage of GDP 2.5 Pensions as Percentage of GDP 2.6 Growth Rates of Social Income and GDP 2.7 Structure of Household Income 2.8 Distribution of Labor Income by Type: Primary Income of Active Earners, Secondary Labor Income and Other Income by Deciles, 1989 2.9 Effect of Taxation: Distribution of Gross Personal Income, Taxes and Net Personal Income by Decile, 1989 2.10 Distribution of Social Income in Cash: Primary Income of Inactive Earners and Secondary Social Income by Deciles, 1989 2.11 Distribution of Net Personal Income, Social Income in Kind and Subsidies by Deciles, 1989 2.12 Hungary: Dependency Ratios Under Scenario II, 1949-2051 4.1 Hungary: Eligibility Ratio and Pension Expenditure Trends, 1960-90 4.2 Implicit Contribution Rate of Pension Program Under Alternative Retirement Scenarios 10.la Social Expenditure, Social Income and Net Personal Income as % of GDP (Base Case) 10.lb Social Expenditure, Social Income and Net Personal Income as % of GDP (Alternative Scenario) 10.2 Implicit Contribution Rates for Social Insurance 10.3a Cost of Labor as X of GDP and Social Income as % of NPI (Base Case) 10.3b Cost of Labor as Z of GDP and Social Income as Z of NPI (Alternative Scenario) xi 10.4 Social Expenditures as % of Total Government Expenditure II.1 National Health Expenditure as Percentage of GDP xii List of Annex Tables II.1 Age Structure of OECD Population and Hungary, 1980, 1990 II.2 Government Social Security and Welfare Transfers to Households II.3a Hungary: Structure of Social Expenditures (shares of household income and GDP), 1975-90 II.3b Hungary: Annual Growth of Social Expenditures, 1975-90 II.3c Hungary: Structure of Social Expenditures (percentage distribution), 1980-90 II.4 Comparative Health Expenditure, 1986 II.5 Elasticities with respect to GDP II.6 Components of Growth of Social Expenditures, 1981-89 II.7 Social Insurance Contribution Rates at the Income Level of an APW, 1987 II.8 Compensation of Employees as Percentage of GDP, 1980-88 II.9 Take-home Pay Plus Cash Transfers II.10 Personal Income Tax at the Income Level of an APW II.11 Change in Composition of Incomes of the Population II.12 The Growth of Labor Income and of Main Earnings II.13 The Dispersion of Income II.14 Share of Population Deciles from Different Kinds of Income and Taxes, 1989 II.15 Real Value of Pensions Established in 1980 II.16 The Composition of Members and of Income of Households with Different Number of Children (1989) II.17 Factors Determining Income of Poor and Nonpoor Households (1989) II.18 Indicators of Housing of Poor and Nonpoor Households (1989) xiii I1.19 Hungary - Demographic Assumptions, 1986-2051 II.20 Hungary: Age Distribution of Total Population Projections (Scenario II), 1986-2051 III.1 Share of Lower and Upper Deciles from Income, Taxes and Transfers, 1989 III.2 Social Incomes in Kind and Subsidies Ranked by Their Share in the Lower and in the Upper Deciles, 1989 IV.1 Hungary - Decomposition of Pension Expenditure Developments, 1960- 90 IV.2 Hungary: Assumptions for Baseline Policy Scenario IV.3 Eligibility Ratio and Effective Retirement Age V.1 Unemployment Benefit Scheme in Place from 1/1989 to 3/1/91 V.2 Unemployment Benefit Scheme Effective 3/1/91 VI.l Family Allowance: Measures of Importance, 1987-90 VI.2 Family Allowances for Two Children as X of Average Remuneration, 1980 VI.3 Rates of Family Allowance and Number of Recipients, 1990 VI.4 The Distribution of 1988 Marginal PIT Tax Rates on Earnings of Persons in Full-Time Jobs VI.5 The Distribution of 1990 Marginal PIT Tax Rates on Earnings of Persons in Full-Time Jobs VII.1 Maternity and Parental Leave in OECD Countries VII.2 Child Care Allowance and Fee: Recipients and Expenditure, 1987-90 VII.3 Nursery Usage (children aged 0-3), 1987-89 VII.4 Kindergarten Usage (children aged 3-6), 1987-89 VIII.1 Social Assistance: Expenditure and Numbers in Receipt, 1985-90 VIII.2 Social Services: Recipients of Selected Programmes, 1985-89 VIII.3 Households and Families in 1984 VIII.4 Subsistence Minima in 1989 VIII.5 Schedule of Copayment for Social Services for the Elderly, 1990 X.1 Base Case - Projection of Social Expenditures, 1989-2000 xiv X.2 Base Case - Projection of Household Income Aggregates, 1989-2000 X.3 Alternative Scenario - Projection of Social Expenditure, 1989-2000 X.4 Alternative Scenario - Projection of Household Income Aggregates, 1989-2000 xv SUMMARY AND CONCLUSIONS (i) The present study evaluates the overall social policy system of Hungary and proposes reforms aimed at strengthening its role as a defense against poverty, raising the quality and equity of social programs, ensuring financial sustainability of the system, and restoring incentives (in particular relating to demand and supply for labor) suitable to a market economy. In addressing these concerns, the study examines the social policy framework in the context of total household income and the financial relationships among households, enterprises, and government ("the distribution system"), and considers the implications of current and alternative social policies for distribution of income among households. (ii) "Social policy" in Hungary pertains to all benefits to households including: (a) "social income in cash" ( old age, survivors' and disability pensions, sick pay, family allowances, maternity and child care allowances, student scholarships, social assistance, and unemployment compensation), (b) "social income in kind" (health services and pharmaceutical subsidies, social services, education and training, other subsidized activities such as culture and sports, and labor market services), together with (c) housing and other consumer subsidies. Recent economic and sector work by the World Bank, conducted in collaboration with the Hungarian government, has examined in detail the components of education and training, employment services, health care, housing and other consumer subsidies.' To complement this previous work, the present Report analyzes in-depth social income in cash, while exploring the linkages with related issues in the other social programs. (iii) This overview chapter first summarizes the economic context of social policy, with reference also to patterns found in other market economies (Chapter 2 of the main report). The next section looks at how social policy expenditures and the major economic aggregates may develop in the future, as indicated by a simple projection analysis (Chapter 10). Following this broad overview, the discussion summarizes the conclusions and recommendations regarding the needs for reform of the individual social programs, as covered in detail in Chapters 3-9. The final part of this Summary discusses some of the major implementation issues addressed in various parts of the Report, in particular: (a) the new relationships between local and central government in I See IBRD, Hungary - The Transition to a Market Economy: Critical Human Resources Issues, Report No. 8665-HU, August 6, 1990; IBRD, Housing Policy Reform in Hungary, Report No. 9031-HU, November 1990; and IBRD, Hungary Health Services: Issues and ORtions for Reform, Working Paper, August 24, 1990. Analysis of consumer subsidies has been carried out in the context of preparation of the first World Bank Structural Adjustment Loan (approved May 1990). xvi the administration and financing of social policy; (b) the tasks of administering a new "selective" social policy, including problems of means- testing, information exchange, and monitoring of system performance; and (c) the linkages between tax policy and social benefit policy.2 The Economic Context (iv) Structure of social expenditure in the economy. From a macroeconomic perspective, the main structural issues of the social policy framework in Hungary are: (a) the relatively high level of national income and budgetary resources it absorbs - social policy expenditures in 1990 amounted to one-third of GDP and almost 60 percent of general government expenditure; (b) the growing dependence of households on social transfers (comprising half of disposable income in 1990), which dampens work incentives; (c) the financial nonviability of the pension program given the existing expenditure trends (this component alone accounting for almost 10 percent of GDP and 18 percent of the consolidated State Budget); and (d) the heavy burden of social insurance taxation on employers - at 43 percent of payroll, virtually the highest contribution rate in the world - which seriously constrains Hungary's international competitiveness and is inconsistent with the functioning of a modern market economy. Admittedly, Hungary in the past has not devoted a larger share of national income to social benefits (adjusted for comparable definitions of expenditure) than some OECD countries with a "social welfare" orientation. However, given the continued growth of social transfers in Hungary in the last few years and retrenchment of the same in many Western countries, it is clear that Hungary's share of GDP devoted to such expenditures is currently well above what would be expected for its income level, especially when housing and consumer subsidies are also taken into account. Moreover, the orientation of social policy in Hungary (tilting towards pensions) has been unsuitable to address the growing problems of poverty and unemployment or the human resource requirements of a competitive market economy, and will become increasingly unsustainable financially in the face of pressure from the rapid aging of the population. Whereas Hungary surpasses even the OECD "social welfare" states in family benefits, it has (until very recently) remained considerably behind in the share of national resources devoted to health, education, unemployment support, and social assistance. (v) Sources of demand for social policz - poverty and demographic trends. Evidence from household surveys during the past decade indicate that at a minimum, one-tenth of the population lives below the subsistence income level. The incidence of poverty has increased for urban active households, in particular those with three or more children. These and related findings suggest that policies to alleviate poverty should be designed to address the needs of families with children, to protect pensions against inflation, and to provide child care support to working mothers. Longer term approaches are also needed to improve education and training of the vulnerable groups. Rough 2 Technical details of the analysis and supporting data are provided in annexes to each of the main chapters. A glossary at the end of the Report defines the main terminology used. xvii estimates of the "poverty gap" suggest that the cost of raising the poor above the subsistence income level could be met with much less than the existing level of social spending, through better targeting. (vi) Unemployment is not yet sufficiently common to have a measurable effect on poverty, but it is increasing rapidly - already representing about 2 percent of the labor force in January 1991 and projected to reach 4-5 percent by year's end. Unemployment to date has hit hardest the lower income segments of the labor force (e.g. the unskilled), who have the most difficult prospects for reemployment. Planning for the next few years must take account of potential structural unemployment rates of at least 10-15 percent. (vii) Projections of population trends over the next several decades indicate that the share of the elderly in the population--the main factor determining pressures for social policy--will decline briefly during the 1990s and thereafter climb sharply through the middle of the next century. There will be a temporary surge in the working age group in the mid-1990s, which occurs at an unfortunate time for the labor market given the other factors likely to increase unemployment at this time. The conclusions from the demographic analysis are that: (a) Hungary must take advantage of the brief spell in the next few years to put in place reforms of pensions and health care which can be sustained financially during the decades of intense demographic pressure ahead; (b) support for the unemployed will have to assume a greater share of social policy in these next few years; and (c) the prospect of a sharply contracting labor force from 1996 implies a reduced tax base for social insurance, further underscoring the urgency of reforms to curb expenditure in this area. (viii) ComRosition and distribution of household income. The two most significant developments concerning household incomes during the past decade have been the change in the composition of labor income (growing importance of "secondary" employment in the private/quasiprivate sector), and the growth of social income. Secondary employment has markedly increased the dispersion of income, an effect which has been moderated slightly by the progressive income tax introduced in 1988. The potential positive effect of private employment on labor incentives has been dampened by the rather limited degree of industrial adjustment over the decade, which has continued to tie most workers to relatively unproductive jobs in State firms, and by households' increasing reliance on social income. Social income transfers in cash and in kind (excluding subsidies) have had a distinctly equalizing effect on income distribution. They have, therefore, contributed to providing a significant income supplement to the lower income groups, despite considerable leakage of benefits to upper income groups. This effect suggests that the priority for social policy reform should be to strengthen the effectiveness of social transfers in keeping vulnerable population groups above a socially-accepted minimum income level, through better targeting. The transfers to the middle strata have not been sufficient to counteract the erosion in their real incomes, and the best strategy to improve their level of welfare would consist of policies to strengthen economic growth, curb inflation, and ensure a flexible labor market. xviii (ix) A fundamental reduction of social expenditure in Hungary requires a transformation of the socialist "distribution system". Under this system, wages were deliberately suppressed and a share of legitimate labor income was channeled from enterprises through the State budget and back to workers in the form of cash benefits and social services. Reform of social policy therefore has a direct bearing on wage policy, even more than in market economies. It is often argued in Hungary that government policy for economic adjustment should ensure a direct and compensating increase in wages in parallel to any reduction in social income. This Report takes the position that government cannot and should not aim to liberalize wages in a direct attempt to compensate for reductions in social income. Wage gains should rather reflect the development of labor productivity and macroeconomic constraints, including the need for Hungary to achieve competitiveness in international markets. Sound economic policies to adjust to the demands of the market and ensure macroeconomic stability are the best approach to promote growth of incomes. (x) At the same time, it is clear that a major objective of cutting back social expenditure should be to reduce the burden on employers to finance such expenditure, thus permitting improved profitability and greater incentives to absorb labor. It is likely that workers bear most of the incidence of the employers' social insurance contribution, either through higher prices or through lower wages. It would therefore be expected that as this tax on labor is cut back through reform of social expenditure, wage payments to workers would rise to at least some degree and/or prices would fall so that households could realize a higher real income, while employers could enjoy some net improvement in their financial position. Perspectives on Reform over the Medium Term (xi) The key issues for the design of the social policy reform are: (a) how to improve the efficiency of existing social programs so that a higher level of welfare is gained from a given level of expenditure, and (b) how to target the reduced social expenditure to ensure that, while some individuals experience lower transfers, no one is put at risk of poverty or undue hardship. Both comparison with international practices, and analysis of the efficiency and effectiveness of social programs within Hungary, suggest that the main thrust of reform should be to tighten eligibility for benefits to improve targeting, and to increase the responsibility of individuals for their own welfare. Specifically, the priorities for reform include the following: (a) curtailing the growth of total pension outlays through tightened eligibility; (b) effecting structural changes in sick pay, family allowances, maternity/child care benefits, and housing subsidies to reduce their share of expenditures and improve targeting; and (c) increasing the level of social expenditures devoted to unemployment support and social assistance. In health and education, the priority should be improving efficiency in the expenditure of existing government resources allocated to the sectors.3 Once appropriate incentives for improved efficiency are put in place, a modest increase in the 3 References in the present Report to reforms in health, education or housing are based on the analysis and recommendations of the World Bank studies cited earlier, which are not recapitulated here. xix share of public spending for health and education could be justified to redress the past underinvestment in human resources. The Report also recommends that social policy for a modern market economy should entail more adequate burden-sharing between the private and public sector, both through continuing the significant level of copayment by beneficiaries of housing, education, and health, and by increased private provision of these as well as other social services, pensions, and student aid. (xii) The Report assumes that reforms to address these concerns will have to take place in a macroeconomic environment characterized by continued resource constraints and a need for structural adjustment during the next few years. Specifically, it is assumed that, after an initial period of declining or stagnant growth as economic reforms are intensified, the performance of GDP should rebound in the second half of the decade; unemployment will correspondingly rise sharply in the next few years and then settle towards the end of the period to the kind of "steady state" rate observed in many market economies; households' share of national income will remain steady or marginally decrease, while the share absorbed by the government sector will decline distinctly to allow a greater role for private enterprise. (xiii) In this context, the Report considers first the implications of continuing most of the existing programs of social insurance without structural reform. This Base Case demonstrates a trend of pension expenditure which would imply a sharply worsening financial deficit after year 2000--and an alternative assumption of lower GDP growth would require already in the 1990s an increase in the contribution rate to balance the pension program. Summary indicators from this and the alternative projection are given in the following Table. Even assuming some reform in health expenditure during the present decade, the Base Case indicates no possibility for the overall social insurance contribution rate to be reduced, but rather a further increase is needed to accommodate growing demand for unemployment insurance. Total social income as a share of GDP or of household income would not fall as recommended. The total cost of labor to employers would be likely to increase to finance the additional social insurance expenditure. In short, the financial burden of social policy would not be relieved, making it less likely that the economy would rebound and achieve the kind of performance desired. (xiv) The Alternative Scenario considers a major structural reform to curb expenditure on pensions, as well as partial reform of family benefits and a shifting of certain expenditure components out of social insurance (with most of sick pay charged to employers and maternity/child care benefits to the general budget). As in the Base Case, some expenditure increases are assumed in other programs which the analysis suggests are justified (unemployment insurance and labor market programs, social assistance, social services).4 This Alternative Scenario permits pension expenditure to decline by 1996 by about two percentage points of GDP, and the implicit contribution rate for pensions falls from 30 percent of payroll to 24 percent from 1990-96, returning to a maximum of 34 percent in year 2036 (versus 52 percent in the 4 Both scenarios assume the same significant cuts in housing and consumer subsidies. xx PROJECTED MACRO IMPLICATIONS OF SOCIAL POLICY REFORMS 1990 1991 1992 1993 1996 2000 BASE CASE (Low Reform Effort) (% of payroll) Social Insurance Contrib. Rate a/ 53 55 58 59 53 55 (paid by employer) 43 45 47 48 45 43 (% of GDP) Total Social Expenditure 33 32 34 34 28 27 as % of General Govt. Expend. 57 55 60 62 56 61 Total Social Income 28 28 31 32 27 26 as % of Net Personal Income 41 40 43 44 41 40 Labor Cost to Enterprise 63 66 66 66 63 63 (gross wages & social insurance) ALTERNATIVE SCENARIO (Higher Reform Effort) (% of payroll) Social Insurance Contrib. Rate b/ 53 55 52 52 43 43 (of which paid by employer) 43 45 38 37 26 25 (% of GDP) Total Social Expenditure 33 32 32 32 26 24 as % of General Govt. Expend. 57 55 57 59 51 53 Total Social Income 28 28 29 30 24 23 as % of Net Personal Income 41 40 42 43 39 37 Labor Cost to Enterprise 63 66 64 63 58 58 (gross wages & social insurance) BOTH SCENARIOS c/ GDP Real Growth Rate (%) -4.7 -1.3 -2.1 0.9 5.8 3.0 General Govt. Expenditure 58 58 57 55 51 45 (% of GDP) Unemployment Rate 2 5 10 12 6 3 (% of Labor Force) a/ Covering pensions, all sick pay, health care/pharmaceuticals, maternity/child care benefts, unemployment insurance, and administrative costs. t)f Same as above but excluding maternity/child care benefds and most sick pay. c/ For simplification, the macroeconomic assumptions are the same for both scenarios. xxi Base Case). The implicit contribution rate for all social insurance programs combined, including the new unemployment insurance introduced in 1991, would decline by 12 percentage points to 43 percent by 1996. The total cost of labor to employers, which is greatly affected by the social insurance contribution requirement, would fall by 8 percentage points (as a share of GDP) between 1991-96. The reduction of social income in total household income would be less dramatic (about 4 percentage points over the whole decade); if a larger compression of this share is needed to improve work incentives, social benefits would have to be cut further until growth in labor productivity allows for more rapid growth in wages. It should be noted that the Alternative Scenario does not include the effect of all of the reforms recommended below, certain of which (e.g., taxation of family allowances) would generate additional budgetary revenues. In addition, by changing incentives, some of the recommended reforms would reduce the demand for certain social expenditures (e.g., a lower level of child care benefit would reduce its use). (xv) The shares of social income and of social expenditure in GDP would each drop about 3 additional percentage points by 2000 compared to the Base Case. The share of social expenditures in total government expenditures would vary with the assumed speed of contraction of the government sector; the Alternative Scenario indicates that social spending would still represent more than half of the total government expenditure by 2000. This outcome is logical if activities which can be financed and managed privately are shed from the public sector and the social welfare function remains one of the government's main responsibilities. Assessment of Social Programs and Recommendations for Reform (xvi) The preceding paragraphs suggest that the structure as well as the size of social expenditures will have to be quite different in the future. The basic emphasis of social policy needs to change from providing income and social services directly to the entire population, to guaranteeing an effective minimum level of the same, while promoting the economic conditions for improved welfare and increased private participation in the financing and delivery of benefits. The reforms recommended in the individual social programs to achieve this fundamental change are summarized below. Estimates of the potential budgetary implications of these recommendations are provided in the respective chapters. (xvii) The following recommendations raise a further set of issues regarding trade-offs and interrelationships (as well as apparent inconsistencies) among the separate program reforms. The first of these trade-offs, and the most troublesome, concerns the problem of growing unemployment and the perceived need to increase exit from (and reduce entry to) the labor force at almost any cost. The Report argues that a viable pension program requires delaying retirement at full pension, and that disability benefits, early retirement provisions for older workers, and child care leave are not appropriate instruments to withdraw persons from active labor force status. Although this position may result in a higher level of open unemployment, the authorities should not attempt to suppress this outcome by measures to "buy off" workers with commitments to long term benefits which xxii can become financially unbearable. In addition to their financial cost, such measures are less economically efficient than letting unemployment "select out" workers in less productive activities and providing appropriate unemployment insurance, other forms of income support, and programs such as retraining to promote labor redeployment. (xviii) This leads to a second apparent trade-off between the pressure for increased expenditure on income support, especially social assistance and unemployment compensation, to protect individuals from undue hardship as a result of economic restructuring, and the need for overall reduction in social expenditure and promotion of work incentives. The Report argues that a well- designed and -administered program of social assistance would actually promote greater efficiency in other benefit instruments, in particular by making possible a more actuarially-based pension program and by permitting unemployment insurance to be of shorter duration to encourage the earliest possible return to work. (xix) Finally, numerous benefit programs are discussed in the Report which provide income support to families and the two main approaches, family allowances and mateinity/child care allowances, may appear redundant. While it is recognized that cash benefits (or conversely tax benefits) may have a higher value to recipients than subsidies which are specific to goods, services, or labor market status, the Report favors a variety of policy instruments supporting families--provided they are well targeted to those with lower income. For example, a reformed system of child care benefits and State support to nurseries is justified on the grounds that quality child care is a legitimate public good supporting labor market flexibility as well as private needs. (xx) Pensions and Sick Pay. The main deficiencies of the pension program are threefold: the system is financially unsustainable; despite the high outlays, it does not provide the required income support; and the system substantially distorts resource allocation. The Report recommends that: To secure the financial viability of the pension program into the middle of the next century, a tightening of eligibility through an increase in the effective (average) age of retirement is essential. This change would not necessarily require an increase in the legal minimum retirement age for both sexes, at least over the next one to two decades. A tightening of eligibility for disability pensions is equally necessary. Increasing the effective retirement age will require (a) harmonization of the legal minimum retirement ages for men and women, (b) introduction of a common and higher standard retirement age (when full pension benefits are granted), and (c) reform of the pension formula to induce later retirement as a deliberate decision of individuals. Transitional measures are needed to ensure the short term health of the social insurance account over the next 2-3 years before the impact of the above reforms is felt. The most important of these measures include flat-rate indexation of pensions (the Government has already proposed to xxiii introduce this, linked to growth in net wages), the reintroduction of retirement income testing, and the introduction of financial incentives to postpone retirement. Reform of sick pay should involve shifting a very large share of the costs to employers (e.g., the first 30 days of sick leave). The social insurance contribution covering the remaining periods of sick leave could be differentiated to recognize and reward efforts by different enterprises to reduce risks to health and safety. (xxi) Unemployment Support. Labor market programs in Hungary include both unemployment compensation and services in kind (e.g., retraining); the latter programs are not discussed here but in another recent report on human resources issues (World Bank, 1990a). Budgetary expenditure on both components combined has grown more than ten-fold between 1989-91 although the anticipated major increase in unemployment has not yet occurred. The conclusions of the present analysis are as follows: - The new unemployment insurance introduced in March 1991 is an improvement in many respects over the former benefit system; however, the maximum duration of unemployment insurance (2 years) could contribute to longer periods of unemployment by providing a disincentive to reenter the labor force, as well as being costly to maintain. The total period of unemployment insurance should not exceed 12 months. If a second year of benefit is retained, it should be a flat-rate payment less than the minimum wage, or be means-tested. An adequate benefit system for the unemployed should include a means- tested benefit of last resort, which could be achieved through reform of the existing social assistance scheme (para. xxv). This could - supplement unemployment insurance for those with family circumstances requiring special help as well as provide primary support when'insurance benefit is unavailable or has expired. - The treatment of part-time work and benefit for older workers need to be reconsidered, as the present schemes may discourage return to regular employment and be unnecessarily costly. Preferable to the new "pre- pension" would be a longer-duration, flat-rate unemployment benefit for older unemployed workers, possibly means-tested, to span the period to retirement age. (xxii) Support for the Familv. Family allowances are a very large item of budgetary expenditure in Hungary (accounting for 3 percent of GDP, or half of social cash income excluding pensions), and represent a major supplement to the income of poorer families. The large budgetary cost and the fact that the allowances are also received by high income families argues for their reassessment. However, it should be recognized that there are advantages of a universal family benefit (i.e., one delivered without ex ante regard to income). These advantages include the low administrative cost, lack of stigma, automatically high coverage ("take up") of the needy population, and avoidance of high marginal tax rates as occur at the cut-off level for means- tested benefits (the "poverty trap"). The challenge for reform of the family xxiv allowance is to improve targeting to lower income families while preserving as much as possible the advantages of a universal benefit. (xxiii) The Report evaluates a number of reforms of family allowance and tax policy and does not recommend a specific package, but urges that the following elements be included: retention of universal family benefit in some form, which is justified to provide some compensation for the private costs of raising children. If the present cash allowance is retained, it should be included in taxable income to improve targeting ex post on poorer families. An equally acceptable alternative would be to shift to a refundable tax credit which can be scaled back for higher incomes. abolition of the existing tax allowance for large families. If additional support to large (also generally the poorest) families is necessary, a larger cash allowance or refundable tax credit for them would be preferable, financed by cutting or eliminating the present allowance for one-child two-parent families (this would amount to greater differentiation of support by family size than presently exists). The family allowance system should in any case be buttressed by a means-tested benefit for poverty relief which is provided through social assistance (para. xxv). payment of the allowance in every case to the mother, except where she does not have custody, to better ensure that the benefit is spent for the child regardless of income-sharing between the parents. To enable a child tax credit to be received by the mother, independent taxation of husband and wife should be retained. (xxiv) Maternity and Child Care Support. The review of maternity and child care benefits in Hungary, together with day-care provision for children of nursery age, suggests that substantial reform should be undertaken here as well to make the system more suitable for an economy with private employers, while increasing support for lower income families and permitting a high level of participation by women in the work force. The Report recommends that a reform package contain some or all of the following elements: - reductions in the length of paid maternity leave and in the earnings replacement rate during maternity leave, both provisions being well above the levels found in OECD countries. - a switch from earnings-related to flat-rate support after maternity leave. - a curtailment of the total length of paid child care benefits (including maternity leave) and statutory unpaid leave to at most two years. - continued State financial support for the public day-care facilities for children of nursery age. xxv encouragement of employer and private nursery facilities via the tax system. (xxv) Social Assistance and Social Services. An essential component of the social safety net in Hungary is represented by social assistance and social services, both of which are largely administered and financed by local councils. It is increasingly likely that these programs will fall short of the actual need unless there is a major overhaul of the system, which was never designed to address the scope of demand expected in the next few years. The Report recommends: Social assistance should be an entitlement, as of right, for those individuals who can demonstrate need via a means test which takes full account of all dependents. Persons receiving social insurance benefits such as pension or unemployment insurance should be eligible for social assistance if their income including such benefits falls below the designated cut-off level. The level of entitlement should be determined in line with budgetary resources. The administration of social assistance and the means test should follow central guidelines to promote equity of treatment across the country. Procedures and criteria for benefits should be transparent, well- publicized, and backed by a formal appeals process. A fair intra-governmental distribution of the financial responsibility for social assistance needs to be devised (discussed below - para. xxx). The presently separate schemes for social assistance and child welfare support should be merged, at least for persons not in work, in the interest of simplifying administration; for persons in work with dependents, however, a supplementary means-tested child-benefit may need to be designed separately to address circumstances of very low earned incomes but with a concern to maintain work incentives. (xxvi) Social Services. The vast unmet need for social welfare services, particularly for the elderly, calls for a higher level of support from both central and local government as well as private resources. This support should not be a net addition to existing social expenditures but a reallocation to address urgent social needs which have gotten little public attention in the past. The Report recommends: Some cost-sharing of social service expenditures between central and local government should be arranged (para xxx). Social work and health services should be better integrated to provide more options for appropriate care for the elderly and follow-up of persons discharged from treatment centers. The central government should consider reinstating some subsidy of a school meals or milk program in collaboration with the local governments, as an effective means of delivering financial and nutritional support to the population most at risk of poverty. xxvi (xxvii) Student aid. The reform of student aid introduced in 1990 allocates all State scholarship funds to the universities on a per capita basis, with the institutions having full discretion in their use (e.g. merit- based or means-tested). If the universities use these funds to outbid each other for the best students, the result could be less access to higher education for the needy students. The Report recommends: - To achieve greater equity and efficiency in the use of student aid, consideration should be given to making all State scholarship assistance conditional on demonstrated financial need, with the funds awarded to the student not to the university. The universities could still administer the student aid program on the basis of minimum eligibility criteria determined by the central government. - An increasing share of State financial aid to students should take the form of loans and work-study arrangements. Issues for the Implementation of Social Policy Reform The roles of central and local governments (xxviii) In 1990 Hungary enacted a major decentralization of government with the passage of the Act on Local Government. The main issue raised by this decentralization which is relevant to the reform of social policy concerns the financing of the local councils, and in particular, the availability of resources for social assistance and social services. The overall level of funding for local government is part of the issue; in 1991, the budgetary transfers from the central government to the councils have been prepared with the assumption that local government as a whole would retain a constant share of GDP. If inflation exceeds the planned level in.1991, however, the councils could experience a real decrease in their resources. r.hat makes this prospect of particular concern is that many of the functions of local government, such as social assistance, are likely to require a real increase in funding to alleviate the hardships created during the transitional period of economic restructuring during the next few years; these functions are also not those that can be passed in major part to the private sector. The small size of many of the new councils would exacerbate this problem (almost half of the new local governments having populations under 2,000). An event such as the closure of one major employer could present a local council with an overwhelming demand for social assistance. The councils in poorer areas are also those with the least recourse to other revenues, such as taxation of local business. Moreover, the 1990 local government act removed the intermediate tier of government at the county level, which previously had redistributed resources among local councils to match needs with available resources. (xxix) The risk of underfunding of the local governments is of particular concern given that, already in 1990, there was evidence that many of the councils were unable to meet the requests for social assistance presented to them. The Report argues here that making social assistance a secure linchpin of the social safety net requires that the program become one of true entitlement, which essentially implies a secure source of funding. In most xxvii market economies, social assistance-type functions are not the financial responsibility of local governments alone, although they are administered locally; such activities normally require access to resources redistributed at some higher level of government. In 1991, the non-earmarked revenues transferred from the central budget to local governments, which represent 44 percent of planned local government revenues, are determined by capitation formulas reflecting primarily the demographic composition of the localities. The Report recommends that this revenue-sharing arrangement be revised so as to take into account as well the large economic disparities among settlements in terms of their local revenue base and incidence of poverty. (xxx) Several alternative financial mechanisms might be considered. At the more general level, the formulas used for revenue transfers to the local governments could simply be expanded to compensate for local economic conditions, based on regional indicators such as per-capita income or personal income tax revenue. The councils would remain free to use the funds at their discretion, as with the present formula grants. Another option would be block grants, which could be provided by the central government for certain specified activities such as housing rehabilitation, provision of social service facilities, or social assistance. Block grants preserve the freedom of local authorities to decide on the specific use of the funds, but within a general purpose agreed with the State. This approach might also be conceived as central-local cofinancing of particular activities, with the State's contribution determined during the annual budgetary negotiations. A cost- sharing (matching grant) arrangement might be offered with quite specific criteria for access to the funds--for example, State funding for nurseries could be made conditional on the introduction of a fee structure based on families' ability to pay.' (xxxi) These proposed funding arrangements should not be seen as a violation of the spirit of local autonomy. Rather, they would be a way of ensuring that certain essential social obligations of the local governments have a more secure source of funding and are met according to common minimum standards across the country. The Administration of a Selective Social Policy (xxxii) Targeting. The Hungarian authorities have expressed the intention to move towards a "selective" social policy system in the future, that is, less universal and more targeted provision of benefits with well-focused objectives. There are alternative ways of targeting, some of which have already been used in Hungary: (a) Targeting by geographic area - e.g., provision of a subsidized service only in designated low income districts. There is probably some scope 5 Similarly, cofinancing of a housing allowance scheme should require the local governments to agree to the introduction of market rents and to a common set of eligibility criteria for the housing allowance. See World Bank, 1990c and The Urban Institute, "Integrating State Rental Housing with the Private Market: Designing Housing Allowances for Hungary," November 1990. xxviii for this, although it is critical that subregional price and wage data be collected to indicate differences in the real standard of living across the country. (b) Selection by demographic characteristics - This is the targeting approach most used in Hungary, e.g. focusing on large families, single mothers, students, the handicapped and elderly over age 70. The evidence on the poverty profile, and the analysis of the actual incidence of social benefits by income group (Chapters 2 and 3), indicate that this emphasis has been reasonably effective in delivering benefits to low income recipients, but it is not sufficiently sensitive to use as the main allocation mechanism for all social programs. (c) Providing in-kind subsidies on goods and services consumed predominantly by the target groups (and removing subsidies on items consumed by the wealthier population) - The Incidence Study prepared for this Report provides evidence that the social transfers related to recreation, culture, housing, and higher education are the most concentrated on upper income groups, while programs directed at children and the elderly provide the largest supplement to the incomes of the poorer groups and are the most redistributive in their effect. This analysis provides some justification for introducing copayments in tertiary-level health services and higher education, and suggests that a meal subsidy in primary schools would be redistributive even if not means-tested. More study is needed of the use of specific services by various groups before this approach can be widely used. The evidence on the highly inequitable allocation of State rental housing indicates the main problem with this method of targeting, however - ensuring that the subsidized item stays with the target population.6 (d) Self-selection. The idea here is to take advantage of differences in preferences to design programs that give those in target groups a greater propensity to participate. A rather unfavorable example is provided by application procedures that are so time-consuming, complicated, and stigmatizing that only the most needy would participate--but such approaches would also achieve a low take-up by the deserving, and so would be self- defeating. A more relevant example from Hungary is that social assistance applications in some rural areas are decided by communal consensus. This practice may appear more acceptable than a faceless bureaucracy, but is impractical on a large scale as well as intimidating. (e) Means-testing. In principle, this is the most accurate targeting mechanism. Such means testing as has been tried in Hungary (for some student aid, for rent rebates in 1990, and for social assistance) is based only on income from main employment and some ad hoc inquiries about family circumstances. It is recognized that in the present culture in Hungary, with the prevalence of second economy earnings, the customary belief in privacy of 6 It is for this reason that the Bank's 1990 housing study (op. cit.) has recommended that subsidies for rental housing be provided directly to families on the basis of demonstrated need rather than to housing units which can be traded on a parallel. market. xxix financial assets and tax statements (curr'ently protected by law), and the large exemptions from personal income tax (PIT), no great reliance can be attached to income declarations. Many countries have faced similar problems yet have become accustomed to means-testing. Making access to financial benefits contingent on an investigation of personal finances is not inherently different than the approach used by any normal commercial bank to assess the creditworthiness of a loan applicant. Still, it must be acknowledged that there are administrative costs to means-testing, as well as a risk that it will miss the most deserving population. Therefore, it is essential that the appropriate approach to targeting be adopted for each type of social benefit. As argued below, an administratively simple and highly effective approach, where possible, would be to subject cash benefits to income taxation. (xxxiii) In designing a means-testing system for Hungary that could be used for a variety of social benefit programs, the following points should be considered: Applicants for some types of benefits could be required to present their income tax returns, or sign a waiver permitting a check of their returns, to certify their declared income.7 Alternatively, a change in the income tax law could permit such data to be accessed by the authorities to determine benefit eligibility. In this connection, the income tax authority (APEH) has proposed that full exchange of information be permitted between the tax files and social security files, to permit such verification. The present use in Hungary of a unique identification number for individuals for income taxation, social insurance, and unemployment benefit claims would allow cross-checking of income from such sources for purposes of means-testing. Personnel involved in means-testing should become trained in-objective procedures of income verification. Given the staffing constraints in the local governments, means-testing for locally-administered benefits should be conducted in collaboration with the social insurance offices. Significant penalties for false income declarations should be instituted; these could be enforced on the basis of random checks of some sample of applications. Recipients' eligibility for benefits should be recertified on a regular basis. Means-testing should not be so extensive that it presents a disincentive to employment by other family members. Related to this point, the unit of assessment should be the nuclear family (individual and spouse, together with any dependent children) rather than the "household" including other related or nonrelated adults; however, since this change may imply a higher rate of measured poverty, it will have an impact on 7 Proposals under consideration to reduce the exemptions from the PIT and broaden the tax base would increase the usefulness of tax returns for income verification. It must be recognized, however, that the timeframe of tax returns (historical annual income) is not relevant for the determination of current need for social assistance. xxx expenditure for social assistance and should be introduced as budgetary means permit. The current practice of taking account of relatives outside the household to determine eligibility for social assistance should be discontinued. The income ceiling for means-tested benefits should allow for benefits to be withdrawn gradually as incomes rise, to avoid high implicit marginal tax rates which would also discourage employment. To facilitate income and benefit determination at least for a transitional period, when the accuracy of data is especially uncertain, means-testing could be designed around several broad income categories and a scale of benefits assigned accordingly, with variations in amount linked to family size. (xxxiv) As a final point, for the first few years of operating a system of means-testing the authorities should devote at least as much attention to ensuring a high level of take-up as to guarding against access by the nondeserving. An agency of government--logically the Ministry of Welfare-- should be responsible for conducting an energetic campaign of public information to ensure that the needy population is fully aware of available benefits and is adequately served. (xxxv) Other issues of information. Two additional points are relevant here to the administration of social policy. First, the Report recommends that the Central Statistical Office (CSO) continue to calculate and publish objectively-defined subsistence income levels as a basis for determining the incidence of poverty. However, for the judgement regarding entitlement to means-tested benefits, the CSO minima should not be taken as given. As noted above, in principle the income levels used for means-testing should be based on the nuclear family and exclude housing costs (which should-rather be assessed separately, given the great variety of housing conditions). (xxxvi) Second, the effective design and administration of social policy requires regular monitoring of data on both demand for and use of benefits. Much of the analysis of the present Report would have been impossible without the micro data used for the Incidence Studies. In the present period of major changes in the economic conditions underlying social policy, it would be extremely useful for the Household Budget Survey to be made annually and the Income Survey at least every 3 years, rather than every 2 and 5 years, respectively. The Report also recommends that administrative data, for example those collected routinely by the Employment Offices, be used for analysis of the structure and incidence of unemployment, as well as the performance of the benefit system. In addition, sample survey data, such as from the Labor Force Survey initiated in January 1991, should be acquired regularly to determine trends in nonregistered unemployment, sources of income of the unemployed (important to assess the adequacy of unemployment benefits and their disincentive effects), and the actual use of social services in kind. xxxi Linkages between tax and benefit policy (xxxvii) As a general point, this Report argues for integrating social cash benefits into the income tax base, mainly in the interest of removing distortions of incentives and increasing equity rather than as a way of enlarging tax revenues. As mentioned, taxation of benefits would preserve the advantages of universal provision while ensuring the desired targeting. However, for taxation to function as a practical "means-test", it is necessary that the tax structure be effectively progressive - that is, with a wide distribution of tax payers across the tax brackets. The available information on the distribution of actual earnings in 1988 and 1990 indicates that a very large number of tax payers are paying the same rate--implying that the effective (rather than notional) progressivity is low. As Hungary reduces reliance on cash benefits as an instrument of social policy, greater consideration should be given to exploiting the redistributive potential of income taxation. (xxxviii) The recommendation is also made that pensions be included in the personal income tax base, while pension contributions should be deducted from taxable income for those currently employed. This change is desirable to reduce the present incentive for individuals to retire at the earliest possible age, and would make a smoother transition between employment and retirement (whether full- or part-time). The measure would likely have a net expenditure effect, which would be justified by the rationalization of incentives. Concerning taxation of family benefits, the Report notes that the revenue effects and the extent of targeting implied would depend on whether the allowance is received by the mother or father, who may face different tax rates under individual income taxation. Paying the allowance to the mother is recommended to increase the effectiveness of the benefit as income support to children, even though this would probably imply lower tax revenue, (xxxix) The Report recommends the use of refundable tax credits as an alternative to the present tax allowances for large families, which are regressive. It is noted that introducing a general system of refundable tax credits (negative income tax) would also be an approach to providing a minimum safety net even to non-taxpayers. However, such an instrument could not cover the variety of circumstances which can be addressed in means-tested schemes and would require much higher marginal tax rates than those presently in place. Finally, the Report notes that the role of nongovernment providers of social services (such as child care and elder care facilities) could be encouraged substantially through the tax system, as is the case in many market economies. (xl) In conclusion, the Report presents a range of recommendations for reform of social policy which aim to achieve greater equity, better protection from poverty, and a more appropriate framework of incentives, while reducing the burden on public finances. Some of the measures also imply increased expenditure or reduced revenues for the government but are justified to address areas of increasing social need. For reference, the following matrix provides a full listing of the main recommendations for reform, both those arising from the analysis of this Report and from the related Bank studies cited earlier. xxxii (xli) It must be acknowledged that these reforms will be difficult for any government to introduce, since they involve reduced benefits for some population groups in the short term, even though in the medium to longer term the reforms will contribute to a more dynamic market economy. Sustained implementation of the recommended reforms will only be possible on the basis of a national concensus on the need for such changes, which is further grounded in an understanding by the Hungarian people of the issues and stakes involved. It is hoped that the present report can contribute to this understanding. SUMMARY OF RECOMMENDED REFORMS OF SOCIAL POLICY Measures Which Reduce Government Expenditure Raise Revenues Increase MeasudesWhic R Ret_%ri,c[ Cut leccease User Privatize Teax Reduce ,gtibir ity/ Benefits tttClency Cnarges/ Services Benefits Tax. arget (Reduce Uaste) no ts xpenditures txpeni tures C ,nnta xze Financing) Pensions Raise eilt on Revise pension Encourage private pensions. Tax Gross-up pension. Exemptpeosion effective Denetlis (and toromula t,o remove pension. contribution retirement.age, contributions). disincentives to trom PIT. iphgh,n gClteria aPor spy tor Isa Ility Dension. introduce retirement income test. Sick Pay Reduce overuse. Make "entoyces Shift major ear hI-st few costoto dys. empoyers. Health and Reward savings Legalize co- Privatize some Pharma- in uSe ot drugs, payments. services. ceuticals services. Unemplo ment Reduce benefits Shift to flat- S t re nd Us Imen or art-tim tn t in tucrnncronctiv Support tor rattm, Wae beneTit i rnngtya,osuprt worters early nase I1 ot private sector. r e ret remeknt- Vnemployment norten total insurance. X MaternitY Reuce materg E1pEtLyer to H- Chflo!UCr. alt Iance stt Ia trst,month H Benerits t.t rt maternity . care eave. inate fd year. Faity Differentiate tl owanceby Tb famity R,e,ve 3rd Ryplace caph Attouance niter o chit rn c means- atowance. ch .d tax a 1gtnfy Dhi test portion of Meieit. aMiowance. re ared itd Sociat Define Unify social rovido tenef1t Assistance nti 'de apistanf? and entitement c I d awefare p incipre. cr, erMat ahd SDct tO means-test. :dhTA rtn Nurserjes/ Introduce Promote Introduce tax Socit lee sc4Le zriyate lnQentives tor Services ~a I to socics private nurseries. Educationi Define.central Copayment for Privatize soame Sbot ar- auide ines.for sohe education education and siips ikans-testing. Services and training. s;g Sta$e schotrshs to students directlyntrouce studnt tsaon__inan iat need. toans/work-stuay. S"hgLol Meall partia VsupDsi?y in Subsidy R~rysce oo n iMeergartens. ConsuTer and Eliminate Raise rents in p,ubli,ly-owned Commerciql banks Hake housini benefits introduce targeted Housin9 geneCat tlats to market IeveLs. tointfoluce tr1p firms to ept oyees nousing alIowances IuDslales 1?dies on fLexible mort- subject to PIT or cor- to re lace inerest rates age istruments porate profit tax. renret'subsidies. Igoods. ertSt rat P ta IC ?IsOUS1 ra mret prce PART ONE: BACKGROUND CHAPTER 1 INTRODUCTION 1.01 The present study of the social policy and distribution system in Hungary is intended to address four concerns. The first, and most urgent, is to develop a system of social support which can raise from poverty those individuals - approximately one-tenth of the population - who are estimated to be in this state and to ensure that others do not fall into poverty, particularly as a result of the changes brought about by economic restructuring. Secondly, there is a growing consensus in Hungary that fundamental reform is necessary to improve the quality and equity of basic social programs for social security, housing, health care, and education. Thirdly, it has become evident that the existing social policy has been sustained by a level of public expenditure that is increasingly unaffordable, given other pressures on the government budget, and that social insurance as presently designed is financially unsound. Finally, reform of social policy is seen as part of an overall reduction in the role of the State which is necessary to develop the structure of incentives consistent with a modern market economy. This last point implies the creation of new financial relationships among the central and local government, enterprises, and the population. 1.02 "Social policy" in Hungary pertains to all benefits to households including: (i) "social income in cash" (old age, survivors' and disability pensions, sick pay, family allowances, maternity and child care allowances, student scholarships, social assistance, and unemployment compensation), (ii) "social income in kind" (health services and pharmaceutical subsidies, social services, education and training, other subsidized activities such as culture and sports, and labor market services), together with (iii) housing and other consumer subsidies. Recent economic and sector work by the World Bank, conducted in collaboration with the Hungarian government, has examined in detail the components of education and training, employment services, the health sector, and housing subsidies.1 To complement this previous work, the present Report analyzes in-depth social income transfers in cash, while exploring the linkages with related issues in the other social programs. 1.03 The present social policy in Hungary is an artifact of the traditional distribution system of the socialist economy. Under this system, labor received a relatively small share of income from production in direct wages, while the State withdrew a large share of income from the enterprises and distributed it back to the workers through subsidized goods, publicly- 1 See IBRD, Hungarv - The Transition to a Market Economy: Critical Human Resources Issues, Report No. 8665-HU, August 6, 1990; IBRD, Housing Policy Reform in Hungary, Report No. 9031-HU, November 1990; and IBRD, Hungary Health Services: Issues and Options for Reform, Working Paper, August 24, 1990. Analysis of consumer subsidies has been carried out in the context of preparation of the World Bank Structural Adjustment Loan (approved May 1990). 2 funded social services and cash transfers. Full employment and suppression of earnings differentials was intended to create equity in income and welfare, so that social policy was not designed to redistribute among population groups or to provide a residual protection of incomes except in rare cases of hardship. Although the State's monopoly of production and of the labor market has since disappeared in Hungary, workers' dependence on the State for much of their income and consumption has remained. 1.04 In the ongoing transition from socialism to a full-fledged market economy, Hungarians have to create a new "social contract". This Report argues that the basic emphasis of the social policy system needs to change from that of providing directly income and social services to the entire population, to guaranteeing an effective minimum level of the same, while promoting the economic conditions for improved welfare and increased private participation in the financing and delivery of benefits. In the short term, the priorities for reform would be to develop a secure safety net for the poor, the near-poor and unemployed; to reduce budgetary expenditures for social policy, as part of the effort of macroeconomic stabilization and adjustment; and to remove the most obvious disincentives to economic growth inherent in the existing system. Although these objectives could in theory conflict - e.g., increased social assistance raising expenditure, unemployment compensation potentially reducing work incentives - this Report points to ample scope for changes in the short term which would serve these objectives consistently. 1.05 The pace and extent of change in the social system over the medium to long term will depend on the growth trend of the economy, the liberalization and distributional changes in labor incomes, and the development of the private sector. Over this longer term perspective, Hungary could choose to evolve towards any of the several systems of welfare provision which exist in modern market economies and which vary, for example, in the degree of private sector financing and provision of social services and in the use of the tax system as an instrument of redistributive social policy. This report does not present a blueprint of the ultimate social policy system for Hungary, but outlines the changes and choices which will have to be made in the near to medium term if Hungary is to compete successfully with other market economies. 1.06 The performance of the existing social policy system, and the alternatives considered here, are evaluated according to four broad criteria. The effectiveness of each social program is determined by the success in meeting its objectives, be they income replacement and the maintenance of a reasonable standard of living (for social security programs, such as old-age pensions and sick pay); alleviation of poverty (social assistance); or the provision of a satisfactory level of social service (e.g., health care and housing). Most of the social policy programs in Hungary are intended to serve too many objectives, which tends to weaken their effectiveness as measured against any one of their aims. In brief, the criterion here is how well the program is serving its intended population. 3 1.07 Taking into account as well the cost of a given benefit gives an indication of the internal efficiency of each social program. One specific measure of efficiency is the degree of wastage ("leakage") of benefits outside of the target population. To the extent that social policy is intended to achieve income redistribution, efficiency can also be measured by the success in achieving this objective. Where a social program is not targeted but universal, the issue of efficiency is somewhat different and can be summarized as the total (net of administration) cost of this approach as compared to a narrower, more focused expenditure of resources to serve the same general purpose. 1.08 The affordability of social programs is the next question; given the impact on public finances, are the programs sustainable? Where the programs are financed by earmarked contributions, the issues are whether the programs are financially viable at the given contribution rates, and whether the implied taxation poses an undesirable trade-off with other objectives such as maintenance of employment and international competitiveness. A further consideration in evaluating alternatives to particular social programs is the degree of burden-sharing involved and the financial implications for the population, in terms of their ability to share costs and to sustain reduced benefits. 1.09 The final issue, already alluded to above, concerns the incentive effects of the social programs, in particular regarding the supply of labor (incentive to work), demand for labor (cost to employer of hiring), and incentives to individuals to save. These effects determine the allocative efficiency of a given social policy, and will become increasingly important as the economy begins to operate more on market principles. 1.10 The institutional context of social policy is already undergoing considerable structural change through, first, the relative shift within the public sector from central to local administration and financing; second, the increased involvement of private voluntary organizations (and the legal recognition of private for-profit providers of some services, such as health care); and third, the reduction of workers' reliance on one employer for both earnings and delivery of many social benefits. The Report gives attention to issues for the administration of social policy which arise from these trends and from the increasing need for support to the unemployed and alleviation of poverty. These issues include in particular the practical and informational problems of determining financial need, the links between tax and benefit policy instruments, and the potential roles of private sector enterprises in the future social policy system. 1.11 Chapter 2 provides a background on the macro and microeconomic context of the social policy system in Hungary. Chapters 3 through 9 assess the performance of each of the existing social programs and argue the case for specific alternatives. Chapter 10 attempts to summarize the implications of these alternatives in terms of the future structure and growth of social expenditures and of household incomes. 4 CHAPTER 2 THE ECONOMIC CONTEXT OF SOCIAL POLICY AND EXPENDITURES 2.01 The first section of this Chapter reviews the size and structure of social expenditures in relation to those of market economies to which Hungary is frequently compared. Section B looks at trends in both labor and social incomes and their impact on income distribution. Section C considers the nature of existing poverty and its measurement as a basis for understanding the "demand" for social policy, while Section D presents the demographic prospects underlying this demand in the future. The analysis is provided in more detail, with the supporting data, in Annex II; the text below presents only the main observations and conclusions. A. International Comparison of Hungarian Social Expenditures 2.02 International comparisons of social expenditure are notoriously difficult to make and to interpret, as the social welfare systems of the developed market economies are very diverse and the demands on them represent many factors such as demographics, economic structure, and level of national income. Broadly speaking, the social welfare systems found in the West can be categorized as three broad types: (i) the Continental European (Bismarck) model of compulsory, employment-related social insurance (best represented by Austria and Germany); (ii) the social democratic state, featuring universal benefits financed by general taxation (typified by Sweden); and (iii) the "modern liberal state" (most notably, the United States) characterized by a residual role of government in social welfare, considerable private sector provision, heavy reliance on means-testing of transfers, and use of the income tax system for social support through tax reliefs and tax allowances.' Most countries, including Hungary, are a hybrid of these models. Hungary has compulsory social insurance covering old-age, disability and survivors' pensions, sick pay, and maternity and child care allowances, health care and pharmaceuticals (as of 1990) and unemployment compensation (as of 1991). Health care and employment services also obtain financing from general tax revenues to cover noncontributors. General tax revenues finance universal family benefits and other social benefits in kind; and a few minor transfer programs (social assistance and some student aid) are means-tested. In recent years, even the social insurance benefits in Hungary have been extended to noncontributors through a generous definition of "time in service" - hence, the social policy has tended increasingly towards universality. 2.03 In relative size and scope, Hungary's social policy provision increasingly resembles that of the "social welfare" states of Western Europe (considered here to include Belgium, Denmark, Finland, France, the Netherlands, Norway, and Sweden). Since it has been theorized that the share of national resources devoted to social benefits increases with the level of income, it is also interesting to contrast Hungary's social policy situation 1 An analytical description and evaluation of these models is provided in Esping-Andersen and Micklewright (1990). 5 with that of the "lower income" group of countries in the OECD (Greece, Ireland, Portugal, Spain, and Turkey) which are nearer to Hungary's level of development.2 Making these international comparisons is not to imply that there is a "correct" model of social welfare policy, and Hungary's position relative to other countries on various indicators does not in itself suggest any value judgements about the country's current social system. Since Hungary is evolving towards a market economy, however, it is useful to situate her current system within the diversity of social welfare policies of the OECD countries. 1. Size and Structure of Social Expenditures 2.04 To gain a proper perspective on the importance of social expenditures in Hungary relative to Western countries, it is useful to examine the overall size of the government sector in each case. As illustrated in Figures 2.1 and 2.2, government intervention in the Hungarian economy (both in terms of budgetary expenditure and tax withdrawals) has been in a different league from that of other West European countries. 2.05 Despite this relatively high level of government activity, Hungary has not devoted a larger share of national income to social benefits than comparator countries when measured by strictly equivalent definitions. Comparing to OECD data on "government social expenditure", and adjusting Hungarian figures for differences in definition, Hungary's social income transfers relative to GDP in the mid-1980s were almost equivalent to those of the Lower Income States, but well below those of the Social Welfare States (Table 2.1).3 The composition of this expenditure is not the same across the 2 It would also be useful to compare Hungary's social expenditures with those of Latin American countries which are closest to Hungary's per capita income level (e.g., Argentina, Brazil, Panama, Trinidad/Tobago, Uruguay, and Venezuela). This has not been done in this report because of difficulties in obtaining comparable data bases. 3 The OECD definition of "government social expenditure" includes public transfers to households in cash (social security pensions, unemployment compensation, sickness benefits, family benefits, other social assistance), and publicly-financed goods and services consumed by households (mainly education and health care). This definition is comparable to the Hungarian accounting of "social income in cash and in kind", except that Hungarian social income includes certain items not recognized by the OECD. In Table 2.1, about three percentage points of Hungary's "other social expenditures" represent social welfare services such as day care, and cultural and sports services, which are not included in the OECD definition and must be deducted from the total to derive strictly comparable total expenditure figures. Many of the "other social services" are provided by State enterprises in Hungary but included in social expenditure data; estimated items amount to about 3 percent of total social income. Hungarian data shown here have not been adjusted to remove these items, since for some OECD countries as well, a small amount of private expenditure (e.g., for noncompulsory pensions and sickness benefits) are included in the reported data for government social expenditures. 6 Figure 2.1 Current Government Expenditure as Percentage of GOP Prcentage 60 55 - s o _ . --"''------------------------- ------ --'---- ----- --------- OECD Social Welfare Stales 45 - OECD Lower Income Slsats 40 - 35 Sources: OECD: Natio,al Accounts, 1975-1588. Detailed Tables. Hungary, IMF Recent Economic Developm.nle. 30 I I I I 1980 1981 1982 1983 1984 1985 1986 1987 1988 Figure 2.2 Total Tax Revenues as Percentage of GDP Percentage 55- s0 OECD S.ollW., ^rssl 45 __ 40 36 OECD LsIcmSs , ................ 30 26 1980 1981 1982 1983 1984 1985 1986 1987 1988 Sourc. trnua S .sIlIloc orCO .Dmb., CoftrI.l-1005-87 HurcOr. IMP Raest economIc OoorSt.. Table 2.1: Social Expenditure Trends in OECD Countries and Hungary, 1980, 85, 86 Govermnent Social Expenditure as X of GDP Total Education Health Pensions Unemployment Other Social Expenditure 1980 1985 1986 : 1980 1985 1986 1980 1985 1986 1980 1985 1986 1980 1985 1986 1980 1985 1986 AustraLia 17.3 18.4 18.4 : 5.9 5.7 5.7 : 4.1 4.9 5.3 : 4.9 4.9 4.3 : 0.7 1.3 1.3 : 1.7 1.6 1.8 Austria 26.0 28.8 28.8 * 3.8 4.4 4.4 : 5.4 5.3 5.3 : 13.5 14.5 14.6 : 0.4 0.8 0.8 : 2.9 3.8 3.7 BeLgium 33.9 35.8 34.9 7.7 7.3 7.3 : 5.3 5.5 5.5 : 11.9 12.0 : 2.8 6.0 * 6.2 23.0 4.1 Canada 19.5 22.6 22.3 5.9 5.9 5.6 : 5.5 6.4 6.5 : 4.4 5.4 5.5 : 2.3 3.3 3.2 1.4 1.6 1.5 Denmark 35.1 33.9 33.9 : 7.8 7.2 7.2 : 5.8 5.2 5.2 : 9.1 8.5 9.3 : 3.0 3.2 3.2 9.4 9.8 9.0 Finland 22.9 22.8 23.8 : 6.0 6.2 6.3 : 5.0 5.6 5.8 : 6.5 7.1 6.3 : 0.4 0.5 0.8 5.0 3.4 4.6 France 30.9 34.2 34.2 : 5.7 6.1 6.1 : 6.0 6.8 6.7 : 11.5 12.7 12.5 : 1.6 2.8 2.8 : 6.1 5.8 6.1 Germany 26.6 25.8 25.2 : 5.1 4.4 4.5 : 6.3 6.4 6.3 : 12.1 11.8 11.4 : 0.9 1.5 1.4 * 2.2 1.7 1.6 Greece 12.6 19.5 19.5 : 2.2 3.3 3.3 : 3.6 4.1 3.7 : 5.8 10.7 10.6 : 0.3 0.4 0.5 : 0.7 1.0 1.4 Iceland : 3.7 : 5.7 6.4 6.9 : 0.2 Ireland 23.8 25.6 26.1 : 6.0 6.4 6.4 : 8.0 6.9 7.0 : 4.5 5.4 5.6 : 1.9 3.6 3.7 3.4 3.3 3.4 Italy 23.7 26.7 26.4 : 5.6 5.9 4.9 : 5.6 5.4 5.2 : 12.0 12.6 12.2 : 0.5 0.8 0.6 0.0 2.0 3.5 Japan 16.1 16.2 16.0 : 4.9 4.3 4.2 : 4.6 4.8 4.8 : 4.4 5.3 5.1 : 0.4 0.4 0.4 1.8 1.4 1.5 Luxembourg : 6.2 * Netherlands 31.8 30.7 30.7 : 7.2 5.6 5.4 : 6.5 6.5 6.6 : 11.0 10.5 10.5 : 1.7 3.3 3.3 5.4 4.8 4.9 New Zealand 22.4 19.8 21.0 : 5.8 4.1 4.1 : 6.0 4.4 6.1 : 7.6 8.1 8.2 : 0.5 0.6 0.9 : 2.5 2.6 1.7 Norway 24.2 23.5 24.8 : 6.2 5.6 6.1 : 6.5 6.2 6.6 : 7.9 8.0 8.8 : 0.2 0.7 0.4 : 3.4 3.0 2.9 Portugal 17.3 : 4.4 4.4 : 4.3 4.0 4.0 : 6.1 7.2 7.2 : 0.4 0.3 0.4 * 2.1 Spain 15.6 15.2 17.0 : 2.8 2.2 1.5 : 4.3 4.3 4.3 : 7.3 8.6 7.6 : 2.2 2.1 2.5 . 1.1 Sweden 33.2 32.0 32.0 : 6.5 5.9 5.9 : 8.8 8.5 8.3 : 10.9 11.2 11.4 : 0.4 0.7 0.8 : 6.6 5.7 5.6 Switzerland 19.1 20.5 20.5 : 5.5 5.5 5.3 : 4.7 5.4 5.2 : 8.0 8.1 8.5 : 0.1 0.3 0.3 : 0.8 1.2 1.2 Turkey United Kingdom 20.0 20.9 20.6 : 5.5 5.0 5.1 : 5.2 5.2 5.3 : 6.3 6.7 6.6 : 1.0 1.8 1.6 * 2.0 2.2 2.0 4 United States 18.0 18.2 18.2 : 5.5 5.3 5.3 : 3.9 4.4 4.5 : 6.9 7.2 7.2 : 0.6 0.4 0.4 . 1.1 0.9 0.8 Hungary 21.2 23.8 24.4 3.6 4.0 4.0 * 3.3 3.9 4.1 7.8 8.9 9.1 * 0.0 0.0 0.0 6.5 7.0 7.1 ("Soci t Incwom") OECD Lower 17.3 20.1 20.9 3.9 4.1 3.7 : 5.1 4.8 4.8 5.9 8.0 7.8 : 1.2 1.6 1.8 1.3 1.6 2.9 Income States OECD Social 30.3 30.4 30.6 : 6.7 6.3 6.3 : 6.3 6.3 6.4 : 9.8 9.7 10.1 : 1.4 1.6 2.5 : 6.0 6.5 5.3 Welfare States Sources: OECD: OECD Social Data Bank (OECD in Figures - Supplement to the OECO Observer). Hungary: CSO Statistical Yearbook, Social Security Administration, other government officials and Staff estimates. Notes: 1. For Hungary, 'other' social expenditure does not include consumer and housing subsidies, but includes social welfare, cultural and sports services not included in OECD definition. 2. Hungary: Education expenditure includes both scholarships and education, kindergarten. Health expenditure refers to health services and pharmaceutical subsidies. 3. OECD definition: Education - expenditure on pre-primary, primary, secondary, tertiary, education affairs and services and subsidiary services to education. Health - expenditure on hospitals, clinics and medical, dental and paramedical practioners, public health, medicaments, prostheses, medical equipment and appliances or other prescribed health-related products, and applied research and experimental development related to health and medical delivery systems. Pensions - expenditure on old-age, disability or survivors' benefits, other than for government empLoyees, and government employee pensions. Unemployment Compensation - expenditure on social insurance and other government schemes to individuals to compensate for loss of income due to unempLoyment. Other Social Expenditure expenditure on sickness, maternity or temporary disablement benefits, family and child allowances, other social assistance and welfare affairs and services. 4. OECD lower income states: Greece, Ireland, Portugal, Spain and Turkey. 5. oECD sociaL welfare states: Belgium, Denmark, Finland, France, NetherLands, Norway and Sweden. 8 comparator groups, however. Hungary surpassed the Lower Income States in expenditure on pensions and the residual "other" expenditure (mainly sick pay and family benefits), but fell short on health services and had no unemployment assistance during the period examined. In education spending, Hungary about matched the Lower Income States but remained well below the Social Welfare States (see Figures 2.3-2.5). These differences in spending cannot be explained by differences in the demographic profiles of the various countries. Hungary's demographic composition would call for significantly higher health and education spending than it has had; its share of elderly population partly explains its greater pension expenditure relative to the OECD Lower Income States, but policy not demographics accounts for most of the growth in Hungary's pension expenditure as explained later in Chapter 4. 2.06 Looking at the second half of the decade, all categories of social payments in Hungary have continued to grow significantly relative to GDP, with the exception of sick pay and maternity benefits, which have remained steady. Although there are no data available on the more recent trend in social spending by all OECD countries, most of them had experienced a marked decline in the growth of this spending in the early 1980s compared to the previous two decades, and indications are that their expenditure shares have at best held constant since 1986. Thus, by the present day, Hungary's overall ratio of social income transfers to GDP is close to that of the Social Welfare States--in other words, well above what would be expected for its income level. Whereas Hungary surpasses even this group in family benefits, it remains considerably behind in the share of total resources devoted to health, education, unemployment support, and social assistance. 2.07 The data bases for the OECD do not include consumer subsidies or total housing subsidies within social expenditure. While the former item is probably not significant in most OECD countries, it has been an intrinsic element in the "distribution system" in Hungary in the past (Table 2.2). Counting these subsidies, the "grand total" of social expenditure in Hungary currently amounts to one-third of GDP, half of household disposable income, and almost 60 percent of general government expenditure (Annex Tables II.3a- C.) 2. Growth of Social Income Transfers 2.08 The annual growth of Hungary's social income (excluding consumer and housing subsidies) in the 1980s has consistently exceeded that of GDP (Figure 2.6), with an income elasticity well above that evidenced by most OECD countries even in the two previous decades. In examining the sources of this growth for the three largest social programs (education, health, and pensions), it appears that most of the increase in total education and health expenditure is attributable to growth in the average real benefits per capita, while the expansion of eligibility for pensions accounted for the greater share of expenditure growth by this program. In brief, it was policy developments, rather than the simple demographic trends, that explain the expansion of social income transfers during the past decade. 9 e a w .1 oW._t Nalls 5jlt,,, r~~~~~~~~~~~~~~~~~~~~~~~~~ Om .~~~~~~~~~~~~~~~~~~~~~~~----- ........ . . . p ..... *~~~~~~~~~~~~~~~~~~~~~~~- ---- ............,.. I~l -w I -t-~ ia--.-- I--- .- ....|--'-T-----' . I .... ...... 1WO T W-w I z ~~~OECDO- W _1 io --------------------------~-- - E '-' 6 ~ ~ | ~ ~ ~ ~ ~ ----------- .... . IIKT .t 2 -- - ------..... - ll100 0,I - ,_t -i OW ._-_ ane. 00O N _ 1 Go.O T_.M[ Table 2.2 - Hungary: Social Expenditures Bln Ft ----------------------------------------------------------------------__-----__---------------------------------------------------------------------__-----__------------------ 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 -----------------------_-____----------____________________--------__--------__---------- _--------____________________--------------------------__--------__------------------ Pensions 27.1 32.3 36.3 40.0 48.3 56.0 61.2 68.5 75.0 84.1 91.7 99.3 110.0 130.0 156.4 202.5 Sickness benefit 6.9 6.7 6.9 7.6 8.1 8.5 9.1 9.7 10.7 11.0 11.6 12.8 13.3 16.7 21.4 24.5 Family allowance 6.4 7.5 8.5 8.8 10.5 13.6 14.6 14.9 16.0 19.4 21.3 21.5 23.2 37.0 52.8 63.2 Pregnancy/maternity benefits 1.7 1.9 1.9 1.9 1.9 1.9 1.9 1.8 1.8 1.9 2.0 2.6 2.8 3.7 4.1 4.8 Childcare allowance 3.0 3.5 3.8 3.7 3.7 3.9 3.7 3.5 3.4 3.4 2.8 2.0 2.0 1.9 2.7 3.2 Childcare fee 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.6 3.4 4.1 7.1 8.3 10.7 Scholarships 0.7 0.7 0.8 0.8 0.9 1.0 1.0 1.2 1.1 1.4 1.6 1.6 1.7 1.5 1.8 2.6 Other cash social income 1/ 1.6 1.5 1.7 1.8 2.0 2.1 2.4 2.8 3.8 4.4 4.7 4.9 5.7 8.1 8.6 17.8 o/w: regular social support 0.3 0.6 0.9 0.9 1.0 1.2 1.4 1.9 irregular social support 0.05 0.2 0.9 0.6 0.8 0.9 1.6 1.7 Employment Fund benefits in cash 6/ 0.3 1.0 6.4 Social income - cash 47.4 54.1 59.8 64.5 75.4 87.0 93.9 102.4 111.8 125.6 137.3 148.1 162.8 206.3 257.1 335.7 Social Income in kind 36.0 38.9 43.0 49.5 53.8 66.1 74.6 83.0 89.2 97.6 109.0 117.1 128.5 151.5 179.8 228.7 of which: Health 2/ 10.8 11.6 12.6 14.3 16.0 23.6 27.4 30.7 32.6 35.9 40.6 45.1 50.0 57.2 71.0 93.4 of which: Pharmaceutical Subsidies 5.4 6.4 7.2 8.4 9.7 12.0 13.8 15.8 18.2 19.3 26.5 Social services, nurseries 2.0 2.3 2.6 2.9 3.3 3.9 3.6 4.3 4.7 5.3 5.9 6.4 7.2 8.9 7.6 9.3 Education, kindergarten 15.3 16.2 17.7 21.1 22.1 25.3 28.3 31.6 33.3 36.4 40.1 42.0 45.6 55.8 75.1 94.0 Culture, sports 3/ 3.9 4.6 5.6 6.3 7.1 8.0 9.6 10.4 12.2 13.3 15.0 16.2 17.5 20.6 19.1 19.9 Employment Fund benefits in kind 6/ 0.8 1.4 3.6 Total Social Income 83.4 93.0 102.8 114.0 129.2 153.1 168.5 185.4 201.0 223.2 246.3 265.2 291.3 357.8 436.9 564.4 Consumer Subsidies 4/ 58.7 67.7 69.8 72.7 65.6 67.6 76.3 84.5 59.2 58.6 41.0 Housing Assistance 5/ 5.9 5.1 6.7 7.1 7.6 9.1 14.1 16.4 28.6 64.7 68.8 Total Social Expenditures 217.7 241.3 261.9 280.8 296.4 323.0 355.6 392.2 445.6 560.2 674.2 Memorandum Items: Total Household Income 306.9 327.6 359.5 389.6 422.6 481.8 .501.2 539.2 581.3 634.0 688.5 740.6 807.0 920.7 1066.1 1390.4 GDP (Bln of Current Ft) 482.7 528.9 582.0 629.7 682.3 721.0 779.7 847.9 896.4 978.4 1033.7 1088.8 1226.4 1409.5 1716.7 2041.8 GDP (Bln of 1981 Prices) 640.0 662.2 708.8 739.3 750.5 751.0 772.6 794.5 800.3 821.5 819.4 832.0 865.7 865.2 863.5 807.4 GDP Deflator (1981 - 100.0) 75.4 79.9 82.3 85.2 90.9 96.0 100.9 106.7 112.0 119.1 126.2 130.9 141.7 162.9 198.8 252.9 Sources: Data for 1975-88 from CSO Statistical Yearbook; for 1989-90, Social Security Administration, other government officials and Staff estimates. For growth rates and percentage distributions of above data, see Tables II.3a-c. 1/ Includes funeral allowance. 2/ Does not include compensation for pharmaceuticals, therapeutic equipment in 1975-79. Data on subsidies for 1980-87 from IMF Fiscal Affairs Dept, "Social Security Reform in Hungary". 3/ Includes expenditure by enterprises up to 1988. 4/ Data up to 1987 from Hinistry of Health & Social Affairs, February 1989. The 1988-89 data from SAL Preparation Mission November 1989; 1990 figure from Ministry of Finance. Includes subsidies for milk and dairy products, household energy, transportation, water supply and sewerage, and culture. To make comparable with data from earlier years, rental subsidies are also included here. 5/ Data up to 1987 from Ministry of Health & Social Affairs, February 1989. The 1988-89 data from SAL Preparation Mission November 1989; 1990 figure from Ministry of Finance. The figure for 1989 includes Ft 11 bln in interest subsidy on pre-1989 housing loans carried over from 1988. 'Assistance' includes interest subsidy on pre-1989 and post-1989 housing loans and 'social policy' grants for housing construction. Amortization of state-owned dwellings is included in Total Social Income in kind. 6/ Created in 1988. Fund covers unemployment compensation, retraining allowance, public works for job creation, early retirement, interest and risk guarantee for loans for new entrepreneurs, and administrative costs of Employment Services, Youth Employment Scheme, job creation subsidies to investors, and special funds for regions of critical unemployment. 29-Jan-91 14:43:48 NEWEXP.WK1 (Actual Values) 11 Figure 2.6 Growth Rates of Social Income and GDP Pe"rentag 81 Social Income Cash and Kind 6 -2 GDP -4 -6 -8 l l l l l 1982 1983 1984 1985 1986 1987 1988 1989 1990 Source: Table 11.3b. 12 3. Financing of Social Income Transfers 2.09 The average social insurance contribution rate for employers in Hungary already exceeded that of almost every other OECD country before 1988, when the rate was raised to a unified 43 percent of payroll--a level approached only by Italy and Spain. Hungary's contribution rate for employees is about the average for OECD countries, but added to the employers' rate gives Hungary a higher total contribution requirement than any OECD country (Annex Table II.7)4 2.10 While social insurance contributions have risen to cover about two-thirds of total social income payments in 1989 (up from 54% in the mid- 80s), revenue from personal income tax (PIT - introduced in 1988) is insufficient to make up the difference.5 As indicated in Table 2.3, considering the sum of employees' social insurance contributions and personal taxation relative to total social income received, the population in Hungary has been receiving a net social transfer equivalent to 17 percent of GDP, as compared to only 11 percent for the Lower Income and Social Welfare States. If the social insurance contributions of employers are considered as well (on the grounds that they represent at least to some degree foregone wages), the net transfer to households is 6 percent of GDP, still higher than the other country groups. 2.11 In sum, social income received by households in Hungary, as a share of GDP, is about at the middle of the range represented by the Lower Income and Social Welfare States of OECD. At the same time, personal taxation is the lowest in Hungary, and employers' social insurance contribution is the highest, which indicates that the social policy system in Hungary is directly financed more by taxation of employers than in the other countries. The ultimate incidence of this taxation is an important question which determines the related impact on incentives for workers to supply labor and for employers to hire. The argument of the "socialist distribution system" (para. 2.13 below) would imply that withdrawals from enterprises to finance social income transfers resulted in the reduction of wages--in other words, the incidence of the social insurance tax falls mainly on labor rather than on enterprise profits. It is not possible to settle this question empirically for Hungary, but it is probably safe to assume that given the traditional full employment 4 Spain's total contribution rate in 1989 ranged from 37-53% . See IMF (1990), Table 36. 5 Personal income tax accounted for only 9.5% of total taxation in 1988-89, compared to an average for OECD Lower Income States of 25% and 33% for the Social Welfare States during the 1980s. Hungary's PIT represented 5% of GDP, as opposed to 7% and 15% for the other two country groups, respectively, in the same periods. (OECD, Revenue Statistics of OECD Member Countries, 1965-87.) 13 Table 2.3: NET SOCIAL INSURANCE TAX AND TRANSFERS tX of GDP) Total SI Individual SI Total Personal TotaL Total Individuals Contribution Contribution Income Taxes Social Net Net Flow (Employees (Employees Only) Expenditures * FLow + EmpLoyers) % of GDP X of GDP X of GDP % of GDP X of GDP X of GDP (1) (2) (3) (4) (1+3)-4) (2+3)-4 OECD Lower Income States Average (1986) 8.3 2.5 7.4 21.0 -5.3 -11.0 OECD Social Welfare States Average (1986) 12.0 5.0 15.0 31.0 -4.0 -11.0 Hungary (1989) 15.2 4.0 4.5 25.4 -6.0 -17.0 Source: For OECD "Revenue Statistics of OECD Member Countries, 1965-87". For Hungary, Official Statistics. * As defined by OECD. ("Social Income in Cash and Kind" for Hungary, from TabLe 2.2). 14 and lack of competitive factor or product markets under socialism, workers bore the brunt of payroll taxation.' 2.12 To complete the analysis of how Hungary's social income transfers compare to those of other countries, it is necessary to examine not only their relationship to GDP, but also (i) the underlying structural differences in the economies, in particular concerning the share of wages in national income and (ii) the tax/benefit position of a comparable group of workers in each economy. These additional factors are considered below. 4. Social Incomes and Total Household Incomes 2.13 As mentioned in Chapter 1, a basic feature of the socialist system was the suppression of wages and redirection of profits from enterprises to the government budget and back to households through social cash transfers and subsidies on goods and services. Advocates of social policy reform in Hungary frequently argue for the reversal of this process, that is, for wages to be increased as social transfers are reduced, so that restructured wages "contain" the costs of social goods and services (e.g., housing) which are presently offered with substantial public subsidy. To evaluate this argument, the structure of labor compensation in Hungary has been compared to that of nonsocialist countries (Annex Table II.8). The theory of the "socialist distribution system" would suggest that the direct compensation of labor (gross wages and salaries) as a share of GDP is lower in Hungary than in most developed market economies. The analysis shows that for most of the OECD countries and Hungary, this share has been steady or falling through the past decade; and by 1987-88, labor's share of national income in Hungary was significantly lower than that of some advanced market economies (U.S., U.K., Canada, Japan, Switzerland), but about the same as the average for the Social Welfare States and well above that of Spain. Taking account of employers' contributions on behalf of workers for both social and private insurance-- considering, that is, the total cost of labor to employers--Hungary's share is still less than that of the Social Welfare States or the other high income states with a large practice of private pension contributions (e.g., the U.S., U.K., and Switzerland), but is greater than that of some of the Lower Income States which are closer international competitors with Hungary (Greece, Spain, Portugal). 6 Economic theory holds that when labor and product markets operate freely, the relative incidence of payroll taxation is determined by the relative elasticities of labor supply, labor demand, and the supply and demand for the product. If the demand for labor is less elastic than the supply, employers will bear more of the burden of the tax. On the other hand, if product demand is less elastic than product supply, employers can shift the burden to consumers through increased prices. Given Hungary's very high labor force participation rate and traditional full employment policy, the former condition (less elasticity of labor demand) may actually prevail. However, product markets have also been monopolistic, lacking even viable foreign competition, which would suggest that the tax can be passed on to consumers. The net effect may be that households bear most of the tax, either through reduced product supply and higher prices or through lower wages. 15 2.14 The conclusions to be drawn from this analysis are, first, that there is not a strong argument to be made, at least on the basis of international comparison, for the share of gross wages and salaries in the Hungarian economy to be increased as social expenditures are reduced. Secondly, the high rate of employers' social insurance contributions in Hungary does not raise inordinately the costs of employment as compared to many wealthier countries; however, given Hungary's income level and corresponding productivity-of its labor, the total cost of labor should be distinctly lower if Hungary is to effectively compete internationally, particularly with the other middle-income countries of Europe. 2.15 Analysis has also been made of the importance of social transfers relative to total incomes of a comparable group of workers in Hungary and in OECD countries (Annex Table II.9). The comparison reveals that the sum of take-home pay and cash transfers (equivalent to disposable income) was 106- 109 percent of gross earnings in Hungary in 1988-89, while only one other OECD country had a value above 100 percent. This imbalance is mainly the result of Hungary's relatively high family allowances, as well as its significantly lower average rate of personal income tax for the workers concerned. 5. Conclusions 2.16 The preceding international comparisons indicate that Hungary's social income transfers are indeed relatively high for its level of income, and especially so if the broadest definition of social expenditures (including consumer subsidies and housing assistance) is considered. These transfers are financed less by direct taxation of households than in comparator countries, implying that the population receives a very large net transfer from government. The large direct taxation of employers to finance social income transfers places Hungarian firms at a competitive disadvantage relative to enterprises in the lower income European countries, and this will become increasingly the case as a true market economy develops in Hungary and creates effective price competition. The reform of social policy must therefore reduce the rate of social insurance taxation of employers. However, it is not certain to what extent the share of gross wages in output would or should rise unless there are commensurate increases in productivity. To the extent that producers have been passing the burden of the social insurance contribution to workers, reducing this tax would permit wages to rise (if a competitive environment exists) while still preserving or improving the firm's profit margin. To the extent that employers have passed the burden to consumers, prices could be reduced so that workers realize higher real (inflation- adjusted) incomes. 2.17 If the Hungarian population chooses to maintain a fairly high rate of social income transfers on ideological grounds (and thus continue to resemble the model of the Social Welfare States of Northern Europe), it is reasonable to expect that households should bear more of the cost through higher social insurance contributions and/or higher personal income taxes. The net social transfer would then be reduced to a rate more in line with that in other countries of the social welfare orientation. Chapter 10 looks at the macroeconomic relationships among income aggregates and social expenditures as implied by various scenarios of social policy reform. 16 B. Sources of Growth in Household Incomes - Implications for Income Distribution' 2.18 During the past decade, the two most significant developments concerning household incomes were (i) changes in the composition of labor income (an increase in the importance of secondary employment in the private or quasiprivate sector), and (ii) the growth of social incomes. Total income derived from labor has declined throughout the 1980s, while social incomes continued their impressive growth, reaching 41 percent of household income by 1989. (Figure 2.7) 2.19 The spread of supplementary employment since 1982 has markedly increased the dispersion of incomes. As illustrated in Figure 2.8, secondary (supplementary) income, which accounted for the greater part of income growth during the past decade, is heavily concentrated in the upper income groups, as is "other income" (mainly from financial wealth). The introduction of the PIT in 1988 moderated the dispersion of income slightly, as only earnings from main employment were grossed-up (Figure 2.9). Among social income in cash, so-called "primary incomes of inactive earners" (pensions, childcare allowances, and unemployment benefit)8 are fairly evenly distributed across the population deciles, while "secondary social incomes" (family allowances, student aid, other cash transfers) favor the lower third of the income distribution (Figure 2.10). Social income in kind is very evenly distributed, in contrast to subsidies (Figure 2.11); taken together, these two categories of transfers moderate further the income distribution. Chapter 3 examines the incidence of each of the categories of social income (both cash and in kind) in more detail. 2.20 The effect of each source and type of income on the overall income distribution is indicated by the Gini coefficients summarized in Table 2.4 below. While the dominant source of income for most of the population ("Main Earnings") shows only a modest degree of concentration (coefficient of 0.24), incomes from self-employment, tourism, other second and third economy activities, and financial wealth increase inequality by 30 percent (coefficient rises to 0.32). The progressive personal income tax corrects this inequality only to a slight degree, with an incremental impact of 7 7 This section draws heavily on the 1990 Incidence Study, Chapter 1. 8 The classification of persons on unemployment benefit as inactive is contrary to international practice. 17 Figure 2.7 Structure of Household Income 100% 75% 50% 4 25% 0% 1976 1978 1980 1982 1984 1986 1988 Social Income Cash Social Income Kind [ Labor & Other Income Source: Table 11.3a. 18 Figure 2.3 Obatrlbutlon ol Labor inconm by Typo: Primary Income of Actlve Earner. Seeondary Labor Income and Other oncomn by Deole., 1BBtl Percentage 35 30 Note, Thea. 3 oiegaories oomprlIs *arned' labor income. I.e.. exclude caolul Inoon. exccpt sick pay hlob la I nluded among primary Incmne. 25- 20 1 2 3 4 6 B 7 8 9 10 Decile - Primary Income Secondary Labor Other Income (Supplemental) Source: Incidence Study 1990, Table 1.2.1/b. Figure 2.9: Effect of Taxation: Distribution of Gross Personal Income, Taxes and Net Personal Income by Docile 1 989 Percentage so 20 -/ 15 25 20 1 2 3 4 6 6 7 a 9 10 Decile - Gross Per Income Taxes Net Personal Income Source. Incidence Study 1990i Table 1.2.1/b Notes to Figures 2.8-2.11: Vertical axis indicates the percentage share of total amount of each type of income or transfer. Absolute equal distribution of each type of income or transfer would be represented by a horizontal line at ten percent on the vertical axis. Horizontal axis is deciles of per capita net personal income. For data on which figures are based, see Annex Table II.14. 19 Figure 2.10: Distribution ofi Social Income In Cash Primary Incom, of Inactive Earners and Secondary Social Incomc by Decilel Percontag1 18 Seaondary Social Income 4(Family allowances, studentg and other cash aid) 14 12 1 2 4 P mary Ineosonal Incoet ESnca nc I KPinsaon childre ullowasnese Punemployment benefis) 25 20 t 2 3 4 6 6 7 8 9 10 Dccile Soure: Incidence Study, 1990, Table t2.1tb. -ilgure Z 11: ulStribUtionl ot Net PeP sonal incoie, Social Income in Kind and Subsidies by Deiilea Std 9 Percentage 25 O ... . ....... . 1 2 3 4 5 6 7 a 9 10 Docile - Personal Income - Soclal Income (Klnd) --Subsidieas Souirce: Incidcnco Study, 1990, Table 1.22Vb. 20 percent between income definition (E) and (F). The largest reduction in inequality is effected by social income in cash, which reduces the Gini coefficient of Gross Earnings (C) from 0.32 to 0.25 for Gross Income (D) - a redistribution of 21 percent. The addition of social income in kind produces a further 18 percent reduction in inequality, while subsidies have only a negligible effect. Table 2.4 Gini Coefficients for Components of Household Income INCOME SOURCE GINI COEFFICIENT A. Main Earnings 0.2938 of which: Earning from employment 0.2445 Income of self-employed 0.6831 B. Main earnings PLUS Secondary Labor Incomes 0.3045 of which: Secondary jobs 0.4771 Tip, (reported) "black" income 0.3616 Income from small scale agriculture 0.2026 Income from tourism 0.5533 C. Main earnings plus Secondary Labor Income PLUS Other Incomes 0.3158 of which: Other agricultural income 0.0857 Interest received 0.4275 D. Gross earnings (C) PLUS Social Incomes in Cash - Total Gross Income 0.2483 Taxes 0.3357 of which: Social Insurance Premium 0.2789 Personal Income Tax 0.4133 Other taxes and duties 0.1841 E. Gross Income (D) LESS Social Insurance Premium 0.2465 F. Gross Income (D) LESS Social Insurance Premium and Personal Income Tax 0.2299 G. Gross Income (D) LESS All Taxes 0.2311 (continued on next page) 21 Table 2.4 Continued Transfers among Households of which: Received family support 0.3103 Given family support 0.1482 Received child maintenance - 0.1587 Given child maintenance 0.1787 H. Net of Tax Income (C) LESS Net Transfers - Total Personal Net Income 0.2313 I. Personal Net Income (H) PLUS Social Income in Kind 0.1907 J. Personal Net Income (H) PLUS Subsidies 0.2201 R. Personal Net Income (H) PLUS Social Income in Kind and Subsidies 0.1859 Source: CSO calculations based on 1989 and 1990 Incidence Studies. Notes: The Gini coefficients of components of social income in cash are provided in Table 3.4, and of social income in kind and subsidies in Table 3.5. (The Gini coefficient is equal to twice the area between the 45-degree line and the Lorenz curve of income distribution. When each individual receives (or pays) exactly the same amount of income (or tax), the Gini coefficient is equal to zero; at the opposite extreme, if the entire payment is made to (or by) only one individual, the coefficient is equal to one. The higher the absolute value of the Gini coefficient, the greater the concentration of income flows. Negative values indicate that lower income groups receive higher absolute amounts of the payment in question than richer households.) 2.21 Differences across households in average earnings and average pensions received play a relatively modest role in determining differences of income per capita. The dispersion of incomes per capita is explained mainly 22 by differences in the number of active earners per household, and the larger number of children in the lower income quantiles.9 2.22 To conclude this section, the distribution of final personal income in Hungary has widened during the past decade mainly due to the impact of second economy earnings. Households have been able to maintain their real incomes only to the extent that they can participate in such supplementary employment and have surplus labor on which to draw (i.e., women becoming active earners). The positive impact on labor incentives that could be expected from the growing importance of informal and private sector employment has been weakened by two factors. 2.23 First, the amount of industrial adjustment that has taken place over the decade in the socialist sector (State enterprises and cooperatives) has been quite limited, so that most workers remain tied to the traditional places of employment despite low wages and often lower productivity. The reallocation of labor into fulltime private sector employment has therefore been limited as well, and such employment merely as a supplementary activity has had a marginal effect in the income of most households, albeit accounting for whatever improvement in labor income they experienced. The second factor dampening work incentives has been the growing importance of social income, especially cash transfers, in total household incomes. Over the past decade the authorities have attempted to compensate the population for declines in real wages by increasing social policy expenditures (including substituting cash transfers in place of many consumer subsidies). This approach has contributed to weakening structural reform by sustaining a large government role in the economy and slowing necessary adjustments in the allocation of labor and of consumption. 2.24 The social transfers in cash and kind have had a significant equalizing impact on income distribution across income groups, while the consumer subsidies in the aggregate have not had this effect. The main contribution of social income transfers, however, may have been to provide a significant income supplement to the lowest income groups (discussed in Chapter 3), despite the considerable leakage of benefits to the upper income groups. The transfers to the middle strata have not been sufficient to counteract the erosion in their real labor income. 2.25 The priority for social policy reform should be to strengthen the effectiveness of social transfers in keeping the lower income groups above a socially-accepted minimum, through better targeting. 'For the middle and upper income groups, the best strategy to improve their incomes would consist of 9Another recent study draws similar conclusions that the income differences among households according to the occupation of the earners in the household have diminished over time, and instead such differences are explained more by the life cycle of the family and its participation in the second economy. "1... the decisive factors of actual differences of income.. .are (increasingly) outside the world of employment." See Julia Szalai, "Poverty in Hungary during the Period of Economic Crisis", Background Paper prepared for the 1990 World Development Report, November 1989, p. 19-20. 23 policies to strengthen the economy and the flexibility of the labor market. Ultimately, the surest way of achieving a sustainable growth in private incomes and overall welfare is through a resolute implementation of structural reform, thereby promoting the mobility of labor in response to opportunities for improved productivity, coupled with macroeconomic policies to reduce inflation. The existing imbalance in earnings between the "main employment" (essentially the socialist sector) and private employment is largely an artificial distortion created by the past restrictions on private activity and the rigidity of the large enterprises. As mentioned earlier, a popular argument in Hungary is that the wages in main employment should be ratcheted- up to accommodate any new social policy system. The only way that such wage adjustment could occur would be through the effective restructuring of enterprises in the socialist sector to reduce overmanning, thereby permitting wages and productivity of the remaining employees to rise; at the same time, labor would shift to the private sector, where increased activity would reduce the scarcity rents which currently appear as relatively high wages for private employment. The simple conclusion is that labor incomes will improve with progress in the overall process of economic restructuring. C. The Profile of Poverty 2.26 One of the principal reasons for interest in the current distribution of income in Hungary is its implications for the prevalence of absolute poverty, as defined for the country. The methods used to estimate the minimum subsistence levels are discussed in Annex II. 1. Changes in share and composition of poverty group over time 2.27 A recent study by Szalai (1990) compares the results of the nationwide Income Survey (conducted on 20,000 households every five years) and Household Budget Surveys of 1977 and 1987 to determine the changes in poverty over time. The data series reveal that whereas the share of the total population living below the subsistence minimum has not changed greatly over the decade (ranging from 10-12 percent from the Income Survey and 14-17 percent on the basis of the Household Survey), what is more striking is that the incidence of poverty in households with active earners has marginally increased while in inactive households poverty has decreased. Moreover, in 1987, a higher proportion of households were poor in urban areas than in rural areas; the ratios of those living below the poverty line as a share of the respective population were 17.3 percent and 14.5 percent for urban active and inactive households, respectively, versus 10.7 percent and 9.6 percent for rural active and inactive households. The explanation for this shift is believed to the be better access of rural households to earnings from private economic activity, namely agriculture. 2.28 The risk of poverty is greatest and is increasing for children, who in 1987 comprised 39 percent of the poverty group as against 27 percent of the total population. Persons on child care benefits were represented twice as much in the poverty group as in the total population, which is consistent with the observation that poverty often reflects the life cycle (in this case, the lower earnings capacity of women raising children). Pensioners comprised 15 percent of the poverty group in 1987 as opposed to 25.5 percent in 1977, 24 while their share in the total population increased. The Income Surveys also showed clearly that the risk of poverty increased with the number of dependent children in the household. The Household Budget Surveys indicate the same trends, revealing, for example, that whereas 14 percent of the total population lived below the subsistence level in 1987, 40 percent of urban persons on childcare benefit and 28 percent of urban children were in poverty.'° 2. Description of Poverty Group in 1989 2.29 The 1989 Household Budget Survey provides an opportunity to examine in some detail the sample population below the poverty line (the 1989 subsistence income level) in comparison to the sample population at the average income level. The HBS reveals that the risk of poverty varies with the number of dependents in the household and with age. 2.30 Table 2.5 compares the age distribution of these two sample populations. As indicated, children comprise a significantly higher share of the population in poverty in both active and inactive households. Among poor inactive households, persons of childbearing age and over 70 years of age are also overrepresented. 2.31 To conclude, the evidence on the incidence and nature of poverty suggests that corrective policies need, in the first instance, to be targeted to children; in this light, the emphasis of some of Hungarian social policies on support to large families is not misplaced, particularly in the absence of more refined approaches to determining need. Secondly, policies to redress poverty should aim for better protection of pensions against inflation. Thirdly, measures to permit women with children to remain in the labor force if they so wish should be strengthened, since the availability of a second income in households is obviously an important factor in keeping members above the poverty line. A very rough estimation of the "poverty gap"--the total amount of income by which the poor population falls below the subsistence level--indicates a figure of only about Ft. 10 billion in 1989, or 0.6 percent of GDP. This calculation, although certainly an underestimate of the current poverty gap, suggests that at least some share of existing social expenditure could be redirected through tighter targeting on the needy without causing substantial hardship to the population at large. Of course, an approach to address the more fundamental causes of poverty should include an enhanced program of education and training to increase earnings capability of the vulnerable groups within the working age population. 10 From Szalai (1990) Tables 4,5,7-13. 25 Table 2.5: AGE DISTRIBUTION OF POOR AND NONPOOR NHOSEHOLDS Active households Inactive households with average with income under with average with income under Age income the poverty Line income the poverty tine 0 - 5 4.4 12.9 0.8 3.1 6 - 14 14.1 27.7 0.4 6.2 15 - 44 47.3 47.5 1.6 9.9 44 - 69 32.2 10.2 73.8 42.7 over 70 2.0 1.7 23.4 38.1 Total 100 100 100 100 Source: Ministry of Finance, Central Statistics Office, and World Bank, "The Hungarian Soial Policy Systems and Distribution of Incomes and Households," Budapest 1990, Table 1.2.1/a. 26 3. The Special Case of Unemployment 2.32 Unemployment as a normal economic phenomenon has only recently been recognized officially and openly observed in Hungary."1 In May 1990, the national total of registered job vacancies for the first time fell below the number of the registered unemployed. The registered unemployed totaled about 100,000 workers in January 1991 (about 2 percent of labor force). 2.33 Unemployment is likely to increase the incidence of poverty and the associated welfare problems in Hungary during the next two-three years of the structural adjustment. There is already evidence that unemployment to date has hit hardest the lower income segments of the labor force, in particular the unskilled (see Chapter 5). Workers with difficult family and social situations, such as single parents and persons in poor health, who already have a tenuous economic status were among the first to be affected by workforce reductions in many firms. Unemployment is the highest in those counties (Borsod, Szabolcs-Szatmar, and Hadju-Bihar) where the number of households below the poverty line was 1.3-1.6 times the national average in 1989. These counties are characterized by a higher than average percentage of unemployed with only primary school education and a relatively low share of job vacancies corresponding to these qualifications.12 Thus, a significant share of the unemployed to date face particularly difficult prospects of reemployment. Adding to these objective constraints, recent sociological research in Hungary has found that the society's unfamiliarity with unemployment has led to a commonly pejorative attitude towards the unemployed "i The labor force in Hungary is defined statistically as equivalent to the number of "active earners"; persons wishing to work but not actually employed have never been counted in labor force statistics, due to the official nonrecognition of unemployment. The number of active earners within the working age group (15-54 for women, 15-59 for men given the usual age of retirement) was 4.7 million in 1989, implying an age-group participation rate of 80.7 percent for men and 74 percent for women. However, it is clear that disability pensions, sick pay, and childcare allowances have been used to some extent as a means of withdrawing some of the potential workforce from active status who would otherwise be counted as unemployed. The number of so-called "inactive earners", persons within the working age group receiving these transfers, was over 600,000 in 1989. Taking half of these inactive earners as putative members of the labor force (those who might be drawn back into the labor force if these social transfers became more restrictive, as recommended later in this Report), the total labor force is then estimated as 5.0 million in 1989. The implied total participation rate is therefore 83 percent (85.2 for men, 80.0 for women). These would have to be considered maximum participation rates for the future. 12 Regional unemployment data supplied by (former) State Office of Labor and Wages in November 1989. Regional poverty information from 1990 Incidence Study. 27 individuals, which is likely to worsen their real or perceived decline in welfare and contribute to psychological and health problems."3 2.34 Over the next few years, the rate of total unemployment is expected to increase considerably. The Ministry of Labor is projecting the rate to more than double to 5 percent by end-1991, and to reach 450-500,000 (10 percent) by 1993. Other projections put the peak rate of structural unemployment due to restructuring (not counting frictional unemployment) in the next one to two years between 7-14 percent, depending on estimates of actual redundant workers, the pace of restructuring, and the average duration of unemployment (World Bank, 1990a, p. 73). It is reasonable to expect that Hungary would have to face long term steady-state unemployment rates at least as high as those in the smaller OECD countries which have been fairly successful in containing unemployment (e.g., Austria and Sweden)--about 3 percent. Therefore, planning for the transition period of the next few years will have to take account of a range of total unemployment between 10 to 15 percent (500,000-750,000 workers). Prolonged and repeated episodes of unemployment are also likely to become more common. Chapter 5 evaluates the existing and proposed programs of support for the unemployed in light of these prospective requirements. D. Demographic Prospects - the Demand for Social Policy 2.35 The future demands for social income transfers will be determined by demographic trends, growth in incomes, and not least, by the population's expectations based on past experiences. The demographic factor is the easiest to predict with some certainty. In the present Report, considering the population trend over the next decade is sufficient for the analysis of reform requirements and options for all areas of social policy except pensions, for which longer term (sixty year) projections are needed. The parameters of population structure which are relevant to social policy are mainly the size of the youth population (affecting in particular needs for child care, family assistance, and education); the retirement age group (determining needs for pensions and health care); and the working age group (relevant for potential unemployment and for the tax base to finance social policy). The key relationships are represented by the youth, old age, and total dependency ratios--generally speaking, the higher these ratios, the greater the demand for social benefits and the corresponding financial burden. Figure 2.12 projects the trend of dependency ratios over the next six decades according to a "medium" set of assumptions about fertility and life expectancy (summarized in Annex Table II.19). Alternative projections based on higher and lower estimates of these parameters indicate in each case that the total dependency ratio will bottom-out by the middle of the present decade, after which it resumes a steep and almost steady climb through about the year 2040. The youth dependency ratio reaches its lowest point around 1995, with only slight change thereafter. The share of the elderly population is what mainly 13 Such attitudes and the self-perceptions of the unemployed have been documented in research by Maria Lado and Ferenc Toth, "Workforce Reductions and their Effects on the Survivors", mimeo, 1990 and also by Karoly Fazekas and Janos Kollo, "Dismissals, Unemployment, State Care," Valosag, No. 11, 1985. Figure 2.12 Hungary: Dependency Ratios Under Scenario II, 1949-2051 1/ Pecnt of ctive popuation 100 - 100 90- - 90 80- Total dependency ratio - 80 70 - - 70 60- Old-age dependency ratio 60 50- - 50 Youth dependency ratio 40 - .- .. / - 40 ....... 30 - / ' 30 20 ~~~~~~~~~~~~~~~~~~~~~~~~~~-20 1951 1961 1971 1981 1991 2001 2011 2021 2031 2041 2051 Source: Research Institute for Population Sciences, Central Statistical Office; and Fund staff estimotes. 1/ Defined, respectively, as population aged 0-14, 55+, and 0-14 plus 55+, to active population aged 15-54. 29 determines the growth path of the total dependency ratio, and it is clear that under all reasonable scenarios Hungary will face a difficult task to manage the sharply growing demands for supporting its aged population. 2.36 The main conclusion to be drawn from these perspectives is that Hungary currently enjoys a brief spell of reduced demographic pressure, and must take advantage of the next four to five years to put in place a set of social transfer policies that can be financially sustained during the much more demanding period ahead. The purely demographic pressures from the youth population will subside, but the need for greater enrollment coverage of secondary and higher education and for quality improvements will keep up the pressure for expenditure in this sector. The sharp dip in all the dependency ratios around 1996 reflects a temporary surge in the working age group (the denominator). This short-lived increase occurs at an unfortunate time for the labor market, when prospects of structural unemployment will already be severe. Therefore, support for the unemployed will have to assume a much greater importance among components of social policy during the next few years. Since labor participation rates have already reached their maximum (see Footnote 11), the projected decline in the working age group after 1996 will imply a contracting labor force and thus a reduced tax base for social insurance. This perspective underlines the urgency of the pension reforms discussed in Chapter 4. 30 PART TWO: ASSESSMENT OF THE SOCIAL POLICY PROGRAMS AND RECOMMENDATIONS FOR REFORM CHAPTER 3 THE DISTRIBUTIONAL IMPLICATIONS OF SOCIAL TRANSFERS 3.01 As indicated in the Introduction, the broad criteria for evaluation of each social program are the effectiveness and internal efficiency in meeting its objectives (be they income support or service delivery), financial viability, and impact on economic incentives. This Chapter gives an overview of one dimension of effectiveness and efficiency which can be measured consistently and quantitatively for each of the social transfer programs. In this discussion, "effectiveness" is defined in terms of the importance of the transfers in relation to the household income of population groups--that is, how much each group is benefitting by the transfers relative to its other sources of income. "Efficiency" is viewed here in terms of the distributional impact of each type of transfer--who is benefitting more relative to others. The indicators cited below, taken from the 1989 and 1990 Incidence Studies, convey an objective description of the data on transfers and do not themselves suggest who should benefit from them or to what degree. There are also more specific definitions of effectiveness and efficiency pertaining to some of the individual social programs which will be discussed in the later chapters. A. Effectiveness of Social Transfers 3.02 The Incidence Studies provide a comprehensive breakdown of social transfers in cash and kind in relation to total income of households at different points on the income distribution. The studies thus convey a numerical picture of the relative importance of transfers within the household budgets and, in this limited sense, the effectiveness of each program in supplementing income from other sources. Table 3.1 refers to all the components of cash income, including social income in cash. In the two lowest deciles of the income distribution, social transfers (Primary Income of Inactive Earners and Secondary Social Incomes) provide about 40 percent of gross personal income, with family allowances accounting for 12-18 percent, pensions about 16 percent, child care allowances 3-4 percent, and social assistance 2 percent. In the upper two- tenths of the income distribution and for the population as a whole, only pensions represent an important share of gross personal income (although some of the smaller items, e.g., sick pay, unemployment benefit, and student grants, comprise practically the same share of gross income across the distribution). Although social cash transfers thus represent a large income supplement to the lower deciles, it should be remembered that some ten percent of the population remains below the poverty level on the basis of their income including such cash transfers. 3.03 The relative contribution of social income in kind and of subsidies to household incomes is shown in Tables 3.2 and 3.3. Social income in kind, especially education and health services, represent a significant (32-42 percent) increment to the net personal income of the lowest fifth of the income 31 Table 3.1: THE RELATIVE IMPORTANCE OF DIFFERENT KINDS OF INCOMES, TAXES AND TRANSFERS IN THE LOWER AND UPPER DECILES, 1989 (in percent) 1 2 9 10 Totat decile I decile I I decile I decile o t Main earnings 43.4 48.9 55.0 62.4 54.8 Sick Payment 2.3 2.1 1.8 1.8 2.0 Primary incomes of active earners 45.7 50.9 56.8 64.2 56.8 Pension 15.7 16.9 11.8 6.4 13.4 Child care aLl./fee 4.1 3.4 0.3 0.2 1.0 Unempl. benefit 0.1 0.1 0.0 0.0 0.0 Primary incomes of inactive earners 19.9 20.3 12.1 6.6 14.4 Sebondary (supplementary) labor incomes 10.1 10.1 18.0 16.8 14.8 Family allowance 18.1 11.9 2.0 1.1 5.0 Grants for students 0.2 0.3 0.2 0.1 0.2 Other benefit, aid 2.4 1.6 0.4 0.3 0.8 Secondary social incomes 20.7 13.8 2.6 1.4 6.0 Other income 3.5 4.8 10.5 11.0 8.1 Gross personal income 100 100 100 100 100 Pension premiums 5.6 5.2 6.3 5.8 5.8 Personal income tax 3.7 5.0 9.9 12.2 8.5 Other taxes, duties 3.0 2.6 2.0 1.8 2.2 Taxes 11.2 12.8 18.3 19.8 16.5 Household transfer received + 2.0 1.7 1.4 1.3 1.4 Household transfer given - 3.2 1.7 1.3 1.3 1.4 Household transfers -1.2 0.0 0.1 -0.1 0.0 Net Personal Income 87.6 87.2 81.8 80.2 83.5 Source: Ministry of Finance, Central Statistics Office, and World Bank "The Hungarian Social Policy Systems and Distibution of Incomes and Households," Budapest 1990, Table 1.2.1/e 32 Table 3.2: THE RATIO OF SOCIAL INCOMES IN KIND TO NET PERSONAL INCOME IN THE LOWER AND IN THE UPPER DECILES, 1989 (in percent) 1 | 2 | 9 10 Total decile I decile deciLe I decile Kindergartens 4.7 2.9 0.4 0.3 1.1 Primary schools 11.9 8.3 1.7 0.8 3.5 Secondary schools 3.4 2.9 1.1 0.4 1.9 Higher education 0.6 1.4 0.3 0.6 1.0 Other trainings 0.3 0.3 0.3 0.2 0.3 Education Subtotal 20.8 15.9 4.3 2.3 7.8 Hospitals 6.4 5.1 2.5 1.9 3.5 Other health 3.8 3.1 1.5 1.2 2.0 Health Subtotal 10.2 8.2 4.1 3.1 5.5 Nurseries 2.1 0.6 0.0 0.1 0.4 Social items 4.0 2.8 0.0 0.0 0.5 Vacation 0.3 0.5 0.5 0.4 0.5 Culture, sport 3.5 3.0 2.3 2.0 2.5 Canteen bonuses 0.6 0.6 0.6 0.4 0.5 Depreciation of state dwellings 0.5 0.5 0.4 0.3 0.4 Other social income in kind, Subtotal 11.0 7.8 3.8 3.1 4.8 Social income in kind, Total 42.0 31.9 12.2 8.5 18.1 Personal (net) income 100 100 100 100 100 Source: Ministry of Finance, Central Statistics Office, and World Bank, "The Hungarian Social Policy Systems and Distribution of Incomes and Households," Budapest 1990, Table 1.2.2./g. 33 Table 3.3: THE RATIO OF SUBSIDIES TO PERSONAL NET INCOME IN LOWER AND THE UPPER DECILES, 1989 (in percent) 1 ?2 9 I 10 I Total decile I deciLe decile i decile i I Rent 1.1 1.1 0.9 0.6 0.9 Heating 2.9 2.7 2.0 1.4 2.2 Water, sewage for state dwellings 0.3 0.3 0.2 0.2 0.2 Water, sewage for owner occupied dwellings 1.3 1.0 0.4 0.3 0.6 Milk 0.3 0.2 0.1 0.1 0.2 Dairy products 0.4 0.4 0.2 0.2 0.3 Transportation 2.5 1.9 1.3 0.7 1.5 School books 0.1 0.1 0.0 0.0 0.0 Theater, etc. 0.2 0.2 0.3 0.3 0.3 Consumption subsidies, Subtotal 9.1 7.9 5.6 3.7 6.2 Medicine subsidies 3.0 2.4 1.3 0.9 1.6 Mortgage payment 5.6 5.3 6.9 5.8 5.5 Other housing inv. subsidies 4.5 2.2 1.0 1.0 1.3 Housing invest. subsidies, Subtotal 10.1 7.6 7.8 6.7 6.8 All subsidies, TotaL 22.1 17.9 14.7 11.4 14.5 Personal (net) income 100 100 100 100 100 Source: Ministry of Finance, Central Statistics Office, and World Bank "The Hungarian Social Policy Systems and Distribution of Incomes and Households," Budapest 1990, Table 1.2.2./h. 34 distribution. Subsidies add about another 20 percent to their income. These transfers have an even greater importance in the total net personal incomes of certain subgroups of households. It appears that the larger the number of children in a household, the greater the dependence on social incomes in kind. Education benefits alone, for example, represent 25 percent of personal net incomes per capita in active households with three or more children. Health benefits alone are 15-20 percent of personal net income of inactive households in the lower fifth of the income distribution. Within the economically vulnerable subgroups (large active families and inactive households), almost 10 percent of the households below the second income decile receive as much or more income from noncash benefits and subsidies as they do from earnings or cash transfers.' 3.04 In a purely quantitative sense, the above analysis indicates that social transfers (particularly noncash) serve as a significant supplement to other components of household income for households with relatively low earned income. This observation is not to suggest that social noncash income has the same value to the households as cash income; indeed, economic theory would hold that transfers in kind are less valued unless it can be assumed that they coincide exactly with the households' consumption preferences. Noting that noncash transfers at present represent a larger supplement to household incomes than cash transfers, households' utility could be increased by shifting the balance more towards the latter. This in fact has been the recent trend of social policy in Hungary as subsidies have been cut since 1988 with partial compensation transferred through pensions and family allowances. The recommendations made in this Report and other recent World Bank studies (1990a,b,c) for introducing copayment and some private provision in health and other social services, and for replacing housing subsidies on State flats with housing allowances allocated to households, are also consistent with the objective of permitting greater consumer choice and increasing the utility of social income transfers. B. Efficiency of Social Transfers 3.05 In one sense, social transfers may be considered to be distributed efficiently if they reach the population to which they are intended. Universal transfers in principle should be allocated more or less evenly among the entire population, while targeted benefits by definition should be received by those designated as eligible and none other. In Hungary, as explained in Chapter 2, only a few benefits are formally "universal" and most are contributory--although with very high labor participation and a generous definition of "service periods" qualifying as contribution to social insurance, social insurance benefits are also practically universal. Only very limited social policy transfers have been explicitly targeted on the basis of income, that is, with the objective of redistribution from richer to poorer groups. Under socialism, Hungary's wage policy was deliberately egalitarian and thus social policy was not intended to correct for inequality. In reviewing below the actual distribution of social transfers among income strata, we do not make the assumption that redistribution (reducing inequality) was an objective of social transfers; in fact, certain Incidence Study, 1990, Chapter II, Sections 1.3 and 1.4. 35 transfers (notably some housing subsidies) may have been used deliberately as a kind of reward system to counteract the suppression of wages and thus increase inequality. In the future, however, as wages are liberalized and social policy loses its role in the socialist-style "distribution system", redistribution to reduce income inequality and maintain an income floor would be expected to become the primary objectives of social policy as in most market economies. In light of the changing rationale for social transfers in Hungary, it is helpful to assess the actual distributional impact of the existing transfer policy. 3.06 In the description below, "progressivity" of social transfers (as of taxes) is defined as an increase in their rate with increase in income levels. Similarly, "regressive" policies are those in which the rate of transfers (or taxes) is inversely related to income levels. Thus, if reduction of inequality is an objective of social policy, transfers should be regressive and taxes progressive. 3.07 The distribution of social income in cash across income deciles is indicated in Annex Table III.1. The cash transfers are also listed below in declining order by Gini coefficients, i.e., in increasing order of their effect in promoting income equality. Aside from "other aid", only the child care allowance/fee and family allowances are distinctly regressive (promoting redistribution to the lower income groups), which results from their being broadly related to numbers of children in a household. The strength of regressivity of the childcare allowance/fee presumably reflects the fact that more lower income women take advantage of the benefit, because the opportunity cost to them of foregoing earnings for fulltime childcare is relatively low. Sick pay and pensions, as income replacement programs linked to prior earnings, are predictably progressive; however, they are less concentrated on upper income groups than are other income aggregates, and so on balance have some redistributive impact as well. As a more steeply degressive schedule of social insurance pensions has recently been adopted (see Chapter 4), and as private pension savings becomes a common practice, the public pensions would assume a greater role of income equalization. The Gini coefficient for student grants is lower than that for higher education (Table 3.5), indicating that the means testing used in allocating some of these transfers within the student population has some redistributive effect. Table 3.4: Gini Coefficients for Social Income in Cash Sick pay 0.2214 Pensions 0.0954 Student grants 0.0925 Other aid a/ - 0.1147 Unemployment benefit b/ - 0.1584 Family allowance - 0.2042 Child care allowance/fee - 0.3125 Source: Central Statistics Office calculations, based on Incidence Study 1990. a/ Includes regular and irregular social assistance and funeral grants. b/ Note that the small number of observations of unemployment benefit recipients in the Household Budget Survey (only 60 households) makes this item statistically unreliable. 36 3.08 The distribution of social incomes in kind and subsidies is summarized in Annex Table III.2. In both categories, transfers related to recreation, culture, housing, and higher education are the most concentrated on upper income groups, while transfers directed to young children and the aged are the most "income-leveling". These effects are apparent in the ordering of Gini coefficients in the following Table 3.5. 3.09 Income in kind through vacations, canteen bonuses, culture and sport is provided to a large extent by State enterprises and especially to their salaried employees, which accounts for the progressive distribution of these items over the population at large. Education benefits appear regressive overall, even at the secondary level, which is surprising given the relatively low enrollment rates for post-primary education. Health benefits appear neutral with respect to income, which would be expected of a universal social benefit except for the fact that lower income groups and the (low income) elderly tend to exhibit the worst health conditions. Evidence on the physical distribution of health infrastructure and the incidence of health morbidity and mortality in Hungary confirm that health services are not provided equitably in all regions of the country or according to actual health needs. (See World Bank, 1990b) It must be acknowledged that methodological issues (see Annex III) make the interpretation of the incidence analysis of social income in kind more tentative than that of transfers in cash or subsidies. 37 Table 3.5: Gini Coefficients for Social Income in Kind and Subsidies SOCIAL INCOME IN KIND Vacations 0.2119 Canteen bonuses 0.1744 Depreciation of State flats 0.1498 Culture, sport 0.1426 Higher education 0.1304 Hospital care 0.0402 Other health care 0.0396 Secondary school - 0.0723 Other training - 0.1997 Primary school - 0.2014 Kindergarten - 0.2465 Nurseries - 0.3216 Social items - 0.7382 Total Health 0.0399 Total Other Services 0.0314 Total Education - 0.1198 Total Social Income in Kind - 0.0311 SUBSIDIES Cinema, theatre 0.2597 Mortgage payment 0.2573 Water supply, sewage disposal in rented State dwelling 0.1511 Rent 0.1500 Residential heating 0.1128 Dairy (cheese, butter) 0.1006 Transportation 0.0649 Medicine 0.0454 Social policy allowance for housing 0.0314 Milk 0.0305 Water supply, sewage disposal in owner-occupied dwelling - 0.0120 School books - 0.1349 Total Housing Investment Subsidies 0.2147 Total Consumption Subsidies 0.0981 Total Subsidies 0.1467 Source: Central Statistical Office calculations, based on Incidence Study 1990. Notes: Services for vacations, canteen bonuses, culture and sport include those financed by enterprises for their employees. "Social items in kind" include noninstitutional care for elderly, handicapped, etc. and welfare homes. "Social policy allowance for housing" is a grant for down-payment on purchase of dwelling based on number of children in household. 38 CHAPTER 4 SOCIAL INSURANCE PENSIONS AND SICK PAY A. Introduction 4.01 The main deficiencies of the social insurance pension program are threefold. First, the system is financially unsustainable. Second, its effectiveness in maintaining a minimum living standard for the elderly is highly inadequate, as indicated by the sharp deterioration during the 1980s in the real value of pensions for most retired persons (Annex Table II.15.). Finally, the pension program is a source of allocative inefficiency in the economy - the system encourages workers to withdraw from the labor force at the earliest possible age, presents no incentives for individuals to save for their own retirement, and, by imposing an extremely high tax burden on employers, puts Hungary at a competitive disadvantage internationally as seen in Chapter 2. 4.02 In the absence of even minor reforms of the pension system, the worrisome trend in pension expenditure continued in 1990, exacerbated by the deteriorating economic conditions. In 1990, pension expenditures as a percentage of GDP rose by some 0.8 percent--an annual increase almost unmatched during the last three decades. As a consequence, the Social Insurance Fund (SIF) achieved only a small balance in 1990, compared to a surplus of some Ft. 22 billion in 1989. 4.03 Such a deterioration of the financial position of the SIF, of which the pension program is the main determinant, is clearly unsustainable for a number of reasons: (i) the already high contribution rate of 53 percent, which must be reduced in order to ease the distorting effects on the economy and on the labor market in particular; (ii) the additional pressure on employers to finance other programs, such as unemployment insurance and the rising health outlays, which start from a very low level; (iii) the restricted possibilities of the State budget to provide additional funds, for reasons of budgetary consolidation and the need to contribute financial resources in the area of social welfare; and (iv) the prospective aging of the population. 4.04 Since the central reason for the rapidly rising pension expenditure continues to be the growing number of retirees under the old-age and disability program, this Section concentrates largely on the policy options and financial effects of a reduction in the number of beneficiaries; more details of this analysis are provided in Annex IV. A comprehensive discussion of the current deficiencies of all three pension programs (old age, disability and survivors) and the available reform options is extensively dealt with in the IMF Social Security Report on Hungary.' 4.05 Section B below highlights very briefly the past expenditure trend of the pension program and the impact of the rising number of beneficiaries. I For details see the IMF (1990): Social Security Reform in Hungary, Fiscal Affairs Department, Chapter II. 39 Based on a long-term projection model, section C investigates the necessity and financial effects of reduced eligibility in the pension program in order to achieve a financially sustainable balance. What appears unavoidable is a substantial increase in the effective retirement age, essentially resulting from a rising standard retirement age and a change in the pension formula.2 Section D comments on the few measures envisaged for 1991 and proposes additional measures in order to contain the expenditure growth in the transitional period till the effects of a more comprehensive reform become effective. Section E addresses the necessary reform of sickness benefits, and Section F summarizes the recommendations. B. Pension expenditure and eligibility trends 4.06 Expenditure on pension benefits are without doubt the fastest growing expenditure item in the budget of general government in Hungary. Whereas at the beginning of the 1960s, pension expenditure as a percentage of GDP accounted for only 2.5 percent, this share almost quadrupled and is expected to reach 9.8 percent in 1990. This share is already well above the average expenditure share of the OECD countries which are characterized by a much higher income level and broader contribution base than Hungary. 4.07 A quantitative analysis of the determinants of past expenditure growth reveals that the dominant factor has been the increase in the eligibility ratio (i.e. number of beneficiaries compared to population aged 55 and over) (see Annex Table IV.1). The strong interrelation between pension expenditure and the rising number of beneficiaries is further highlighted in Figure 4.1, which graphically juxtaposes the trends in pension expenditure (as a percentage of GDP) and beneficiaries (as a percent of population aged 55 and over). The chart reveals not only the close relation between pension expenditure and beneficiary trends but also the steep upward tendency of both developments, exacerbated in recent years. 4.08 In 1990, the total number of beneficiaries of all pension programs (old age, disability and survivors) amount to 97.5 percent of the total population aged 55 and over. The vast majority of beneficiaries are eligible for old-age and disability pensions; therefore, their average (effective) retirement age must have decreased considerably over the recent years. Since the minimum retirement age (in Hungary equal to the standard retirement age) for old-age pension is 60 years for men and 55 for women, not only their effective retirement must have decreased close to the minimum age but also the effective retirement age of disability pension must have been substantially reduced, in particular in recent years. Such a development is consistent with the evidence that at present, 40 percent of all newly granted pensions are disability benefits. 2 The "effective retirement age" is the average age of retirement according to actual practice and reflects individual's response to regulations (e.g., on the legal minimum retirement age) and incentives to remain in the labor force. The "minimum retirement age" is the minimum age of eligibility for a public pension. The "standard retirement age" is the age at which full pension benefits can be received. 40 FIGURE 4.1: Hungary: Eligibility Ratio and Pension Expenditure Trends, 1960-90 100%- 10% -9% 90%X 8% 80%-~~~~~~~~~~ 80%~~~~~~~~~~7 0~~~~~~~~~~~~~~~~~~ B 70% 6% 'gb~~~~~~~~~~~~~~~~~~~~~~~~~' 60%- -4% 50%3 -3% 40%- , I I I I I I I I I I I I I I I-I I 2% 1960 1966 1972 1978 1984 1990 Year - Benefic.tPop 55 + -+- Pen.Exp./GDP 41 4.09 Even if there would be no further decrease in the effective retirement age for these two pension programs, the eligibility ratio is likely to continue to increase for some time till the higher inflows of beneficiaries per age cohort are distributed over all pension ages. Yet, and in addition, a further decrease in the effective retirement age must be expected if the current policy with regard to the awarding of disability benefits continues and the new pre-pension option for unemployed is widely implemented (para 5.32). Consequently, without major reforms of the pension program and strong measures which dramatically reverse the trend in the effective retirement age, the eligibility ratio will continue to increase for some time. Even if a stable eligibility ratio is finally reached, the aging of the population, which accelerates again soon after the turn of the century, will put additional pressure on the pension system. C. Financial scenarios and retirement reform options 1. Projections and magnitudes 4.10 In order to quantify the financial operations of the pension program under a baseline scenario (no reform) and under different retirement age scenarios, a long term projection model is used.3 The assumptions of the baseline scenario are summarized in Annex Table IV.2. It assumes a further increase in the total eligibility ratio till 2000, essentially through rising ratios of old age and disability pensioners, and constancy thereafter. For old- age pensioners this implies a further decrease in the effective retirement age from 59.7 in 1990 to 57.6 years in year 2000. As regards the other major determinants of the model, it is assumed that from the year 2000 onward, full price indexation of pensions is granted. All calculations are based on the medium demographic scenario (see Chapter 2). The annual real growth rate of GDP per employed is assumed to rise to 4 percent from 2000-on, although the implications of lower growth (2 percent per annum) were also explored (para. 10.10). 4.11 The financial result for the baseline scenario is presented in the uppermost line on Figure 4.2. The target indicator is the implicit contribution rate, that is, the rate of taxation of the wage bill required in each year to cover the financial operation of the pension program. The implicit contribution rate for the pension program alone amounted to 30 percent of the wage bill in 1990 (out of the total social insurance contribution rate of 53 percent). 4.12 The baseline scenario starkly demonstrates that an unreformed pension system is unsustainable. On its own, an unreformed pension program implies that the required contribution rate would dip slightly during the mid-1990s but return by 2000 to the 1990 level. Thereafter, the contribution rate would soar to 36 percent by 2006, to 51.6 percent in year 2036 when the population aging is expected to peak, and then experience a slight decrease reflecting some easing of demographic pressure. Clearly, such rate levels are inconsistent with a market oriented economy and would result in major economic distortions and flourishing parallel market activities. It has to be kept in mind that this 3 For details see IMF (1990), op.cit., Chapter IX and Appendix II. 42 Figure 4.2: Implicit Contribution Rate of Pension Program Under Alternative Retirement Scenarios 52% Base Case 47% - scenario I 42%- 37% 32% 27% 22% 1989 2000 2011 2022 2033 2044 Note: For derivation of Scenarios, see Annex Table IV 3. 43 contribution level does not include the financing requirements for other programs of social insurance such as unemployment benefits, the implications of which are examined in Chapter 5. In summary, even allowing for some mis-specification of the model, the magnitude of the projected cost of the existing pension system is in total contradiction to any market oriented economy and calls for a rapid and comprehensive reform as well as strong transitional measures. 4.13 The following analysis investigates the impact of alternative increases in the effective retirement age on the implicit contribution rate of the pension program; the measures to achieve such an increase are also outlined below. The detailed assumptions of the alternative scenarios are provided in Annex Table IV.3. To fit such an investigation into the model structure, it is assumed that the awarding of disability pensions is restricted considerably, resulting in a long-term decrease in the eligibility ratio of disability pensions from 22 percent in 1990 to 18 percent by 2021. The decrease of the total eligibility ratio is achieved primarily by an increase in the effective retirement age for old-age pensions. Five alternative policy scenarios are considered, which are differentiated by the steepness of the assumed decline in the total eligibility ratios and increase in effective retirement age. 4.14 Returning to Figure 4.2, it is seen that in order for the implicit pension contribution rate to stay roughly constant on average over the fifty- year projection period (Scenario V), a substantial change in the effective retirement age for old age pensioners has to take place (besides a tightening for disability pensions). Already by the year 2000 an increase in the effective retirement age by 2.7 years is required. Then, and until 2021, a total increase in the effective retirement age by almost 8 years is necessary, supplemented by an additional increase of around one year to the end of the projection period. As to be expected, the lower the projected long-term rate of economic growth, the larger will be the required increase in the effective retirement age to maintain the implicit pension contribution rate at about its present level. 2. Measures to increase the effective retirement age 4.15 In the past, the Hungarian authorities have avoided any serious consideration of an increase in the retirement age. This attitude is partly based on the given pension system in which the minimum retirement age coincides with the standard retirement age. Thus in the present circumstances, enforcing a one-year delay in retirement, for example, simply means that individuals would forego one full year of benefit, which would obviously be highly unpopular. However, many other countries do not provide full pension benefits at the legal minimum retirement age. If Hungary were to introduce such a distinction, an increase in the effective retirement age would not necessarily imply an increase in the minimum retirement age, at least over the next one or two decades. It would require, however, a change in the pension formula to induce later retirement as a deliberate individual decision. In Hungary, in order to reduce the incentives for retirement at the earliest possible age, it also requires a consistent change in the tax treatment of contribution and pension benefits. The remainder of this chapter elaborates on these issues and makes appropriate recommendations. 44 4.16 The standard retirement age in Hungary (55 years for women; 60 years for men) is one of the lowest in the industrialized world. Although the health status of Hungarian workers is notoriously bad, and, in consequence, morbidity and mortality of the elderly are very high, an improvement must be assumed over the next decades as a result of the changing economic structure and life styles. Such an assumption also underlies the demographic projections provided by the Hungarian authorities. Hence, an increase in the effective retirement age is feasible and should be based on the following considerations: (i) harmonization of the minimum retirement ages for men and women; (ii) introduction of a common and higher standard retirement age; and (iii) reform in the benefit formula. 4.17 The harmonization of the minimum retirement age through raising the retirement age of women to that of men should take place in a very short period, say, till 2000. Starting in 1992, the minimum retirement age for women should be increased by 1 year every two years. As a result, such a harmonization may contribute to an increase in the effective retirement age (for men and women) by some 1.5 to 2 years.4 4.18 The increase in the standard retirement age for both men and women should complement the harmonization and lead to a further increase in the effective retirement age. To this end, it is recommended to raise the standard retirement age, also starting in 1992, by 1 year every 4 years to reach 65 by 2008. If supplemented by an adequate reform of the pension formula, such a change will achieve the required increase of the effective retirement age by some 4 years over the same period. Whether a further increase in the effective retirement age will be necessary in the following years depends on the structure of the pension formula at the time, the wage rates for elderly (which determine their opportunity costs of retirement), the employment chances of the elderly and other economic, social and demographic variables. There is some likelihood that a further increase in the minimum and standard retirement age by some two years will be necessary. Such an envisaged change may be announced at an early stage, but finally postponed if the financial operations of the pension program are eventually better than projected. 4.19 If the minimum retirement ages for women and men are not harmonized in the proposed period, this would require an even faster increase in the common standard retirement age, reaching 65 years already in 2000 and further increases by some 3 to 5 years during the following decade. 4.20 In order to be financially and economically effective, an increase in the standard retirement age has to be supplemented by a comprehensive reform of the pension formula. The aim should be to make the benefit formula as neutral as possible in order to reduce distortions in labor supply and retirement decisions. To this end, the following major changes are recommended: (i) use of lifetime income as a pension base (with appropriate indexation of past 4 The exact impact in years depends on the distribution of retirement ages for men and women on which no data was received. It is understood that as of mid-1991, the Hungarian authorities were considering a similar approach of raising the minimum retirement age for women by one-half year every year, starting in 1993. 45 earnings); (ii) adoption of a linear accrual rate, with actuarial decrements and increments for earlier and later retirement (measured from the standard retirement age); (iii) reduction in noncontributory service periods; and (iv) introduction of a ceiling on contributions and benefits.5 3. Tax treatment 4.21 Concurrent to the change in the benefit formula, it is also recommended to eliminate the inconsistent tax treatment of contributions and benefits through an inclusion of benefits into the personal income tax base (for retirees) while excluding the contributions of the population from the tax base (for employed). Such a solution would substantially reduce the distortions which the current rules imply for individuals' incentives to retire. However, reforming the tax treatment of pensions would likely result in net expenditure for the general government finances. A full evaluation of the net expenditure effects would require detailed knowledge of the income distribution of both employed and retirees. Preliminary estimates in (Annex IV) suggest that the net expenditure effect in the short run would be minimal, but would rise to a significant level in the longer run as all retirees received pension benefits based on gross earnings. In order to compensate for this net expenditure, some additional adjustment in the pension formula will be needed to reduce the pension level for new retirees. The disadvantages of such a reduction would be outweighed by permitting a more equitable and efficient pension system. 4. Other issues concerning implementation of the new system 4.22 During the introduction of the above reforms, including in particular the reduction in noncontributory service periods and provision of more actuarially-based pensions, some special provisions would be required for certain groups of individuals. A more actuarial pension system, for example, would imply lower pension levels for some retirees. In the short run, for those individuals whose pension would thereby fall below a minimum subsistence level, the government may wish to make up the difference (provide a guaranteed minimum pension compensation) which should be financed by the State budget rather than by the SIF, although the SIF could disburse the payment. This provision should be developed as part of the social assistance system rather than as part of social insurance, and further discussion of this proposal is therefore left to Chapter 8. In the longer run, the revamped pension system could have a two- tier structure: a flat-rate component which, at least at the beginning, can be proportionate to the service years (i.e., flat-rate component equals flat-rate amount times years of service), and a strictly earnings-related part (based on lifetime income and linear accrual rate). Later on, the flat-rate component can become independent of service years or even universal, thus replacing the minimum income guarantees proposed above. Such a benefit structure is a very easy way to introduce distributive features while minimizing the distortionary impact on labor supply decisions. The taxation of benefits would also create no problems with such a structure. 5 For details, also with regard to disability and survivor's pensions see IMF (1990), Chapter II. 46 4.23 The new system should also be introduced in such a way as to take account of the varying ability of different cohorts of workers to adjust employment and savings behavior. The exact transitional approach could only be worked out once the full structure of the future pension system is determined, but the following phasing is one possibility. Individuals within five years of the minimum retirement age at the time of the reform could be allowed to retire under the old system (except for women being subject to the increasing minimum retirement age) or the new system, if more advantageous. For individuals between five to fifteen years from retirement, specific transitional arrangements would be established. For all younger persons, the new system would become applicable. D. Transitional measures prior to the major pension reform 4.24 In order to ensure the short-term financial security of the SIF, a number of corrective measures have been made or are under consideration for 1991. Early retirement provisions for the unemployed have also been introduced in 1991, financed from the new unemployment insurance fund (Solidarity Fund). This section comments on these measures and recommends additional actions. 4.25 With respect to the increase in the qualifying period from 10 to 20 years (already decided in 1981 to be implemented in 1991), this measure will certainly increase the pressure to stay on the labor market. Furthermore, in view of the non-actuarial character of the current system, there is some rationale for a long qualifying period in order to avoid speculation and exploitation of the system. However, a gradual phasing-in of this measure would have allowed a better adjustment by the individual pensioners concerned. The 20-year minimum vesting period will result in several thousand persons (mainly women) becoming excluded from any pension; for this reason, the authorities have reintroduced a partial pension in 1991 for persons employed between 10 and 20 years.6 It should be noted that a more actuarial structure for the pension system would not have required a lengthening of the minimum insurance period. Hence, if suitable reforms were made in the retirement age and benefit formula, the relatively high qualifying period for a full pension could possibly be reconsidered. 4.26 The limitation of the pension level is certainly necessary in view of the growing wage differentiation and increasing possibilities of personal retirement savings by the higher income groups. Pensions can be limited either by an absolute ceiling on the amount of pension or by a degressive formula for pension calculation (i.e., the share of the earnings base used to calculate the pension declines more sharply with higher income levels). In 1991, the 6 The partial pension is calculated as follows: the pension which would be received after 20 years of employment is reduced by 2 percent for each year the retiree falls short of the 20-year minimum (e.g., with only 10 years of employment, the retiree would obtain 80 percent of the normal pension). 47 authorities have made the existing degressive formula steeper.7 With this method, it is essential to provide adequate indexation of the income-limits in order to avoid distortions between pension cohorts. In order to achieve neutrality of incentives with regard to labor market supply and retirement decisions, the present non-actuarial character of the pension system has to be reconsidered as well. Finally, whatever solution to limit pensions is chosen should be consistent with the envisaged reform in the pension formula (which should include a lengthening of the assessment period to cover lifetime income). 4.27 The proposed one-year or even two-year postponement of indexation for new benefits may be a convenient short-term financial measure. It has, however, various pitfalls. First, in view of the expectation that inflation in 1991 may exceed 30 percent, a universal postponement for all pension levels may not be acceptable. Second, it creates immediate inconsistencies between consecutive pension cohorts (which retired before and after the adopted threshold date). In consequence, after one year of high inflation, a newly retired person may be much worse off than somebody with equivalent retirement earnings who retired two years earlier and whose pension is partially indexed. Finally, to achieve a lower pension expenditure level, direct measures such as reduced eligibility (discussed above) or lower nominal replacement rates are preferable. 4.28 The shifting of non-contribution based benefits to the central government budget has some rationale under a reformed system in which a clear separation between insurance and assistance-type benefits may be made and in which the remaining benefits under the SIF reflect clear, consistent and rational choices. Without that, the measure would just shift expenditure within the consolidated central budget and is no substitute for real reforms. 4.29 Even if all these measures were implemented, their financial impact is likely to be very modest. Consequently, further measures are needed for 1991 in order to permit even a slight easing of the implicit contribution rate of the pension program in the next couple of years, before the impact of the major structural reforms is felt. Such measures may comprise (i) an effective flat- rate indexation of all pension benefits throughout 1991; (ii) a tightening of disability benefits; (iii) the (re-)introduction of retirement income testing; and, (iv) financial incentives to postpone retirement. 4.30 As regards indexation, the Mission recommends a flat-rate increase for all beneficiaries. This flat-rate amount could be fixed at a level which compensates the low-income beneficiaries for increases in prices in 1991, in particular food, energy and rent, based on a representative basket for this group. During the period of high inflation, the flat rate should be calculated, 7 The degressivity formula worked as follows up to 1991: for monthly preretirement incomes below Ft. 10,000, 100 percent of income was taken into account in calculating the pension; for each increase of the income level by Ft. 2000/month, the share taken into account decreased by 10 percentage points, to a minimum of 60 percent for incomes over Ft. 16,000. As of January 1991, this schedule is extended, such that for each Ft. 10,000 increase in incomes over Ft. 30,000 the percentage is reduced by 10 points, reaching a minimum of 5 percent for incomes over Ft. 80,000/month. 48 and paid quarterly, based on the programmed price changes, with adjustments in the following quarter once the actual price development is known. The Mission was informed that the authorities have proposed to base the indexation on the average increase in net wages (but less than the rate of overall inflation); under debate is whether the indexation would be flat-rate for all pensioners or variable according to the level of the pension. It should be recognized that the latter option would increase the discontinuities among pension levels, as mentioned in para. 4.27.8 4.31 The tightening of disability benefits should be based on three principles: first, the application of medical criteria only in the determination of full disability (67 percent disabled) and qualification for disability benefits; secondly, a more frequent review and monitoring of already granted benefits; and third, the speedy elaboration and implementation of retraining and rehabilitation facilities for disabled at the earliest possible stage of disability. The Hungarian authorities have proposed that persons qualified as between half and two-thirds disabled would receive less than a 100 percent disability pension, but would receive a special allowance or wage subsidy which would supplement any income they could earn from work. (Under the current system there exists only a full disability pension and recipients are not permitted to work.) The Mission fully supports this reform proposed for introduction in 1991. 4.32 The retirement test was discontinued in 1988, apparently for political reasons. However, in the absence of a retirement test, a low retirement age and a non-actuarial structure present a clear incentive to retire at the earliest possible age and to continue labor market participation while receiving full pension benefits. Admittedly, in many cases, continued employment is necessary to supplement a low benefit level, but the existing incentive to retire early is also one of the main reasons the pension program is in such precarious financial shape. It has been argued here that reforms in the pension formula are needed to remove the incentives for early retirement, including through an actuarial reduction in pensions taken before the standard retirement age. Until this change is introduced, retirement income testing will be necessary to discourage individuals who can continue to work from taking their full pension before the standard retirement age. With such a test, persons having higher labor earning possibilities will have to make a choice between full labor force participation (with substantially reduced pension benefits, or no benefits at all) and full retirement. The re-introduction of an appropriate retirement income test as a transitional measure is therefore strongly recommended. 4.33 The individual decision to stay on the labor market may be enhanced through pension increments when retirement is postponed. Actuarial increments (i.e. compensation for postponed retirement which keep the present value of benefit expenditure constant) in the age group 60 to 65 are in the order of 6 8 Concerning compensation for inflation in 1990, the Parliament in November 1990 adopted a plan to grant persons with pension below Ft. 5000/month a lump sum payment of Ft. 4000 in December 1990; pensions between Ft. 5- 10,000/month were awarded Ft. 3400, and between Ft. 10-13,800 were awarded Ft. 2400. Pensions over Ft. 13,800 were not given any lump sum increase. 49 to 8 percent per year. Hence, it is recommended to increase the pension level by, say, 5 percent for every year of postponed retirement for those with, say, 35 years of service at standard retirement age. The short-term saving could be substantial if an important share of elderly employed would use this option. The long-term saving could still be noticeable. 4.34 The expansion of early retirement benefits for the unemployed in 1991 constitutes a substantial potential expenditure burden for the central government. The experience in OECD countries in the 1970s and 80s with similar provisions reveal that the measure is very expensive (compared to alternative means of providing income support to the elderly unemployed), induces special incentives for employers to separate older employees, and reduces also the incentives for older unemployed to search for work (although, admittedly, with lower chances). In view of this experience, in recent years, many European OECD countries have reduced or tightened special early retirement provisions.9 It is recommended that the proposed program in Hungary be reconsidered carefully, and at least not adopted in the general and largely untargeted form as currently drafted (see also discussion of this issue in Chapter 5). E. Reform of Sick Pay'0 4.35 Hungary has one of the highest numbers of compensated sick days per insured in the industrialized world, averaging 29 per year. Besides the admittedly substantial health problems of many Hungarian workers, the incentive structure of this program is also likely to be responsible--mainly, the provision that employers pay for only the first three days of sick leave, with remaining leave (up to a maximum of one year) financed by the SIF. 4.36 A number of proposals for reform of sickness benefit have been under discussion in Hungary. The most effective reform, which would be simple to implement, would be to lengthen considerably the period of sickness for which the employer has financial responsibility. To make practice in Hungary more consistent with that in many other industrialized countries, the employer should be made liable for the full duration of normal sick leave (30 calendar days or four weeks), and the employee could also be asked to bear the cost of an initial few days of each episode. These changes would make both employees and employers more conscious of the costs of unhealthful behavior and lifestyle and reduce the present inordinate use of sickness benefit. Doctors should also be monitored in their approvals of sick leave. 9 For a review of early retirement benefits in OECD countries, see Holzmann, R., "On the Relationship Between Retirement and Labor Market Policies," in Structural Problems of Social Security Today and Tomorrow (Leuven: ACCO 1988). 10 Persons covered under social insurance are entitled to up to one year of sick pay, normally equivalent to 75 percent of average earnings of the preceding year; the replacement rate may be lower if insurance coverage has been interrupted for extended periods, or higher for illnesses longer than 60 days and for occupation-related illness or injury. 50 4.37 Increasing the employer's direct responsibility for sick pay from three days to four weeks would require only longer periods of sick leave to be financed by the SIF; this change would enable the implicit social insurance contribution rate for sick pay to be reduced from 3.5 percent of payroll in 1991 to about one percent in 1992. It would be reasonable to introduce some differentiation in the sick pay-related social insurance rate to recognize the different risks of illness and injury in various activities, and to provide an incentive for enterprises to minimize these risks as well as to enforce only appropriate uses of sick leave. Consideration should therefore be given to varying the remaining social insurance premium for sick pay to reflect experience - e.g., enterprises with a sick leave record above some norm, say, 20 days per employee per year, could be charged a higher contribution rate. F. Conclusions and Recommendations 4.38 These may be summarized as follows: - To secure the financial viability of the pension program into the middle of the next century, a tightening of eligibility through an increase in the effective age of retirement is essential. This change would not necessarily require an increase in the minimum retirement age for both sexes, at least over the next one to two decades. A tightening of eligibility for disability pensions is equally necessary. - Increasing the effective retirement age will require (i) harmonization of the minimum retirement ages for men and women, (ii) introduction of a common and higher standard retirement age (when full pension benefits are granted), and (iii) a comprehensive reform of the pension formula to reduce distortions in labor supply and retirement decisions, and specifically to induce later retirement as a deliberate choice of individuals. - The tax treatment of contributions and benefits should be made consistent through inclusion of benefits into the PIT base for retirees and excluding contributions from the tax base for the employed. - Transitional measures are needed in the current year to ensure the short term health of the SIF as well. The most important of these measures include flat rate indexation of pensions for 1991, the reintroduction of retirement income testing, and the introduction of financial incentives to postpone retirement. - Reform of sick pay should involve shifting a very large share of the costs to employers (e.g. the first four weeks of sick leave). The social insurance contribution covering the remaining periods of sick leave could be differentiated to recognize and reward efforts by different enterprises to reduce risks to health and safety. 4.39 The financial implications of the above reforms in eligibility and in the benefit formula for disability and retirement pensions and sick pay can also be summarized here. Comparing the projected "Base Case" with the most radical reform scenario (Alternative V), the latter would result in an implicit pension contribution rate which is lower by 2.5 percentage points in 1994, 6 51 points by 2000 and 17 points by 2041. Against this reduction, the recommended reforms of tax treatment would result in a net expenditure effect which may require 1.5 percentage points to be added to the pension contribution rate in the short term (the next 10 years or so), rising to possibly 5 points in the longer run (when all retirees are receiving pensions based on gross earnings). The recommended reform of sick pay has been estimated to reduce the implicit sick pay contribution rate by about 2.5 percentage points from 1992. The net effect of these changes would be a reduction in the combined pension and sick pay insurance contribution rate (compared to the no-reform scenario) by about 3-4 percentage points as early as 1994, with a significantly larger reduction from 2016. (See Annex Tables X-2 and X-4 for details of projections through 2000.) This estimate does not include any further reduction that might be achieved in 1992 by the shift of some noncontribution-based pension benefits to the consolidated State budget. 52 CHAPTER 5 SUPPORT FOR THE UNEMPLOYED A. Introduction 5.01 As mentioned in Chapter 2, registered unemployment in Hungary in January 1991 amounted to about two percent of the labor force (some 100,000 persons). This was nearly four times higher than in January 1990, but it is recognized by all concerned that far higher unemployment rates will be experienced over the next two or three years. Expenditure on unemployment benefits and employment services (The Employment Fund) amounted to only Ft. 2.4 billion in 1989, rose to Ft. 10 billion in 1990, and is budgeted at Ft. 29.6 billion in 1991; thus, at the present trend it is becoming one of the largest transfer programmes in Hungary. It is against this background that the current system of support for the unemployed must be assessed. 5.02 Experience in developed market economies has shown that effective programs to address the employment dimension of economic restructuring combine reactive and proactive measures. Reactive measures ameliorate the short-term costs of economic adjustment to individuals through income support, and include unemployment compensation as well as social assistance. Proactive measures seek to add value to human capital through development of new skills, thereby facilitating investment and redeployment of labor to newly created jobs. These programs include employment services which respond to frictional unemployment and assist with redeployment of labor, as well as training and employment generation to address structural unemployment. 5.03 Hungary has had a basic system of unemployment compensation and proactive labor market programs for two years which has provided the authorities with some important experience in the administration of such schemes, albeit at a much lower unemployment rate than will be faced in the future. A major reform of the unemployment support system has been introduced in 1991 as part of a new Employment Law. The functioning of the former system and the reforms are evaluated below. Both the reactive and proactive measures should meet the test of effectiveness (in providing income maintenance and facilitating labor redeployment, respectively); efficiency (in directing benefits to those in need and creating appropriate incentives for labor supply and demand); and financial sustainability. The aims of income support and an efficient labor market are often seen as conflicting, implying a "trade-off" view of unemployment benefit in which more generous support reduces the incentive to work. However, generosity in the benefit system can help as well as hinder the operation of the labor market, by increasing the incentive to accept redundancy and enter unemployment. Successful redeployment of labor depends as much on the proactive measures as on the incentives provided by the benefit system. Both types of measures are ultimately constrained by the impact on the budget and the sustainability of payroll taxation as a source of revenue. 53 5.04 A full assessment of the unemployment benefit system should take into account other legislative reforms affecting household incomes and the labor market. An example of the former is the development of a more comprehensive social assistance program (discussed in Chapter 8), which could provide income support to those with no unemployment insurance cover. This Report does not discuss in detail the proactive labor market programs in Hungary to date (retraining, wage subsidies, start-up loans, job-creating investment, public works employment, employment services), for the sole reason that other recent reports have treated these measures in some detail. (World Bank, 1990a; IMF, 1990) 5.05 Part B of this chapter discusses the emerging picture of unemployment in Hungary. Part C briefly reviews the former unemployment benefit system and Part D comments on the major changes introduced by the recent legislation. Issues which remain under the new system are discussed in Part E, including the treatment of part-time work and early retirement. The costing and monitoring of unemployment benefit is addressed briefly in Part F. Part G concludes and summarizes recommendations for reform. B. Unemployment in Hungary 5.06 Unemployment insurance works best if unemployment is evenly distributed across the working population; however, this is rarely, if ever, the case. There is already evidence of a higher incidence of unemployment among certain groups of the workforce in Hungary and in particular areas of the country. Unskilled manual workers accounted for 1 in 3 of all spells of unemployment benefit receipt in 1989 but for only 1 in 10 of the workforce. Half of benefit recipients were aged under 35 in December 1989 compared to about 40 percent of active earners as a whole; 50 percent of recipients at this time lived in three counties which had only 17 percent of total employment.' These patterns may change as unemployment rises but there is unlikely to be an equal probability of unemployment across the labor force. 5.07 Three groups may be especially vulnerable. Firstly, new entrants to the labor market. The prospects for persons attempting to get their first job may be expected to be particularly poor in the period of economic transformation, especially for those with a low level of education and training (note the collapse of the youth labor market in various OECD countries in the early 1980s following the second oil shock). Their position may be aggravated if employment law or social pressures make it difficult for managers to adjust workforces through dismissal, since employers will then tend to reduce their workforces by cutting down on new recruitment and relying on natural wastage. I Figures from Matoricz (1990), Statistical Yearbook 1989 Tables 4.1 and 32.7., and Statistical Pocketbook, 1989, pp. 27, 29. 54 5.08 Secondly, older workers made redundant during economic re- structuring may find it hard to find new jobs. The experience of various OECD countries has been that the probability of leaving unemployment declines with age, leading to longer durations of unemployment and a concentration of the elderly among the long-term unemployed. 5.09 Thirdly, there will be those who suffer from recurrent unemployment. There is no reason to believe that the usual pattern in Hungary will be a single spell of unemployment followed by several years of stable employment. As in other countries, a sub-section of the labor force may experience repeated spells of unemployment.2 The development of new private sector firms, some of which will fail, may re-enforce this trend, as will the development of temporary jobs, a prominent feature of Western European labor markets. C. The Former Scheme and Pressures for Reform 5.10 The principal features of the unemployment benefit scheme in existence through March 1, 1991 are given in Annex Table V.1. In essence, the scheme paid at most two years of benefit to someone who had spent at least half of the previous three years in employment. In the first year, payments (Unemployment Benefit) were related to past earnings while the second year provided a flat-rate benefit (Temporary Allowance). Voluntary quitters also received benefits, but at a lower rate than persons laid off. Funding was from the state budget via the Employment Fund. Although this scheme has now been replaced, certain key features are discussed below, against which the new unemployment benefit program can be compared. 5.11 In September 1990, one-third of the registered unemployed were in receipt of neither the basic unemployment benefit nor the extended allowance (Table 5.1), although some of this group were involved in the proactive labor market programs. There are no statistics at the national level on the different reasons for non-receipt of benefits (and this is itself a cause for concern). However, it is believed in the Ministry of Labor that the most important reason was the lack of an adequate employment record, and there was no provision for income support in such cases either from the Employment Fund or from social assistance. 2 The proportion of persons experiencing unemployment in 1983 who had more than one spell during the year was 21% in Australia, 24% in Sweden, and 32% in the USA. The proportion of all weeks in unemployment accounted for by multiple spells in these three countries was 21%, 29% and 39% respectively (OECD, Emplovment Outlook, 1985, Table 37). In the UK, one-half of the male inflow into registered unemployment in Autumn 1978 had a spell of registered unemployment in the previous 12 months (Moylan, Millar and Davies, 1984, p.24). 55 Table 5.1: Number of Unemployed and Beneficiaries of Support Programs 1989 September 1990 (Actual figures) Total during year Stock Total Registered Unemployed 28,500 (100%) 57,091 (100%) (as percent of labor force) (0.6%) (1.1%) of which: Received Unemployment Benefit 12,000 ( 42%) 35,919 ( 63%) Received Temporary Allowance none 1,660 ( 3%) Other* or no support 16,500 ( 58%) 19,512 ( 34%) end-1990 (stock) end-1991 (stock) (9/90 projection) (2/91 projection) Total Registered Unemployed 100,000 (100%) 200-250,000 (as percent of labor force) (2.0%) (4-5%) of which: Obtain Unemployment Benefit 76,000 ( 76%) (75-80%) Obtain other* support 28,000 ( 28%) *Breakdown of Employment Fund Programs. 1990 Ft. bn #i beneficiaries (flow) Unemployment benefit/Temp. Allowance 3.0 na Interest subsidies for Start-up Loans 2.5 ** Retraining 1.0 30,000 Subsidies to regions with large unemployment and for job-creating investments 1.0 ** Public works employment 1.0 4,050 Early retirement 1.0 4,200 Wage subsidies for young professionals 0.2 ** Employment services 0.3 na Total 10.0 **subtotal of new jobs created - 10,000 Source: Actual figures supplied by Ministry of Labor, projections from National Labor Market Center. 56 5.12 The feature of the past scheme which would have particularly affected the recurrent unemployed was the restriction of the basic benefit to a total of 12 months in any three years. A person who had a spell of 9 months of benefit receipt would have been able to draw unemployment benefit for a total of only 3 months in the next two and a quarter years; entitlement would soon have expired if subsequent unemployment occurred. 5.13 The very low numbers of persons receiving the Temporary Allowance suggests that expiry of entitlement was not a significant problem in 1990, but this could change as the level of unemployment rises, leading to longer durations and increased frequency of unemployment spells.3 The high probability of denial of cover to those recurrently unemployed through no fault of their own was thus an undesirable feature of the previous system. D. The New Unemployment Insurance 5.14 Annex Table V.2 provides some of the details of the unemployment support scheme introduced under the new Employment Law passed in February 1991. The main form of compensation is provided as unemployment insurance and funded through earmarked payroll contributions to a new Solidarity Fund (1.5 percent paid by employers and 0.5 percent by employees). These payroll deductions come into effect July 1, 1991. The State budget is providing Ft. 9 billion to the Solidarity Fund for its outlays during the first half of 1991, as well as Ft. 12.6 billion to the Employment Fund, which will continue to finance labor market programs other than cash compensation (new benefits for career beginners and pre-pensions - see below - will be covered by the Solidarity Fund). To the extent that job security goes with better pay, the employee contributions will redistribute from persons with higher earnings to those suffering unemployment. The most important changes in unemployment benefits compared to the former scheme are summarized below. 1. Coverage 5.15 Four principal changes have been introduced. Firstly, eligibility conditions,are significantly relaxed, benefit being payable with a minimum of only 12 months work in the previous 4 years compared to the former condition of 18 months in 3 years. 5.16 Secondly, the duration of benefit varies with the employment record. The minimum period of entitlement after 12 months of employment is now six months. The maximum benefit period of two years would apply only to someone who had spent all of the previous four years in insured employment, and the Temporary Allowance is phased out. Whether these changes represent on balance a reduction or increase in coverage compared to the former system is difficult to say without knowing the employment profile of the persons currently facing unemployment. The coverage of repeat unemployment is apparently greater under the new scheme, since persons can requalify for the minimum six-month benefit after one year of employment regardless of their 3 The average period on unemployment compensation increased from 95 days in the first half of 1989 to 128 days in the first nine months of 1990. 57 previous unemployment spells, and the former limit on benefit entitlement within any three-year period is removed. However, it should be noted that the authorities' expressed assumption that 75-80 percent of the stock of registered unemployed will continue to obtain benefit coverage is very optimistic compared to the experience of other countries.4 5.17 Thirdly, a new benefit ("unemployment assistance") for career beginners is proposed, payable for six months without any employment history requirement. This increases access to unemployment benefit for a large group previously excluded. 5.18 Fourthly, under the new law persons quitting their jobs voluntarily would be disqualified from unemployment insurance for 90 days, after which they would be eligible for the same benefits as other unemployed workers. This is an important change to curb overall demands on the limited insurance funds. 2. Benefit Levels 5.19 During the first half of the entitlement period, the insurance benefit is still 70 percent of the earnings during the last 12 months of unemployment; for Phase II, the benefit will be 50 percent of the same earnings base (in contrast to 60 percent under the old system). The maximum benefit is three times the minimum wage as before, but the minimum benefit has been raised to equal the minimum wage, whereas the previous floor was only 80 percent of the minimum wage. Here too, without data on the profile of claimants, it is not possible to determine whether the average unemployed person would fare better under the new or the old system in terms of level of benefit. In the Mission's view, it would be preferable in any case for the benefit to be stated as a flat-rate at least during the Phase II entitlement period. Although this would not necessarily entail cost savings to the Solidarity Fund, it would involve less administrative burden on both employment offices and employers (who at present have to provide the data on past earnings) than an earnings-related benefit. E. Evaluation of the new unemployment benefit system 1. Effectiveness of income maintenance 5.20 On the face of it, the initial gross replacement rate of 70 percent (gross benefit as a percentage of gross previous earnings) offers a reasonably generous level of income support. There are several reasons why the net replacement rate will in general be somewhat different, due to the For example, in Britain and West Germany, only one-quarter and two- fifths, respectively, of the registered unemployed stock were in receipt of insurance benefit in 1988 (Micklewright, 1990). See also Annex V. 58 impact of tax and social insurance deductions.5 Moreover, the gross earnings base used in the benefit calculation is not indexed for inflation.' The absence of any indexing has a negative impact on effective replacement, the purchasing power of the benefit being lower than suggested by the gross replacement rates. The only indexing in the unemployment benefit scheme is of the minimum benefit, in that this is based on the prevailing minimum wage. 5.21 Aside from the ratio of benefit to earnings, the absolute level of benefit payments is relevant to the discussion of both living standards and incentives. Looking at the experience with the former scheme during 1990, the average unemployment benefit was Ft. 5,700 per month; this may be compared to the minimum wage in August 1990 of Ft. 5,600 and the previous figure of Ft. 4,800 set in February.7 In April 1990 over 40 percent of Unemployment Benefit recipients were receiving payments of less than Ft. 4000 per month (IMF, 1990, p.90). These levels of benefit may also be compared with the yardstick of Ft. 5,122 per capita for the "minimum subsistence" monthly income for households with active earners in 1990 (based on the 1989 CSO estimate inflated to 1990). 5.22 These actual benefit figures suggest that, in practice, many unemployed people were receiving payments that provided them with only a rather modest standard of living. This picture needs to be modified, however, to take into account the presence of other income and other persons in the claimant's household. 5.23 If the claimant has a job in the second economy which continues during unemployment, then this will contribute to living standards while unemployed (and the ratio of total income received to that when fully employed is higher relative to the situation where there is no second job). On the S For example, the social insurance deduction during unemployment is only 5% of benefit rather than the 10% rate on earnings, thus raising the net rate above the gross rate. Moreover, although benefit is subject to income tax, the average rate of tax on benefit and on earnings may differ. On the one hand, the Ft 1,000 per month additional tax allowance for employees may no longer apply when unemployed. If this is the case, the effect will be to reduce the net replacement rate in relation to the gross rate. On the other hand, benefit may be taxed at a lower marginal rate than earnings thus raising the net rate; this would occur if the gross benefit falls in a lower rate bracket than the gross earnings. It should be noted that the net replacement rate would be the same for all if benefit was calculated as a percentage of net earnings, as is the case for example in Germany. a Benefit is calculated on the basis of average monthly earnings during the last 12 months of employment. If the time with the last employer was less than 12 months, then it is the monthly average over the period for which the job was held which is relevant. It is not clear how the benefit would be calculated if the individual did not work during the twelve months immediately prior to the claim. 7 The income tax threshold for employees in 1990 was very close to the minimum wage of Ft. 5600/month. 59 other hand, if the second economy job is associated with the first economy workplace, unemployment may lead to the loss of both jobs; living standards when unemployed will then be determined by.unemployment benefit alone (the drop in income in this case will of course be greater than if no second job was held). 5.24 The presence of other persons in the claimant's family or household (the distinction is discussed in the chapter on social assistance) may influence the unemployed person's living standards either way. For example, assume that an unemployed man receives Ft. 5,000 (net) per month, with previous net earnings of Ft. 7,000. If he has a wife earning Ft. 5,000, then the household replacement ratio rises from 70 percent (in the one-earner case) to 83 percent. With two children and a family allowance of Ft. 4,140 (in the first half of 1990), the replacement rate rises again to nearly 88 percent. 5.25 These sorts of calculations are illustrative of the possible differences which household composition and other sources of income can make. However, they are not a sufficient basis on which to judge policy reform. Experience in OECD economies has shown the importance of using survey microdata on a representative sample of the unemployed containing full information on their family and household characteristics and the different income sources. A pilot labor force survey is being conducted by the CSO in 1991, on the basis of which more detailed analysis should be possible of the nature of unemployment and the effects of unemployment support. 2. Incentives to Work 5.26 As far as the incentive to return to work from unemployment is concerned, the relevant ratio is that of benefit to prospective earnings, and not past earnings. For some people, past earnings may be a good guide to future earnings, in real terms. But the lack of indexing implies that the ratio of benefit to prospective earnings may often be significantly lower than the gross replacement rate used to determine the level of benefit paid. 5.27 In some cases, the past earnings may exceed those which an unemployed person could reasonably expect to get in a new job. A person who is fired on account of a failure to perform in the job may expect to receive a lower wage in the future. A worker made redundant as part of restructuring may not be able to find another job at the same skill level and wage. In this case, an initial gross replacement rate of 70 percent might be viewed as a cause for concern. The practice of calculating benefit solely on the basis of earnings with the last employer (rather than all employment in a previous period) may aggravate this situation. 5.28 Incentives to work are in theory affected by the duration of benefit as well as its level. The total of two years of benefit might result in a disincentive problem and attention should certainly focus on this aspect as much as on the level of benefit. However, the two cannot be seen as independent; the reduction in the effective replacement rate over the benefit period is important to bear in mind. 60 5.29 Finally, it should be noted that there are administrative sanctions against those whom the authorities believe are not making sufficient efforts to find work. Benefit can be withheld from those who "fail to co- operate" with the employment center where they are registered.' If a vacancy is suggested to an unemployed person, then benefit is withheld unless the claimant can demonstrate rejection by the employer. These sanctions may of course be harder to apply in a time of high unemployment when greater demands are placed on the employment center staff. 5.30 In the absence of concrete evidence on the effects of the current benefit levels and entitlement periods on the probability of leaving unemployment, it is not possible to conclude definitively that the scheme presents serious disincentives to work.9 However, considering the potentially high financial cost of the present scheme if many unemployed persons stay on benefit close to the maximum entitlement period, the authorities should give serious consideration to reducing the maximum insurance period to 12 months, and reducing the benefit in Phase II to a fraction (say 80 percent) of the minimum wage. Such a step would require, however: (i) expansion and improvement of the quality of proactive measures, in particular retraining, and (ii) provision of social assistance for persons who remain unemployed after the maximum entitlement period. These two actions are necessary in any case to provide an adequate safety net for workers during economic adjustment, but become even more essential if the total duration of unemployment insurance is reduced below its present maximum. 5.31 There are two other areas of concern with the unemployment benefit system which affect incentives to work. 5.32 Provisions for older workers. The new Employment Law allows older unemployed workers to claim their social insurance pension provided: they satisfy the normal eligibility criteria for a pension, have received unemployment insurance for at least 6 months, and are within 3 years of the minimum retirement age. This so-called "pre-pension" would be calculated under the normal pension rules and paid out of the Solidarity Fund."0 This proposal goes directly against the increase in the effective retirement age recommended in Chapter 4. However, there is a recognized need for special provisions of income support for older workers whose reemployment prospects a This power does not go unused; in January 1990, 445 persons were not receiving unemployment benefit for this reason which may be compared to a total number of recipients at that time of 9,706 (Matoricz, 1990). 9 Much of the evidence from the OECD area on the effects of unemployment benefits on the incentive to work suggests rather small effects but this evidence is often far from robust (Atkinson and Micklewright, 1990). tO This provision is in addition to the existing "early retirement", which remains possible for workers within 5 years of retirement age who have not yet been made redundant. The Employment Fund and the employer negotiate a sharing of the pension costs until the person reaches retirement age. 61 are poor. One less costly approach would be to offer a longer period of benefit for the older unemployed, as occurs in some other countries, as a flat-rate payment. A means-tested benefit to span any remaining gap between the period of unemployment insurance and the minimum retirement age could also be considered. As discussed in Chapter 4, introducing an actuarially-adjusted pension for retirement before the standard retirement age should also be part of the solution. 5.33 Part-time work. At present there is little part-time work in Hungary. However, this may well change in the future, particularly for women; part-time work could be an important way of reconciling the needs of the family and the economy. The new legislation, like the former system, allows an unemployed person taking part-time work to continue to receive the insurance benefit in full on top of the part-time earnings, provided that these earnings do not exceed the minimum wage. There is no tapering of the benefit in terms of the number of hours worked. The rules would seem to present a disincentive to move from part-time to full-time work and an incentive to move from full-time to part-time work via unemployment." 3. Financial Sustainability of the Unemployment Support 5.34 It is impossible to assess the financial viability of the new unemployment benefit system without better information on the nature of unemployment, including the likely share of career beginners and pre- pensioners among total claimants. It is noteworthy that a major component of the cost is the social insurance contribution which the Solidarity Fund must pay to the SIF for recipients of unemployment insurance and for career beginners. Considering this (currently) 43 percent premium, these beneficiaries may cost the Solidarity Fund more than even the pre-pensioners. Based on average benefits expected for 1991, it can be determined that a one percent increase in the eligible unemployed (50,000 persons) will imply an additional requirement of Ft. 5.4 billion for the Solidarity Fund, or roughly a 0.7 percentage point increase in the unemployment insurance contribution rate. Projections of the possible magnitude of expenditure on labor market programs are provided in Chapter 10. 5.35 According to the budget of the Solidarity Fund and Employment Fund in 1991, the latter (which approximately corresponds to the proactive element) will account for 43 percent of total expenditure on labor market programs. This share compares well to the norm for the OECD, where it has been found that countries with a relatively high share of resources devoted to proactive labor market programs (such as Sweden) have a good record of reducing long term unemployment.12 Even though the revenue sources for the two funds are different, with the Solidarity Fund based on earmarked contributions, any t1 The legislation also introduces a new subsidy to employers to encourage shifting workers to part-time status who would otherwise be laid-off. The above criticism of the treatment of part-time work by unemployment insurance is not directed at this new arrangement. 12 Based on a review of 22 OECD countries in 1988. See Scherer (1990). 62 unplanned surge in unemployment benefit claims will put pressure on the general budget. To protect budgetary resources for the proactive programs, the Hungarian authorities may have to reduce the length and generosity of cash benefits as unemployment rises. F. Administration. Costing. and Monitoring of Unemployment SuRport 5.36 The substantial increases in unemployment which must be expected imply that considerable importance should be placed on the implementation, budgeting, and evaluation of the unemployment benefit scheme. Detailed discussion of these issues is provided in Annex V, with the main points summarized below. 5.37 Considerable challenges remain to improve the administration of the unemployment benefit system. Since 1989, the government has made a substantial investment in computerizing both claims for benefit and notification of job vacancies in the regional Employment Offices. However, despite this technology, the Mission was informed from separate sources that 1-3 months would be the normal gap between claim and payment, which represents a considerable delay. Recent hiring by the employment offices has brought the staff-to-client ratio to about 1:87, which is not far from the OECD average of about 1:50-60 (Scherer, 1990); however, the expected level of unemployment in 1991 may bring the staffing ratio back up to an overtaxed level of around 1:150 even with additional hiring planned.'3 To obtain reasonably accurate forecasts of the costs of support services for the unemployed, a more rigorous model of benefit prediction will need to be developed which is based not on an assumptions about the stock of unemployed, as in the present method used, but rather on the flows in and out of unemployment and the various programs of support. 5.38 Experience from the OECD area shows that data collected as part of the administration of unemployment benefit systems are invaluable for monitoring the occurrence and structure of unemployment (such as the frequency of recurrent unemployment) and the performance of the benefit system. The Ministry of Labor should aim to publish at regular intervals (e.g. quarterly) basic statistics on the operation of the unemployment benefit system. Together with these statistics on the registered unemployed, information is needed on those not registered and on family and household incomes of both groups of unemployed. Much of these data should be obtained from sample survey microdata, as intended through the new Labor Force Survey beginning in 1991. "3Staff of the employment offices numbered about 1150 in January 1991, with a stock of registered unemployed of about 100,000. By end-year, the staff is planned to be increased to 1500 while registered unemployment is expected amount to 200-250,000. (Based on verbal communication from Ministry of Labor staff). 63 G. Conclusions and Recommendations 5.39 These may be summarized as follows: The new unemployment benefit system is an improvement over the former scheme in a number of respects; however, the maximum duration of unemployment insurance could contribute to longer periods of unemployment by providing a disincentive to reenter the labor force, as well as being costly to maintain. The total period of unemployment insurance should not exceed 12 months. If a second year of benefit is retained, it should be a flat-rate payment less than the minimum wage, or be means-tested. Unemployment insurance is most effective and efficient when focused on providing temporary income support during relatively brief periods between employment (e.g., up to one year). For longer term structural unemployment, proactive labor market programs must provide the main solution to reintegrating workers into the labor force. A means-tested benefit of last resort should also be made available to the unemployed, and this could be achieved through reform of the existing Social Assistance scheme. The latter could supplement unemployment insurance for those with family circumstances requiring special help, as well as provide primary support for those uninsured or whose benefit has expired. However, the means-testing should be designed in such a way that it does not provide a disincentive to work by other family members. The treatment of part-time work and benefit for older workers need to be reconsidered, as the present schemes may discourage return to regular employment and be unnecessarily costly to maintain. The adequacy of unemployment benefits and their disincentive effects depends in part on the incomes of the unemployed person's family or household, and resources should be devoted to more measurement in this area. The Government is urged to make periodic Labor Force Surveys a regular activity of the CSO, based on experience to be gained through the pilot survey in 1991. More detailed costing and monitoring of the benefit system is needed, using in part data collected as a by-product of benefit administration and job matching. 64 CHAPTER 6 SUPPORT FOR THE FAMILY A. Introduction 6.01 This Section addresses support for families provided through family allowances and the tax system. Benefits in kind and maternity/child care payments are discussed in Chapter 7. The basic rates of family allowance and quantitative details of the analysis are contained in Annex VI. 6.02 The family allowance, a universal benefit paid by the State, is one of the most important cash benefits currently provided in Hungary, whether measured in terms of numbers of recipients, ratio of benefits to incomes, or relative to macroeconomic aggregates. The importance of family allowance has increased sharply since 1987 due to a marked rise in its real value, the number of families in receipt being almost unchanged through 1987-9. Comparisons of universal child support schemes in Central and East European countries and those in Western Europe indicate that Hungary in 1980 had the most generous scheme of all and this generosity has since increased still further. (See Annex Tables VI.1 and 2). 6.03 As discussed earlier in this Report, the different role of wages in planned economies is one explanation for the differences in social policy between Eastern and Western Europe. As wages in the reforming socialist countries begin to perform the role of reflecting productivity differences as in market economies, family allowances can be expected to decline in relative terms. But another explanation is that typically a greater variety of instruments are used to provide support for the family in OECD countries; a focus on just the universal child benefits is misleading. 6.04 This aspect of the comparison with OECD countries is a reminder that future provision of support for the family in Hungary may require a number of instruments to be used. In choosing these instruments the policy goals need to be spelt out. Universal benefits increasing with family size may be thought to serve the traditional pro-natalist policy of centrally planned economies, although the evidence suggests that it was not a success (CSO, 1984). Economic stability may be more important in encouraging population growth. Other goals should be considered more relevant at the present time. These include the provision of some off-set to the costs of raising children (so as to promote equalization of consumption across family size), the partial stabilization of family incomes, and the alleviation of poverty among families with children. 6.05 Currently, family allowances are supported by one other policy, the provision of a tax allowance against the PIT for families with three or more children. Part B looks at the incidence of these two policies, and at the amount by which family allowance reduces poverty in families with children. Part C considers three possible reforms that could improve targeting of family support. Part D discusses the future of family support within the tax system and Part E concludes. 65 B. The Incidence of Family Support and Child Poverty 1. Family Allowance 6.06 The incidence of family allowance is shown for 1989 in Table 6.1. Family allowance is clearly very important at the bottom of the income distribution, its share of total income falling continuously as one moves to higher deciles. Among the poorest ten per cent of the population, the benefit accounts for 20 percent of total income; in interpreting this figure it must be born in mind that one-third of households whose members make up this decile receive no family allowance (column 3) implying that its importance to those households with children is even higher. But the table also shows that a significant amount of the total expenditure on family allowance goes to much higher income households. More than one-sixth is received by the top three deciles; the Ft 9 billion (in 1989 prices) involved represents a very substantial amount of money but made up only about 2 percent of the total incomes of the households concerned, although the rather low proportions of households with any family allowance in this part of the income distribution should be noted. 6.07 On one criterion, this table suggests that Family Allowance is not well targeted, in that a considerable proportion of the expenditure goes to those with higher incomes. This is the situation with any universal benefit and the general picture given by the figures in the table should come as no surprise. The apparent "wastage" of expenditure on high income households must be balanced against the attractions of a universal benefit; these are considered further below. The aims of family support need also to be taken into account. The wastage argument relates most to the goal of poverty alleviation and is less relevant when looking at horizontal equity between those with and without children. 2. Child Tax Allowances 6.08 The personal income tax (PIT) in Hungary embodies the principle of independent taxation of the incomes of married couples; indeed, marriage has no implications for an individual's tax treatment, there being no additional personal allowances for married persons, for example. However, families with three or more children may claim an additional Ft 12,000 per year allowance and this goes to the parent with the higher income (Act on the Personal Income Tax, Section 17). This allowance represents the only recognition of the family within the PIT.1 1 Clarification is needed as to whether the allowance is Ft. 12,000 per child, or per family (as the PIT text implies). 66 Table 6.1: Family Allowance (FAM) Expenditure by Income Decile Shares. 1989 (Net Per Capita Household Income) Decile Share (%) Share (%) of FAM % of Income of Total FAM in Total Net Households Decile Expenditure Household Income Receiving FAM 1 (lowest) 16.3 20.8 64.6 2 14.3 13.7 60.7 3 13.1 11.1 57.5 4 11.1 8.5 50.7 5 9.9 7.0 44.4 6 9.3 5.9 42.6 7 8.4 4.8 38.9 8 7.3 3.7 34.5 9 5.6 2.4 27.5 10 4.7 1.4 20.9 Total 100.0 Avg. 6.0 Avg. 42.6 Notes: The first column in the table shows the share of total expenditure on family allowance within each decile of per capita net household income; this shows how the benefit expenditure is spread across the income distribution. The second column gives the share of family allowance in the total income of each decile; this shows the importance of family allowance at different points of the income distribution. The income distribution is one of final income i.e. including Family Allowance; the table. does not show the share of total expenditure going to each decile of pre-allowance income. Source: Incidence Study 1990, Tables 3,4,7 6.09 As with any tax allowance, it is useful to think of the third child tax allowance as a tax expenditure. Its cash value is equal to the marginal rate of tax paid by the parent with the higher income, multiplied by the value of the allowance. This means that it is worth more the higher the marginal tax rate, and it is worth nothing to the non-tax payer.2 The cash value of the tax allowance is very small in relation to family allowance, even for families in which one parent is paying a high marginal rate of tax. If the average tax expenditure were worth Ft 300 per month then, given the number of three-child families in Hungary, the total tax-expenditure involved would be below Ft 2 billion per year, adding less than 3 percent to the total value of support given to all families by family allowance. 2 The cash value of the allowance would be worth Ft. 150 a month when tax first becomes payable at 15% of income, rising to a maximum of Ft 500 a month to the person paying tax at the top marginal rate of 50%. This may be compared with a monthly family allowance of Ft 6,600 in the first half of 1990 for a family with three children. 67 6.10 This suggests that the targeting of the third child tax allowance is very poor, since it is worth most to households with higher income parents. In practice, however, this depends on the distribution of income of parents with three or more children. If such parents are all tax-payers and are heavily concentrated within one tax bracket with a low marginal tax rate, then the criticism just made has less relevance. The necessary data are not at our disposal but the tabulations of the 1989 Incidence Study provide some indication. Table 6.2 shows the distribution of households with children across per capita net income deciles in 1989. Large families appear concentrated at the bottom of the income distribution: some 60 percent of households with three or more children are in the bottom two income deciles. 6.11 Were there tax-payers in all these low-income households who used the allowance? The per capita net income of a three-child household with two adults, both earning at the level where tax would first start to be paid, would have been about Ft 3,100 per month in 1989 (assuming no other income); this compares with a figure of Ft 4,190 for the first decile in Table 6.2. This suggests that the third-child tax allowance probably was used by most low income families (especially when differences between earnings of the two adults are taken into account). However, similar calculations indicate that even within the bottom two deciles the value of allowance was not constant and there must have been a significant number of households in higher ranges obtaining greater assistance. 3. Poverty among Children 6.12 How much poverty is there among households with children? How effective is family allowance at reducing poverty? Does the picture vary with households of different sizes? 68 Table 6.2: Proportion of All Households with Three or more Children in each Per Capita Income Decile. 1989 Proportion of All Households with Per Capita Income Decile 1 Child 2 Children 3+ Children 1 5.1 11.8 36.1 2 8.1 15.0 23.5 3 9.3 15.8 12.9 4 11.3 13.6 8.4 5 12.5 10.6 6.1 6 12.4 9.4 4.9 7 11.5 7.4 2.0 8 10.3 6.7 1.7 9 10.2 5.2 3.0 10 9.4 4.6 1.4 100.0 100.0 100.0 Source: Incidence Study 1989, Tables p.36. 6.13 Table 6.3 provides estimates of the incidence of poverty for households with different numbers of children.3 The table also shows the proportion of households which would be classified as poor on the same basis if no family allowance were paid. It should be noted that all the figures in the table are estimates based on tabulations (of income distribution) in the 1989 Incidence Study; they are not calculations which have been made with the benefit of access to the relevant survey microdata. The figures should be taken as giving an indication of the magnitudes involved rather than being seen as precise measures. 3 It should be noted that the figures relate to households and not families. Chapter 8 on Social Assistance provides some discussion of this important distinction. A household has been defined as being in poverty if per capita net income is below the relevant minimum subsistence line calculated by the CSO. 69 Table 6.3: Poverty and Family Allowance. 1989 Active Households with 1 Child 2 Children 3+ Children Percent of Households with Income beneath Subsistence Minimum 5.3 6.4 14.5 Percent of Households with 10.0 18.4 41.9 Income beneath Subsistence Minimum if Family Allowance not paid Poverty Reduction Rate 47.0 65.2 65.3 of Family Allowance (percent of households beneath subsistence minima who are raised above the threshold by family allowance) Source: Incidence Study 1989, Tables 36 and 37. Note: Figures interpolated assuming Pareto distribution within each decile with no distinction between urban and rural households; in the case of 3+ children, the subsistence minimum for a household with three children has been used. The proportions of households beneath subsistence minima in the absence of family allowance were calculated by subtracting the per capita values of the relevant allowance (taking account of average household size by decile indicated in the database) from the per capita decile figures and re-interpolating within the range in which the minima then fell (average values of January and August family allowances were used). These latter calculations assume that all the children in one household belong to the same family. The calculations do not distinguish between single parent and other families (a weighted average of the allowance for the two groups was used for single and two child households). 6.14 The first line of figures in the table shows there is a substantially higher poverty rate in households with three or more children, as discussed in Chapter 2. In part this result may be due to applying the subsistence minimum income level for three children to a group including households with larger numbers of children. On the other hand, the result is qualitatively similar to that found in other studies of poverty in Hungary (e.g. Szalai, 1989). 6.15 What effect does family allowance have on poverty? The second line in the table gives an estimate of the proportion of households which would be beneath the relevant subsistence minimum in the absence of family allowance. The third line shows the proportion of households which are poor on a pre-family allowance basis but are pushed above the relevant threshold of income by the 70 allowance. Family allowance clearly has a big effect on poverty, as one would expect in view of its importance in final household incomes in the lower part of the income distribution; two-thirds of households poor on a pre-allowance basis with two or more children are removed from poverty and nearly a half of households with one child.4 6.16 The figures in Table 6.3 can be viewed in different ways. On the one hand, it could be argued that despite an expensive income support system for the family in Hungary, a significant problem of child poverty still exists, this being especially true for large families. It might also be pointed out that pre- allowance poverty among one-child households is fairly low. These arguments would support the view that family allowance is poorly targeted. On the other hand, the table shows that child poverty would be very much higher in the absence of family allowances. Both views need to be taken into account in any discussion of policy reform. C. Three Possible Reforms of Family Allowance 6.17 The reform of the family support system in Hungary is clearly a sensitive issue. Family allowances were only made universal in April 1990; this was an important principle to have introduced, even if the number of additional recipients was small (about 5 percent). Family allowance has been up-rated every six months in recent years and substantial increases in real value which have resulted from this. There can be little doubt that this has generated expectations of continued generous support. Considerations such as these need to be born in mind together with the evidence on the incidence and impact of family allowance described above. 6.18 Three methods of changing family allowance so as to reform support for the family are discussed below: these are more sharply differentiated payments by family size, means-testing, and the taxation of family allowances. These possible reforms are all aimed at improving the targeting of support on those families with lower incomes. 6.19 Improved targeting may be the overriding aim at the present time but the other goals of family support listed earlier should not be forgotten. In light of these goals, it is helpful to review the attractions of universal family allowances. Firstly, a universal benefit is a highly efficient method of providing income - administrative cost is low. Secondly, there is little or no problem of take-up; all those who are eligible will receive. Thirdly, the 4 The calculations of pre-family allowance poverty have been made by withdrawing a calculated value of the relevant amount of family allowance from all households with children. In the case of one-child households there will be some instances where no allowance was in fact being paid, implying that pre- allowance poverty is overstated in the table. In this category will be households with families with one child aged over six and no elder children for whom allowance was previously paid. Nevertheless, the number of one-child households in the Incidence Study database seems broadly in line with the total number of families recorded in administrative data as receiving family allowance in respect of one child. 71 lack of an income-test implies that no poverty-trap exists; there is no reduction of benefit payment as other family income increases. The high implicit marginal rates of tax typically present in means-tested schemes, and the resulting disincentive to work, are therefore avoided. Fourthly, universal family allowance provides an important element of protection against income instability, which may occur if a parent is unemployed and there are delays in receipt of unemployment benefit (as noted in Chapter 5), or in the event of divorce (which is relatively common in Hungary). The challenge for reform of family allowance in Hungary is to improve targeting onto lower income households but at the same time preserving as much as possible the advantages of a universal benefit.5 1. Differentiating Payments by Family Size 6.20 Table 6.2 showed that the greater the number of children, the further down, on average, the position of the household in the per capita income distribution. Associated with this is an increasing probability of being in poverty, especially where there are three or more children. One possible reform of family allowance would be to redistribute income from smaller to larger families by changing the payment per child, reducing it for smaller families, increasing it for larger ones. The change could be effected so as to bring about a reduction in total family allowance spending. 6.21 The family allowance system already embodies some differentiation of payment per child by the number of children in the family (see Annex Table VI.3). However, this differentiation is only moderate and does not particularly favor large families; it is also much less than that in some Western European countries (although it should be noted that differentiation by family size is not prevalent throughout Western Europe). The treatment of single parent families in Hungary is also not especially supportive. In fact, single parents with three children receive the same total child support as families with both parents present. This does not seem a sensible arrangement and any reform of family support in Hungary should aim to improve the position of single parents. 6.22 The most obvious change that could be made to the rate structure of family allowance would be to reduce or eliminate the payment for the single child (a simple failure to index would, at current inflation rates, produce a substantial real cut). In 1990, the age limit of 6 years for one-child families was removed, resulting in an increase in the number of families receiving the one-child allowance from 171,000 to 305,000 (Annex Table VI.3). Consequently, at the level of allowances fixed in August 1990, payments for one child (excluding single parent families) amounted to Ft. 13.3 billion (21 percent of total family allowances in 1990). 5 It is interesting that a recent (1990) study by the Institute for Fiscal Studies in London, "Alternative Tax and Benefit Policies for Families with Children" considers an option for the U.K. of increasing the existing universal family allowance but subjecting it to progressive income taxation, in an effort to improve the targeting while retaining the advantages of universal provision. 72 6.23 Eliminating the one-child allowance would produce a significant saving in family allowance expenditure, but at a substantial cost for families with one child. Despite their higher position in the income distribution relative to households with larger numbers of children, it is still the case that poverty among one-child households (as measured by the minimum subsistence income line) would rise significantly as a result; Table 6.3 suggests a doubling of the incidence from 5 to 10 per cent. Unless accompanied by other changes, this measure would also do nothing to reduce the poverty in large families, which should be an aim of any reform. In this sense, the reform would save money but would not improve targeting; overall child poverty would be lower than with an across the board reduction in family allowance but it would still be unambiguously higher than if no changes were made. 6.24 It is to be hoped that some of the savings from a cut in first child family allowance would be used to relieve child poverty via increased expenditure in other areas, e.g. through an explicit means-tested benefit for families with children or by real increases in the family allowance for large families. The above estimate of Ft. 13.3 billion in potential savings at the August 1990 allowance rate would therefore be less by the amount of any transfer to other measures addressing family poverty. 2. Means-Testing 6.25 Means-testing of family allowance could obviously generate large expenditure savings and at the same time allow a higher level of benefit to be paid to families with income beneath a designated income level. The savings generated would depend both on the cut-off level of income chosen and on the rate at which benefit was withdrawn as income approached this level. 6.26 The drawbacks of means-testing family allowance are in many ways the converse of the attractions of universal provision. Low take-up may especially characterize a means-tested benefit designed to supplement a universal allowance, a possibility for Hungary which we consider below.' Where means- testing is used, it is essential that the authorities aggressively promote the program to ensure public awareness and reduce the stigma (which may be worse, the more restrictive the means-testing). 6.27 The immediate abolition of universal family allowance and its replacement by a means-tested benefit would be a very radical reform which could be expected to be deeply unpopular. More practical on a variety of grounds (both political and administrative) would be a combination of a lower universal allowance with a supplementary means-tested program.7 Quite apart from 6 For example, in Britain in 1985-6, a supplementary means-tested family benefit had an estimated take-up rate of only 50%, despite the scheme having been in existence for some 15 years. (HM Treasury (1990) Table 14.16). 7 In December 1990, a portion of the Ft. 600 cost of living increase in family allowance/recipient which was to have been introduced from January 1, 1991 was reallocated to targeted social assistance. The aggregate amount to be rechanneled in this way during 1991 amounts to Ft. 3.8 billion. 73 practicalities, a combination of programs may in any case be the most appropriate policy aim when designing family support policy. The basic universal allowance could continue to pay a fraction of the costs of a child, with the means-tested supplement paying a substantial part of the remaining costs to those on low incomes. 6.28 Hungary has little experience of widespread means-testing on which to draw and there is the danger that this might be reflected in any scheme which is hastily introduced. There is a strong argument for the introduction of such a supplementary benefit in the future, but the first priority in designing a workable and equitable means-tested scheme in Hungary would seem to lie with Social Assistance. Many of the same problems arise here and valuable experience could be gained which could be applied in the field of family benefits. 3. Taxation of Family Allowance 6.29 Taxation of the family allowance would achieve targeting by an implicit income-test. The use of the PIT to reduce the value of family allowance to those with higher incomes again represents the application of a combination of policy instruments. Moreover, the tax system is an instrument which is already in place. Although there are administrative considerations stemming from taxation of family allowance, there is not the need to set up a completely new income-testing mechanism.8 The PIT represents a potentially powerful tool of redistribution and its use in helping target family allowance deserves careful consideration. 6.30 The taxation of family allowance would in theory have considerable attractions. Benefit would remain universal, continuing to represent society's sharing of the costs of raising children. The take-up problems of means-testing would be avoided. There would be no poverty trap for those on low incomes. The tax system would provide an immediate adjustment in benefit level with changing income, and the role of family allowance as insurance against income instability would remain. 6.31 The problems of subjecting family allowance to tax in part stem from the treatment of the family within the PIT. As mentioned, the tax system embodies independent treatment of husband and wife and tax liability depends solely on individual income. Family allowance, on the other hand, is, as the name indicates, a benefit paid to families. In practice it can be paid to either husband or wife and although this may be of considerable importance to the financial arrangements made within individual marriages, it has no implications at the present time for other parts of the tax and social benefit system. 8 Family allowance is disbursed either by the employer or by the social security office (for unemployed persons). If the tax were to be withheld by the latter, there would be a problem of establishing the relevant marginal tax-rate to apply to family allowance, this depending on other taxed sources of income not paid at that office. 74 6.32 This situation would change with the taxation of family allowance. With the present PIT structure it would become necessary to indicate who would be liable to pay the tax due on the family allowance. The revenue which would accrue as a result would depend on this decision, due to the lower participation rate and lower wages of women. Moreover, as a result of the reduction in number of tax brackets in 1990, redistribution and hence targeting via the taxation of family allowance would probably be much blunter than might appear from a first consideration of this alternative. A very large number of family allowance recipients could be expected to pay tax on the allowance at the same rate. It is also likely that the family allowance would push many recipients into a higher tax bracket. A reasonable estimate of the range of revenue which could be gained from taxation of the family allowance, at 1990 tax and family allowance rates, is between Ft. 15 billion (if wives paid) and 19 billion (if husbands paid). (See Annex VI for calculations.) 6.33 An obvious choice regarding the decision onto which partner the tax liability should fall might be to tax the partner actually receiving family allowance, but to allow free choice to a married couple as to who this should be, as at present. Typically, we would expect families where wives pay lower marginal rates than their husbands (the normal case) to choose that the wife receive the allowance, thus minimizing the total tax paid by the family. However, the distributional consequences of this, in terms of family incomes, depends on the joint distribution of husbands' and wives' employment income. If the wives of high earning husbands have a lower labor force participation rate, then within the present PIT there would be high income families who pay less tax on family allowance than lower income families with two earners. This underlines the problem of applying an instrument designed to modify individuals' incomes to the taxation of a family benefit. If, on the other hand, the principle was adopted of paying family allowance only to the mother (except in the case of a divorce resulting in the father having custody of the children) as a "mother's benefit", then the inconsistency would be less apparent. Taxation of the family allowance would also produce more effective targeting if the tax brackets were set so as to spread the actual earnings distribution widely over the scale of marginal tax rates. D. Family Support and the Tax System 6.34 The possible reforms of family allowance discussed above do not provide for any radical change in the tax treatment of the family; although the taxation of family allowance would represent a significant change in the treatment of that benefit, it would not represent a fundamental reform of the tax system. In this section we consider the future treatment of the family in the tax system in more detail, including the integration of the family allowance and tax systems. 6.35 Could more support for the family be provided through the tax system as well as through the cash benefit system? What forms could the future tax treatment of the family take? A possible move to a family income tax system based on income splitting is already under discussion in Hungary. There are a number of ways of operating such a system. In the scheme which operates in France, for example, the incomes of the family are summed and the same allowances, tax bands and marginal tax rates which are applicable to single 75 persons are then applied to the per capita incomes of each family member (with the first two children counting as one-half in the calculation). This sort of system is at the opposite end of the spectrum of tax treatments of the family from the present Hungarian PIT. It would represent a reform of enormous significance, greatly favoring married couples, especially those with children, and would require considerable adjustment of marginal rates. Any change to such an alternative should only be contemplated after extensive analysis of the distribution of gains and losses which would occur. 6.36 There are, moreover, strong arguments for maintaining the present system of independent taxation of husband and wife and it should be noted that a move away from the present position would be against the general trend in industrialized market economies. Firstly, the attraction due to the administrative simplicity of the system is obvious; when determining individual liability there is no need to take account of the incomes of other family members, a particular problem when those incomes are subject to fluctuations. Secondly, independent taxation allows for privacy of individual finances within marriage, limiting the situation where full details of personal income must be shared to those families receiving means-tested benefits. Unlike a family system, independent tax treatment embodies no assumption about the degree of income sharing which actually occurs within the family. 6.37 Increased support for children, as opposed to married couples, is not incompatible with a continued policy of independent taxation of husband and wife. One way of providing this would be through child tax allowances as described earlier. Greater use of child tax allowances is not a policy which we can support, however, since it would be inherently regressive by providing greater assistance to higher income families.9 It would run counter to any aim of improving the targeting of family support. Indeed, we have recommended that the allowance currently granted to families with three or more children be discontinued, with the additional revenue being provided to large families in another way. 6.38 What other alternatives should be considered to provide additional support to families within the tax system? One option would be to convert the tax allowance into a tax credit; the amount of the credit (which could be equal to the current tax allowance multiplied by the bottom marginal tax rate of 15 percent) being subtracted from tax due, rather than from taxable income. Where no tax was due the amount of the credit could be refunded as a cash payment. In this way the value of the old allowance would now be the same for all. 6.39 A tax credit, refundable in the case of low income, represents an integration of tax and cash benefit systems. This may have both advantages and disadvantages from an administrative point of view. The principle could of course be extended, with the replacement of the family allowance itself by a tax credit; there would certainly be little point in the co-existence of family allowance with a restricted system of child tax credits for large families. This 9 Although it should be noted that allowances can be designed that are progressively withdrawn as income rises. This is the case with child tax allowances in the US. 76 reform would have two implications. Firstly, a child tax credit system would require independent taxation of husband and wife to continue if the principle of providing family support to the mother is to be further strengthened. Secondly, since tax-credits have the same value independent of income, they would not necessarily improve targeting unless the credit is partially withdrawn at higher incomes. 6.40 There is one final alternative to be considered: the replacement of the personal tax allowances with personal refundable tax credits. This would have far-reaching consequences for all tax payers, not least the changes in marginal tax rates which would be needed to finance the credit refunds. The implications for family support would include the additional income for parents who do not use all their personal tax allowance since they are not at work or because they have low earnings. E. Conclusions 6.41 This Chapter began by pointing to the importance of family allowances in Hungary. This importance implies that great weight should be attached to analyzing the effects of any possible reform, and in particular the distribution of resulting gains and losses. 6.42 It is suggested that reform may require a number of policy instruments being used to provide support for the family, this being the case in many Western European countries where universal family allowances are less important in comparison with Hungary. The reform package could include several elements discussed above, including more sharply differentiating payments by family size, means-testing a portion of the allowance, taxing the family allowance, and reforming the tax treatment of children. These should not necessarily be seen as alternatives. A real cut in family allowance for one- child families and increases for second and third children could be accompanied by a means-tested scheme providing supplementary support for low income families. Income testing could also be provided via the tax system as an alternative requiring no new administrative machinery. The financial implications of these various options have been roughly estimated as follows, although the net overall savings would depend on the combination of measures used (at 1990 rates and percent of total family allowance expenditure): -elimination of 3rd child tax allowance ........ Ft. 2 bn revenue increase (3%) -elimination of family allowance for one-child two-parent families ............ Ft. 13 bn expenditure saving (21%) -taxation of family allowance .......... Ft. 15-19 bn. revenue increase (24-30%) 77 6.43 The Report does not recommend a specific package but urges that the following elements be included: (i) retention of universal family benefit in some form, justified as providing some compensation for the private costs of raising children. A cash allowance should be included in taxable income to improve targeting ex post on low income families; an equally acceptable approach would be to shift to a refundable tax credit which can be scaled back for higher incomes; (ii) abolition of the existing tax allowance for large families. If additional support to large (also generally the poorest) families is necessary, a larger cash allowance or refundable tax credit for them would be preferable, financed by cutting or eliminating the present allowance for one-child two-parent families (this would amount to greater differentiation of support by family size than presently exists). The family allowance system should in any case be buttressed by a means-tested benefit for poverty relief which is provided through social assistance (paras. 8.17-8.19); (iii) payment of the allowance in every case to the mother, except where she does not have custody, to better ensure that the benefit is spent for the child regardless of income-sharing between the parents. To enable a child tax credit to be received by the mother, independent taxation of husband and wife should be retained. 78 CHAPTER 7 MATERNITY AND CHILD CARE SUPPORT A. Introduction 7.01 Hungary has a comprehensive system of maternity and child care payments which provide income support to families with children aged under 3. These permit lengthy periods of paid parental leave from employment. Three of these benefits we consider here' (data pertaining to this discussion are attached as Annex VII): (i) Maternity Allowance, providing 24 weeks of taxable benefit, at 100 percent of gross earnings in the previous calendar year, for women with at least nine months employment in the two years prior to the birth, (ii) Child Care Fee, providing a continuation of the maternity allowance until the child's second birthday, at 75 percent of previous earnings, again taxable (the fee can be paid to either the mother or father after the first birthday), (iii) Child Care Allowance, providing a tax-free flat-rate payment until the child's third birthday for those exhausting entitlement to the child care fee, or as an alternative to the fee for those not qualified by work history, e.g. students (payable except in special cases to the mother). Together with these cash benefits, which are paid from the Social Insurance Fund (the child care allowance is administered by the Fund but paid for by the State budget), this Chapter also considers nurseries providing day-care of children up to the age of three, used by those families where both parents are at work. The remainder of this Introduction reviews the scope and incidence of child care benefits. Part B discusses the rationale for these benefits within the context of the economic and social reform in Hungary. Part C considers some possible reforms to the program and Part D looks at the specific issues of nursery care. Part E summarizes the recommendations. 7.02 The system of cash benefits and nursery care has developed over a number of years. The flat-rate child care allowance was introduced in 1967 but the earnings-related child care fee dates only from 1985. As Annex Table VII.2 shows, it is now the fee which is the more important of the two, accounting in 1989 for nearly two-thirds of the quarter of a million recipients. Women receiving one or other of the two benefits represented over 8 percent of all women of working age in that year.2 Expenditure on the child care fee and allowance totalled Ft. 13.9 billion in 1990, which, added to I There is also a lump-sum maternity grant and a system providing paid leave from employment of parents of sick children. 2 CSO, Statistics Yearbook, 1988, Table 4.1. 79 maternity and pregnancy benefits (Ft. 4.8 billion) amounted to almost one percent of GDP. Maternity and child care benefits therefore represent a significant and costly item in the State budget. In view of the present budgetary difficulties, this alone provides sufficient reason for an appraisal of the present system. The change to a market economy provides a second reason; the present arrangements are unlikely to be compatible with the development of a large private sector. 7.03 Table 7.1 shows the spread of recipients of the child care fee and allowance across the distribution of final per capita household income (i.e. post-fee/allowance). Some 40 percent of expenditure goes to persons in the bottom 20 percent of the distribution and over 25 percent of households receiving family allowance in this part of the distribution also received one of these two child care payments. Even if we allow for the much higher proportion of households with children in the lower parts of the income distribution, it is clear that a disproportionate share of the total expenditure does go to lower income households. This trend, as already noted in Chapter 3, may reflect a number of factors. Higher-income families may use maternity leave less on account of the interruption it makes to a white- collar career. On the other hand, lower-income households have fewer adult earners, and in a sense have lower incomes because of the child care fee and allowance, since earnings replacement is only partial. Table 7.1: Child Care Benefits and the Distribution of Income. 1989 (Net Per Capita Household Income) Decile Share (%) % of Households Households Receiving Income of Total Expenditure Receiving Child Care Allowance Decile on Childcare Allowance Family Allowance or Fee as % of Number and Fee with Family Allowance 1 (lowest) 19.1 64.6 27.0 2 20.8 60.7 24.8 3 15.8 57.5 18.9 4 8.9 50.7 12.3 5 9.0 44.4 13.4 6 7.1 42.6 9.8 7 5.9 38.9 9.2 8 5.2 34.5 7.6 9 3.7 27.5 6.9 10 4.6 20.9 10.3 Total 100.0 Avg. 42.6 Avg. 15.1 Source: Incidence Study 1990, Tables 4 and 7. 80 7.04 The heavy use made of maternity and child care leave is reflected in a low use of nurseries for children aged less than 3. Only 12 percent of children of nursery age were enrolled in 1989 and 40 percent of available nursery places on an average day were unused, suggesting substantial overcapacity. This may be compared with a figure of 86 percent for participation in kindergarten education for 3-6 year olds (see Annex Tables VII.3 and 4). 7.05 A comparison of maternity and child care benefits in OECD countries (given in Annex VII) confirms that Hungary has a distinctly more liberal regime. These countries include several with a very firm commitment to family policy, e.g. Sweden, Norway and France; Sweden and Norway also have high female labor force participation rates. Even allowing for Hungary's relatively high female labor force participation as compared to most other Western countries, the extent of departure from international practices suggests that a reappraisal is needed as Hungary moves in other respects closer to market economies. B. The Rationale for Child Care Payments in Hungary 7.06 As with other aspects of family policy, a case may be made for maternity leave and child care payments on a number of grounds. 7.07 Firstly, these benefits may be intended to serve a pro-natalist policy. As noted in Chapter 6, however, it would be hard to defend the current system of payments in Hungary on grounds of its effectiveness in promoting births. 7.08 Secondly, a maternity leave and child care policy provides important income support for families with infant children and ensures that children can be looked after in the home in the first years of life. On the other hand, income support can be given as effectively through family allowances. These would allow a greater freedom of choice since they are not dependent on the parents' labor force status. It is notable that at present the family allowance in Hungary for children aged under 3 is only marginally higher than at older ages. 7.09 Thirdly, by permitting interruption of work with guaranteed re- employment after a period of paid support, maternity leave and child care payments are an important method for raising the labor force participation rate of married women, and encouraging women to enter careers requiring a substantial period of training. The high labor force participation rate amongst married women in Hungary should be seen as an economic asset not held by many developed countries. 7.10 The question that must be asked with this third argument is what length of leave (and level of earnings-replacement) is required to provide the required incentives. Reportedly, more highly educated women in Hungary tend to return to work after child-birth much sooner than other women, since a period of three years away from the job would be detrimental to their careers even if job rights were formally guaranteed (i.e., promotion possibilities and "human capital" are lost). It is for this reason that feminist movements and 81 trade unions in Western Europe have criticized very extended maternity leave periods; it is arguable that such periods may encourage women to stay away from work for too long.' 7.11 It is also important to consider the incentives on the demand side of the labor market. Employers may be reluctant to hire women if there are lengthy periods of maternity and child care leave. Will private enterprises in Hungary be prepared to guarantee job rights for three years for each birth?4 This may be particularly important if women are not to suffer disproportionately during the expected period of adjustment of labor demand. 7.12 This leads to the fourth argument for leave and benefit payments. Running counter to the argument just made, maternity and child care leave is often viewed as a way of disguising female unemployment during a period of adjustment.5 However, withdrawing the younger and relatively well-educated women from the labor force through this benefit system is less efficient for the economy than providing unemployment support which compensates those women who are "selected out" of the workforce by the process of restructuring. It must also be recognized that during stressful economic times such as the present in Hungary, the government cannot expect to pay high enough compensation to make it financially attractive for many women to forego their labor income for full-time child care. C. Suggestions for Reform 7.13 This section suggests four possible reforms of the present system of maternity leave and child care payments, aimed at making the system more suitable for a market economy. These are a reduction in the maternity benefit, a switch from earnings-related to flat-rate support, a reduction in the total length of leave, and an introduction of partial employer liability for maternity payments. 3 Gordon (1988, p.299). 4 The present system requires the employer to guarantee, for up to three years per instance of maternity/child care leave, the right to return to the same (or equal-paying) job held before leave was taken. 5 It has been argued that the original introduction of the child care allowance in 1967 was intended to reduce female unemployment associated with the New Economic Mechanism (Adam, 1984). 82 1. Reduction in length or rate of maternity benefit 7.14 Both the length of maternity leave (24 weeks) and the gross earnings replacement rate (100 percent) are well above the standards found in the OECD countries, any of which can better afford such liberal benefits than Hungary. The combination of these two features is especially difficult to justify. Given that the earnings base used to calculate the benefit is not indexed, the actual replacement is not total, but a more transparent approach would be to index the base and cut the rate. Assuming that the average monthly wage of new mothers was 70 percent of the average for all workers in 1989, each month of maternity leave would have cost about Ft. 0.6 billion, assuming also that 70 percent of births were to working mothers.6 2. Switch from earnings-related to flat-rate support 7.15 Both the maternity leave allowance and the child care fee provide benefit related to previous earnings. This may be justified on the grounds of income replacement; earnings-related benefit eases the transition to full- time childcare. However, this argument applies with less force after the initial period of maternity leave, especially if the mother does not eventually return to work, something which can be expected to occur more frequently in Hungary once living standards start to rise. 7.16 The arguments for flat-rate support after maternity leave are strong. Firstly, the child welfare rationale for maternity and child care benefit calls for income support that is the same for all children, in the same way as family allowance is a flat-rate benefit. Secondly, flat-rate support decreases the incentive for the presumably more productive, higher earning women to stay out of the labor force. Thirdly, flat-rate benefit would be preferable on grounds of cost. 7.17 In 1989, the average monthly child care fee was Ft. 4,340; this compares with Ft. 2,660 for the flat-rate child care allowance.' Given the number of persons receiving the fee at that time, the replacement of the fee with a flat-rate benefit at the same level as the allowance would have saved around Ft. 3.2 billion in 1989 prices, or nearly 30 percent of total expenditure on fee and allowance. This calculation assumes no change in behavior but it is to be expected that the switch to flat-rate support would encourage some women to return to work earlier than would otherwise be the case; this would imply even greater expenditure savings. 7.18 The option of switching from earnings-related to flat-rate payment in effect represents abolition of the child care fee and a return to the pre- 1985 position. B Calculation based on Statistical Pocket Book, 1989, pp. 18 and 31. If the benefit was less than 100 percent of the previous earnings (80 percent would be a reasonable replacement rate, based on practices in OECD), this estimated expenditure would be correspondingly lower. 7 CSO, Statistical Pocket Book, 1989, p.54. 83 3. A reduction in the length of leave 7.19 At present, child care benefits can be paid until a child's third birthday. One option would be to limit the total period of paid leave (including maternity leave) to two years per child. This would not affect those mothers, often from higher occupations, who return to work by the time their child is aged two. Nor would it reduce the period of earnings-related support (if this remains) since this does not cover the child's third year. Assuming that the number of mothers who would be affected is 70 percent of the total number receiving the child care allowance at any one time, the saving would be nearly Ft. 2 billion in 1989 prices, or nearly 20 percent of total expenditure on the child care fee and allowance in that year. 7.20 The ending of support in the third year could be expected to produce a significant change in behavior, encouraging an earlier return to work. For this reason, the reduction of the period of paid leave should be seen as a policy change rather than simply as an expenditure cut. This change would be re-enforced if any future provision of unpaid leave was also restricted to two years. We have noted that private sector employers, who it is hoped will continue to grow rapidly in number, may discriminate against women in their hiring and lay-off policies if they are required to hold jobs open for three or more years. For this reason, a reduction in the length of the leave period should not be seen as unambiguously bad for women. 4. A partial switch to employer support 7.21 At present, all maternity leave and child care benefits are provided by the State. Funding is largely through the SIF but, since the contribution rate is the same for all workers, there is no direct cost to the employer of maternity leave. Of course, the employer is obliged to keep the job open and this entails important indirect costs as stressed above. Consideration is being given to shifting the maternity and child care benefits from the SIF to the State budget in 1992. 7.22 The cost of a short initial period of maternity leave, say one month, could be born by the employer; this would be consistent with the recommendation for a longer initial period of employer-paid sickness benefit. This benefit should be earnings-related but well below the present 100 percent, as argued above. 7.23 Eligibility for an employer-provided maternity benefit might depend on the length of continuous service with the employer in question. This contrasts with eligibility under the present social insurance scheme which depends on the total length of work within a given period. For this reason, a change to partial employer provision could have implications for eligibility. 7.24 A change to an initial month of employer leave would produce a small but significant saving in government expenditure on maternity benefit at the expense of an additional cost for employers. The magnitude of the saving (employer expenditure) for one month of maternity leave would be the same as that estimated in paragraph 7.14, Ft. 0.6 billion at 1989 prices (excluding 84 the social insurance contribution). 7.25 The four possible reforms suggested here would result in a maternity and child care support system which is more akin to those in the OECD area. The reforms should not be seen as competing alternatives. For example, a reduction in the length of leave would be to the benefit of employers and the partial switch to employer support could be seen as the price which they would pay for this. 7.26 The significant savings in expenditure which could result from the reforms need not be withdrawn entirely from family support. The savings could in part or in whole be used for improvements in the quality of nursery facilities, tax concessions for nursery care, and other income support programmes providing money or in-kind benefits for families. D. Reform of Nursery Care 7.27 Most nursery care is provided by local councils and, as noted earlier, is currently underused; the main reason may be that it is widely perceived to be of low quality. Demand for nursery places can be expected to expand if child care leave is curtailed but there is also the issue of the structure of nursery provision. What should be the relative roles of state, employer and private provision? As a principle, permitting families to exercise a choice among child care arrangements is certainly desirable. 7.28 In 1988, only 10 percent of nursery places were provided by employers, compared to 17 percent in 1980.8 In the short run, the development of private sector employment can be expected to reduce employer provision still further, not least since new private sector firms will remain small for the foreseeable future. In the longer term, employer nurseries may take on a new role since research in other countries suggests that their provision may be important in easing recruitment, and reducing absenteeism and turnover (OECD, 1990, p.145). 7.29 At present, charges in the State nurseries are on a flat-rate basis, regardless of income.9 A fairer arrangement, which existed up until two years ago, would involve a differentiated fee structure based on ability to pay. Nor should State nurseries discriminate against mothers working part- time as apparently was the case until recently. On grounds of equality of access between different parts of the country and between urban and rural areas, a continued State involvement in provision for nursery age children is, moreover, to be recommended.'° State financing could supplement that of the 8 CSO, Statistical Yearbook, 1988, Table 21.20. 9 The charge is Ft. 24/child/day in 1990, compared to an estimated cost for one nursery place of Ft. 6000/month (information from Ministry of Welfare). 10 Continued provision need not be in the current form. One possibility is the merging of nurseries and kindergartens, at present the responsibility of different Ministries (Welfare and Education respectively). 85 local governments through the formula grant system; a capitation grant for nursery enrollments could be added to the revenue-sharing system, possibly with conditions attached such as the introduction of such a differentiated fee system. (See Annex IX, "The Financing of Local Governments".) This would help ensure that nurseries are available to meet the effective demand, especially by low income families, during the difficult period of economic restructuring ahead. 7.30 The role of nongovernment providers--employers and private for- profit or nonprofit organizations--could be substantially encouraged through the tax system. Whatever public financial support is necessary to ensure access to private facilities for low income families could be provided to individuals through tax incentives (preferably in the form of a credit against tax due rather than an allowance against taxable income), although means- tested cash assistance to pay nursery fees is also a possibility. Tax breaks or other subsidization of day care operators to permit reduced fees for low income users might be considered as well.' 5. Conclusions and Recommendations 7.31 The review of maternity and child care benefits in Hungary, together with day-care provision for children of nursery age, suggests that substantial reform should be undertaken. The Report recommends that a reform package contain some or all of the following elements: - reductions in the length of paid maternity leave and in the earnings replacement rate during maternity leave. Employers could also be required to bear the cost of one month of the maternity leave. * a switch from earnings-related to flat-rate support after the period of maternity leave. - curtailment of the total length of paid child care benefits (including maternity leave) and statutory unpaid leave to at most two years. - continued State financial support for public day-care facilities for children of nursery age, possibly through the formula grant system conditional on introduction of a differentiated fee structure. - encouragement of employer and private nursery facilities via the tax system. tt Details of the tax-treatment of nursery and kindergarten facilities in OECD countries are given in OECD, Employment Outlook (1990, chapter 5). 86 7.32 The potential savings in expenditure by the government (whether by social insurance or the State budget) due to these reforms is roughly estimated as follows, at 1989 earnings and benefit levels and in percent of total maternity/child care benefit expenditure: - one month reduction in maternity leave Ft. 0.6 bn 4% (or shift of cost to employer) - reduction of earnings replacement rate for remaining maternity leave (5 months) from 100 to 80 percent Ft. 0.6 bn 4% - conversion from earnings-related to flat-rate child care fee Ft. 3.2 bn 21% - elimination of third year of child care leave Ft. 2.0 bn 13% 87 CHAPTER 8 SOCIAL ASSISTANCE AND SOCIAL SERVICES A. Introduction 8.01 An essential component of the "social safety net" in Hungary for those on low incomes is represented by social assistance and social services, both of which are largely administered and financed by local councils. Social assistance provides cash benefits to persons over age 18, either on a continuing basis or in the form of one-off grants to meet exceptional needs. These latter payments are restricted to a maximum of six per year to any claimant. A parallel system of child and youth welfare aid is administered by the Court of Guardians, providing both regular and exceptional types of support. Total social assistance payments amounted to Ft. 3.6 billion in 1990, and child welfare support totalled Ft. 1.3 billion in 1989. Details of these programs and descriptive data are provided in Annex VIII. Social services provide additional support through a variety of channels. These include (in addition to the nurseries discussed in Chapter 7) free or subsidized meals for the elderly, assistance with fuel costs, and domestic care. 8.02 As noted in Chapter 2, evidence from the household surveys conducted by the CSO indicates that there was already a substantial problem of poverty in the late 1980s; it is estimated that about one million persons in 1989, some 10 percent of the population, had incomes beneath the relevant per capita subsistence level. The numbers of recipients of regular social assistance (less than 50,000) and of regular child welfare aid (80,000) seem very low by this standard. In 1990, exceptional social assistance payments were made to 420,000 persons, or 4 percent of the population. The number of families receiving regular or emergency child support represented 18 percent of the number receiving family allowances in 1989 (Annex Table VIII.1). 8.03 Demand for such support can certainly be expected to rise sharply with the change to a market economy and economic re-structuring. In particular, the need will increase for social assistance to play a major role in providing income support during unemployment for those with no entitlement to benefit from the new unemployment insurance scheme. Evidence of increased demand for assistance is already evident (Table VIII.2). In Budapest, the number of persons receiving social support and the funds disbursed in the first half of 1990 were reported to be higher by 39 percent and 70 percent, respectively, compared to the first half of 1989. Most Councils ran out of their funds earmarked for social assistance and social services in 1990 by the end of August. In contrast to earlier years when applicants for social assistance were mainly elderly persons on minimum pension, during the last year an increasing number of claimants have been economically active, which is consistent with evidence discussed earlier on the changing nature of poverty.' I Article in Nepszabadsag, 26 October 1990. Note that with 30 percent inflation in 1990, the increase in real disbursements approximately matched the increase in number of claimants. 88 8.04 In this situation, the safety net for the poor needs urgent re- appraisal. The first part of this Chapter concentrates on social assistance, making references to social service provision where the issues under discussion are similar. One of these issues is the amount of central government involvement in this key area of social policy, although it is assumed here that local government will retain primary responsibility for the provision of safety-net benefits in cash and kind, in view of the apparent desire in Hungary for strengthened local government and a reduction in central government activity. 8.05 Part B considers the main problems with the current system of social assistance. These include the lack of defined entitlement, the funding structure, and the administration of the scheme. Part C discusses key issues for a reformed scheme of social assistance, while Part D summarizes our recommendations in this area. Section E briefly addresses the specific issues of social services and school meals. B. The Problems of the Current Social Assistance Scheme 8.06 There may be a temptation to see the problems of the current social assistance scheme as having entirely to do with the aggregate level of finance. The total level of expenditure on regular and irregular support to adults and children--Ft. 4.3 billion in 1989 or only 0.25 percent of GDP--is very low given the picture of need outlined above. Social assistance, like social services, is the responsibility of local councils who rely for the great majority of their finance on transfers from central government. (See Annex IX, "Financing of Local Government".) Local councils are likely to obtain a real decrease in their budgetary resources in 1991, and their estimated budgets assume that new local taxes will be levied although most of the governments lack any significant experience in local revenue administration. Viewed at the aggregate level, the main problem might seem one of providing local government as a whole with more money. 8.07 The funding at the micro level of individual local councils is as serious a problem, however, as discussed below. This aspect of the funding issue is bound-up with two other problems which are addressed first, entitlement and administration. 1. Entitlement 8.08 There appears to be no right to social assistance, in the sense that any individual demonstrating financial need (and meeting any other specified criteria) is guaranteed payment, with an appeals procedure providing for redress in the case of an assessment with which the claimant disagrees. This situation contrasts with other cash benefits and other elements of social services.2 The decision to award any benefit appears to be entirely at the discretion of the local council. The ministerial decrees concerning social assistance and child welfare support specify the minimum levels of aid and the conditions on which 2 For example, appeals on unemployment benefit claims are at the county level and Budapest City Council determines appeals on decisions concerning social services provided by the city's districts. 89 individuals may be considered eligible. One of the explicit conditions for regular child-raising support (in addition to evidence that the family's per capita income is below the minimum pension level) is a determination that this financial situation threatens the minor's development, which seems quite an unnecessary as well as arbitrary judgement. 2. Administration 8.09 As a result of local government re-organization in 1990, there are over 3,000 local councils in Hungary, varying in size from around 1,000 persons in rural areas to over several hundred thousand in urban areas. Each one of these councils is responsible for administering social assistance in its area and appears to adopt whatever practice it sees fit. 8.10 On the evidence of the Mission's visits, administrative practice varies enormously, being especially bureaucratic in rural areas. In a Budapest district, it was reported that the determination of social assistance eligibility was entirely in the hands of trained social workers and the process seemed relatively smooth. In a rural area in Barana county, on the other hand, each application requires processing by a variety of people, and the decision on each application requires a meeting of the local council. 8.11 In this situation, there must be a major issue of equity of treatment across the country, especially between large urban and small rural councils. It might be argued that detailed inspection within a small community produces the best assessment of an individual's true financial situation, but the possibility of capricious treatment also exists. Moreover, the process in rural areas, as described to the Mission, is obviously very public with the danger of stigmatizing receipt and depressing the level of take-up (although the very concept of take-up is difficult to discuss where eligibility is not clearly defined). The conditions for assistance as specified in the relevant decrees (see Annex VIII) are very explicit and restrictive for regular social assistance, but nonspecific for irregular assistance and emergency child support. This difference may explain the preponderance of the irregular forms of both adult and child assistance; however, it must be recognized that dependence on irregular aid increases the administrative burden (for repeat assessments) and may reduce take-up. There also appear to be inconsistencies among the aid programs regarding the requirements for officials or recipients to re-certify individual eligibility; the regular child-raising assistance seems to be the most stringent in this respect. 3. Financing 8.12 Local councils have to finance social assistance and social services expenditure almost entirely from their budgetary revenues, over 80 percent of which are funds transferred from central sources either in the form of State grants, personal income tax revenue, or SIF resources. There is also a small amount of copayment and contribution by private voluntary organizations in the case of social services. The importance of social assistance and social service expenditure presumably varies from council to council. In the councils the Mission visited in Baranya county, expenditures on these headings accounted for about half of their total revenue in 1989, indicating that the precise means of 90 financing of these items is a matter of considerable consequence. 8.13 The bulk of grants from the State budget to the local governments is made on the basis of normative formulas, as summarized in Annex IX. These formulas compensate the local settlements primarily according to their demographic composition and enrollments in educational institutions. The formula grants are not designed to account for the highly disparate economic characteristics of the various localities, which are indicated by the fact that the per capita PIT revenue originating in the settlements ranges from zero in the poorest to Ft. 28,000 in the wealthiest. Thus, the grants do not compensate in any significant degree for the unequal capacity of various council to generate their own tax revenues .3 8.14 The small size of many of the new local councils may be expected to greatly exacerbate this problem; almost half of the new 3,100 local self- governments have a population of less than 2,000. An event such as the closing down of one large employer could present a local council with an overwhelming demand for social assistance. The councils in poorer areas are those with the least recourse to other income sources, such as taxation of local business or access to credit to cover unforeseen needs. An additional problem is caused by the removal of the intermediate tier of government at the county level, which up to 1990 had redistributed funds among local councils to better match needs with resources available. 8.15 It would be impossible, of course, for the Ministry of Finance to achieve ex ante revenue sharing in a way that can anticipate all possible needs of individual councils. However, the formula grant system as designed appears to be especially ill-equipped to provide financial means to localities which are likely to have the largest demands for social assistance, i.e., those settlements with a weak economic base. In this way, the current structure of local government financing suffers from a potential problem of redistribution which may affect severely the functioning of programs needed to alleviate poverty during economic restructuring. This problem is in addition to that of the aggregate level of social assistance financing referred to earlier. C. Issues Relating to Reform 8.16 The problems identified above indicate a scheme in need of substantial reform. The Hungarian government has committed itself to preparing a new social assistance law during 1991 to be implemented from January 1992. The key issues to be considered in the preparation of this law are addressed below, along with a number of other problems with the present scheme. The principle of providing entitlement (the "right" to assistance) is generally accepted, but the crucial design questions concern the definition of the eligible population, the unit of assessment, the level of entitlement, and thus the implied financial requirements of the new system. 3 One component of the formula grant which provides a per capita subsidy for the "inactive" population (aged under 18 and over 60) is described as a proxy for the dependency burden in each settlement and thus its economic capacity. 91 1. The Eligible Population 8.17 Most social assistance and social services in Hungary have been provided in the past to a rather narrowly defined group-mainly the elderly without pension (particularly those over age 70), the blind, and disabled. Establishing a guaranteed income support to this population-which would largely amount to a minimum pension as mentioned in Chapter 4-would limit the eligible group to at most a few hundred thousand persons. It is proposed here that the eligible group be more broadly defined. The first issue is treatment of labor force status. Eligibility should not be restricted, for example, to the elderly without pensions or the unemployed with no benefit from the new unemployment insurance scheme; pensions and unemployment insurance benefit should simply be taken into account in the social assistance means-test. Similarly, there should not be need for a separate assistance scheme to provide support for children within families, as long as the means test for social assistance takes account of dependents. 8.18 In the case of the unemployed, registration for work with the local employment office could be made a condition for social assistance, in the same way as unemployment insurance is withheld if unsatisfactory efforts to find new work are made. This would obviously require considerable liaison between the local council administering social assistance and the local employment office. 8.19 The position of those in work is less obvious. Clearly, if social assistance were payable to the employed then a substantial disincentive problem could arise. On the other hand, some people in work, especially those with large families, might be able to demonstrate need according to the social assistance means-test. One solution would be to have a separate means-tested scheme for the employed (as in the U.K.), restricted say to those with dependent children, with a benefit withdrawal rate of substantially less than 100 percent as incomes rise. The aim here would be to supplement earnings rather than cover all remaining subsistence needs. 2. The Unit of Assessment 8.20 In determining eligibility, the current social assistance scheme appears to take into account, firstly, the income of other persons within the claimant's household and, secondly, the existence of relatives outside the household. The latter implies, for example, that the eligibility of an elderly person with no pension is based in part on whether there are grown children outside the home who could be providing support. Whether the children do actually provide support is apparently not questioned. It is strongly recommend that the practice of considering income of relatives outside the household be discontinued on grounds of both equity and administrative convenience.4 8.21 In principle, the nuclear family should be used as the unit of assessment for social assistance, rather than the household. The nuclear family is defined as the individual with his or her spouse together with any dependent 4 The limitation of relatives' responsibility may be thought to mark a transition from traditional Poor Law (Gordon, 1988). 92 children. Other relatives in the household are excluded as are other unrelated persons. A household consisting of a married couple with two dependent children and one grown child (or an elderly relative) would be two families according to this definition. 8.22 The restriction of the means-test to the nuclear family may again be justified on grounds of equity; it reduces the size of the unit in which income-sharing is implicitly assumed. But it may also be justified on grounds of incentives. Firstly, it removes the incentive to "push" non-dependent relatives out of the home so as to reduce household income and increase benefit payments. Secondly, it reduces the number of persons in the household facing high implicit marginal rates of tax as benefit is withdrawn with rising income. 8.23 The importance of distinguishing between family and household depends on the difference between the two in practice. Available Hungarian data do not provide an exact indication since families with grown children living at home are not separately identified from families with only dependent children or no children at all. An unknown additional proportion of households must therefore be added to the 14 percent which can be identified from Annex Table VIII.3 as containing more than one nuclear family. If only one-fifth of employed young persons aged 15-29 were living at home with their parents then the proportion of multi-family households would rise to over 20 percent.' 8.24 The choice between the two units of assessment may make a significant difference to the calculation of the numbers of persons in poverty. It should be noted that the figure of 1 in 10 of people in Hungary below the relevant per capita subsistence income line refers to per capita household income, and not family income. In other words, 10 percent of the population live in poor households rather than poor families. One would expect the amount of poverty to be higher on a family income basis and this has certainly been the conclusion in the UK where there has been a number of analyses of this issue.' The unit of assessment will therefore have important implications for the numbers of persons likely to be eligible for social assistance within the designated target group; therefore, the shift to the nuclear family may have to be made gradually, in line with the availability of financing. 3. The Level of Entitlement 8.25 At present, the cut-off level of income for social assistance is the minimum widow's pension (Ft. 4400/month at end-1990), reflecting the group which historically has formed the bulk of recipients. However, there is no particular reason to continue with this practice. 8.26 An obvious choice would seem to be the subsistence minima income levels calculated for different household types by the CSO; the rates for 1989 are given in Annex Table VIII.4. The purpose of social assistance is to provide 5 In 1988, there were 1,243,500 persons aged 15-29 working in state or co-operative enterprises (Statistical Yearbook, 1988, Table 4.4). e.g., Atkinson et al (1981) and Beckerman and Clark (1982). 93 a minimum income and the CSO calculations are estimates of the minimum levels of income required to buy the mere necessities of life. However, on closer inspection, there are a number of problems with the use of this benchmark. Firstly, the minima refer to per capita household income, which would not necessarily be the appropriate income level if the nuclear family were to be adopted as the unit of assessment. Even with a household means-test there might remain a problem since the operational definition of a household for benefit purposes could depart significantly from the standard survey definition. 8.27 Secondly, there is the problem of housing costs. These vary enormously across households and cannot be adjusted by the household in the short-term. The CSO subsistence minima include an average housing cost element; using this as the basis for benefit assessment would deny benefit to some individuals with greater than average housing costs and provide disproportionately generous payments to others. This would imply problems of targeting and equity. A reformed social assistance scheme should rather pay up to a specified income level which did not include any allowance for housing costs; the housing costs could then be paid as an addition.7 8.28 Thirdly, the subsistence minima take no account of important design of an assistance benefit scheme, such as incentives to work. In particular, the calculated subsistence minima are higher for active households, while on grounds of incentives there would be an argument for social assistance payments for active persons to be lower. 8.29 None of these objections to the use of the subsistence minima as the starting point for the calculations is insuperable. A range of adjustments could be made, e.g. making the family sum of household per capita subsistence incomes the cut-off level (with some adjustment for the different economies of scale), subtracting the average housing cost element and adding actual housing costs, reducing the level paid by a given amount for active households, or for all households on grounds of budgetary cost. However, there is a more fundamental objection to those just made. The calculation (from 1984) and finally publication (more recently) of the subsistence minima have represented important landmarks in the objective analysis of poverty in Hungary. The use of these minima as benchmarks for benefit amounts would run the danger of politicizing the calculations. This would lead to a decline in their reputation and function as objective statistical measures at precisely the time when the need for measurement of low income is growing. 8.30 The above arguments suggest that the level of social assistance benefit should be determined independently from the CSO calculations, although the CSO methodology could obviously play a role in the process. It is recommended that the CSO continue to publish calculations of subsistence minima and that the practice is adopted of regularly publishing estimates of the numbers of persons with incomes below these levels. In this -way, there would be a strengthening of the role of the CSO minima as objective statistical benchmarks against which changes in low incomes in Hungary could be assessed. 7 This is consistent with the World Bank proposal for the introduction of a housing allowance scheme (World Bank, 1990c). 94 4. Costs and Financing 8.31 The above three design issues will determine the cost of a reformed social assistance scheme. If the program is defined in terms of a very narrow population (e.g., at the most limited, persons over 70 without a pension), then means testing is not really needed for targeting, although it may still be necessary to limit expenditure in line with budgetary constraints. Assuming, for example, that the end-1990 minimum widow's pension of Ft. 4400 is granted to the number of persons who currently obtain regular social assistance or regular child welfare (130,000) and to one-third of the recipients of irregular aid (about 200,000), the resulting total cost would be about Ft. 17 billion, or 0.8 percent of GDP. Such expenditure would not be out of line with expected social assistance requirements in a country undergoing restructuring, but many persons experiencing poverty would also not be served.8 8.32 To provide a safety net for a wider range of cases of economic hardship, social assistance should not be restricted a priori to only certain groups in the population; however, the payment by right can then only be made on the basis of a means test, and should be determined as filling the gap between the claimant's proven income and an income ceiling which should be set as low as necessary based on the budgetary means available. Conceptually, the cost of this type of scheme is represented by the "poverty gap" mentioned in Chapter 2. The calculation of this gap for 1989 based on survey data indicated a cost of Ft. 10 billion to bring the estimated share of persons below subsistence (10 percent of the population) up the subsistence income level (see Annex II, para. 24). Inflating the 1989 subsistence income level to 1991 prices, and assuming that the average income of the poor in 1991 is the same proportion of the subsistence level as in 1989 (about 80 percent), meeting the poverty gap for 10 percent of the population in 1991 would cost about Ft. 16 billion, or very close to the above estimate for the more restricted but non-means tested approach.9 If poverty has increased since 1989 to, say, 15 percent of the population, the estimated cost would rise to about Ft. 24 billion (almost one percent of 1990 GDP). Such an expenditure would only be financeable for Hungary if other, less justifiable, social expenditures were reduced in the government budget. 8.33 Financing such a program of social assistance entitlement would be impossible for the local governments to manage on their own, and indeed such programs are not financed solely at the local level in Western countries. The revenue-sharing arrangement between central and local government should in some way take account in some way of the foreseeable needs for the social assistance program. One option would be for the central budget to provide a block grant a Note that in Table 2.2, "other cash assistance" is shown as Ft. 17 billion in 1990--however, this includes a residual amount of social income expenditure which is not identifiable. 9 The estimated subsistence income level in 1991 (after inflating the weighted average of that for active and inactive households in 1989 by the annual CPI increase) would be Ft. 6390/month. Taking 20 percent of this as the average per capita income gap for the poor (Ft. 1278/month or Ft. 15,336/year) times one- tenth of the population (1.057 million persons) equals Ft. 16 billion. 95 to the local councils for the provision of social assistance and related social services, with the amount of this grant determined ex ante on the basis of some objective indicators of economic need in the area (e.g., per capita income, unemployment, business turnover, etc.). 8.34 The issue of financing is closely bound up with the wider issue of defining central and local government responsibilities for social assistance. All aspects of the administration (including determination of eligibility and payment) should remain the responsibility of the local councils, as at present. The central government (and the Ministry of Welfare seems the obvious agent) should lay down a clear set of procedures and guidelines for means testing; these should be enforced so as to provide a minimum standard in administration and to promote equitable treatment across the country. The minimum level of entitlement should be determined at the center, and should take account of regional differences in cost of living and a realistic assessment of the overall level of social assistance which is financially supportable. Individual councils should then be free to provide more generous amounts of funding from their own revenues if they are willing and able to do so. The central government should also actively promote take-up of social assistance; it may be particularly important to reduce any stigma that could be associated with receipt in smaller settlements. 8.35 Finally, there is another method of providing a minimum income safety-net, that is, transformation of the current personal tax allowances into a system of refundable tax credits (negative income tax). Such a system would automatically provide some cash income to those on low incomes and has much to recommend it. However, it would be very hard to completely remove the necessity for some kind of means-tested safety net. It would be difficult to design a system of tax credits which could match the heterogeneity of circumstances which can be addressed in means-tested schemes. And, most importantly, the level of the credit sufficient to remove persons from a situation of need would imply much higher marginal tax rates than those presently in place. D. Conclusions and Recommendations 8.36 This Chapter has discussed the problems with the current social assistance scheme run by local councils and considered ways in which it could be substantially reformed. The expenditure on benefits under such a social assistance program is estimated to cost on the order of one percent of GDP at the scale of poverty measured in the recent past. Chapter 10 presents medium- term projections showing the growth of social assistance in parallel with the growth of unemployment compensation, but in the context of reform in other items of social expenditure. 8.37 To summarize the recommendations of this Chapter: - social assistance should be an entitlement, as of right, for those individuals who can demonstrate need according to a means test which takes full account of all dependents. Persons receiving social insurance benefits such as pension or unemployment insurance should be eligible for social assistance if their income including such benefits falls below the designated cut-off level. The level of entitlement should be determined 96 in line with budgetary resources. the administration of social assistance and the means test should follow certain central guidelines so as to promote equity of treatment across the country. a fair system of distribution of funds for social assistance financing needs to be devised; the system of revenue-sharing between central and local government should be adapted to recognize the large disparities among settlements in economic capacity and likely incidence of poverty. the practice of taking account of relatives outside the household in means testing should be discontinued. As budgetary means permit, the unit of assessment should be made the nuclear family instead of the household. the presently separate schemes for social assistance and child welfare support should be merged, at least for the unemployed, in the interest of simplifying administration; for persons in work with dependents, however, a supplementary means-tested child benefit may need to be designed separately to address circumstances of very low earned incomes but with a concern to maintain work incentives. the CSO should continue to calculate and publish objectively-defined subsistence minima together with estimates of the number of persons falling below these minima. E. Social Services and other Social Income in Kind 8.38 Social services (mainly directed to the elderly and handicapped) are an important part of the health and welfare system in Hungary. These services are financed largely through central budget transfers to the councils. The councils manage the large institutions which are not operated by private (e.g., church) organizations, such as the nursing homes; and the community- based activities are financed and operated jointly by the councils and voluntary groups. For the population groups in question, the social services are designed as a continuum of support intended to allow individuals to remain with their families or live independently for as long as possible. For the elderly, this system consists of: food catering, Elders' Clubs (multiservice, nonresidential care centers), a cadre of home helpers, professional domestic caretakers, and nursing homes. Although admirable in design, this network is extremely thin, as the data on available places indicate it covered only 12 percent of the population over 60 in 1989. 8.39 Expenditure of State and council funds on services for the elderly amounted to roughly Ft. 5 billion in 1989, to which would be added nominal co- payments by most of the users and contributions by enterprises and the community. The schedule of required copayment established by the Ministry of Social Welfare is given in Annex Table VIII.5. Since 1985, the Ministry of Social Welfare has also supported the establishment of Family Assistance Centers to provide a wide range of counseling and assistance to families as a unit; only 114 are in operation to date, and the Ministry estimates that 500-600 are needed. 97 8.40 The Ministry plans to seek additional private support to provide more social services, and is encouraging voluntary organizations and settlements to set up facilities with matching grants from the central and local government budgets. Although such activities are properly derived from local community support, the vast unmet need for social welfare services and the almost complete lack of trained social workers call for a more active financial support from government, both central and local, at least in the early stages of expanding the network of facilities. Additional financial support need not imply a net increase in costs to government, but should be considered in the context of reallocating other social policy expenditures which are currently less well targeted to urgent social needs.'° There may be a case for more explicit cost- sharing between central and local government for certain social services." 8.41 Better integrating social work and health services, to provide more, options for appropriate care of the elderly and follow-up of persons discharged from treatment centers, could ultimately save some resources. The amount of financing that might be needed, for example, to modestly improve the remuneration of community service workers and staff of nursing homes in order to attract more personnel to this work, would be small relative to the other components of social policy and to the cost of hospital care. Such interventions among the elderly and families can make a major contribution to their mental and physical health. Indeed, they may be effective approaches to prevent social welfare problems such as alcoholism, suicides, and divorces which are particularly prevalent in Hungary, as well as provide mutual support for individuals during the particularly stressful period of the economic transition. 8.42 The Mission explored briefly the issue of State support for school meals. The State subsidy was completely lifted within the last couple of years, with the result that many families can no longer afford school meals. The local councils, to the extent of their means, have restored some subsidy for low income families based on demonstrated need. Studies from other countries indicate that school feeding programs can be one of the most cost-effective uses of state funds to support the food budgets of low income families, and to provide nutritional improvement to this vulnerable age group. Local governments in the poorer settlements, which would have the largest numbers of school children needing subsidized school meals, would be unable to continue an adequate level of financial support on their own under the new local government financing arrangements--another instance of the problem highlighted above with social assistance. 10 For example, if the pension of persons admitted to long term nursing homes was taken as a copayment for the total care provided, the institution's financing would be improved with no new cost to government. Put another way, there is scope for increased private contributions to the cost of such care. tt The formula grant from central to local government includes per capita funding for numbers of residents in homes for the elderly, but not specifically for social services such as Elders Clubs or Family Assistance Centers. 98 8.43 It is recommended that the central government consider reintroducing some degree of State financial support for a schools meal program. The cost of complete subsidization would be roughly Ft. 8.1 billion at 1990 prices.12 Full subsidization is not recommended in light of budgetary constraints, although given the evidence on the incidence of primary school expenditure (Chapter 3), targeting would not be much of an issue even with full subsidy. An alternative approach would be to subsidize a school milk benefit for all students, which would be estimated to cost about Ft. 3 billion in 1990.13 12 Estimate based on current nonsubsidized charge of Ft. 600/month (Ft. 30/meal/day), assuming 9 months or 180 days of school year and 1.5 million students aged 5-14. 13 Assuming one daily serving of milk at Ft. 12/serving, with the same number of school days and number of students noted above. 99 CHAPTER 9 SCHOLARSHIPS FOR STUDENTS 9.01 The system of State scholarships for students in higher education is an area where important changes have been made recently. Prior to the 1990- 91 academic year, the central government budget provided three types of financial assistance to tertiary level students for living expenses: scholarships based on test results or academic performance (received by about 60 percent of students); a "social subvention" based on family income (received by another 60 percent and averaging Ft. 1200/month in 1989), and a uniform cost of living allowance provided to all students (also Ft. 1200/month in 1989). Before September 1990, the Ministry of Culture determined the allocation of student aid among the three forms of assistance and divided the funds which were earmarked from the central budget for this purpose among the universities. The funds for income-targeted assistance were allocated among the institutions based on the average income of their student populations. Until January 1990, parents were required to apply for this form of assistance by declaring their previous year's earned income, as verified by the employer (with the employer required to respond to such requests for verification). The awarding of merit scholarships was based on detailed performance criteria set by the Ministry. 9.02 With the 1990-91 academic year, the allocation of central government financial assistance funds has been transformed into a capitation grant system. The total available funds are divided by the total student enrollment, and the resulting average (Ft. 30,000 per capita in the current year) is then transferred to the individual universities solely on the basis of their enrollment. Each institution is now free to determine the allocation of their aid funds among the three main uses, and to set the selection criteria. The universities may therefore choose their own methods for means-testing, and are no longer permitted to require employers to verify the income of applicant families for such purposes. To date, the institutions are beginning to experiment with various income declarations to determine eligibility for the social subvention-type of assistance, and in many cases student committees (a form of peer review) are providing considerable input to the award decisions in individual universities. 9.03 The reform introduced in 1990 is intended to put the student aid system on a "normative" basis--i.e., to reduce the allocative discretion of the Ministry by shifting to an objective capitation grant system. However, the new approach increases the potential inequities within and among the educational institutions. The system is intended to encourage educational institutions to compete for students through the use of financial assistance funds, which each university receives on an equal (per capita) basis. The approach will be disadvantageous, however, to institutions with a larger than average number of low income students, and could lead to reduced access to financial aid for needy students. It is important that some standards for means testing be maintained for the country as a whole, and that the availability of needs-based assistance for low income students not suffer too greatly by universities' efforts to outbid each other for the best students. 100 9.04 It is worth noting that in some countries where financial aid to students plays a very important role (e.g., the United States), the government no longer provides financial assistance based on merit but only based on need, on the grounds that the government's responsibility is to promote educational attendance and guarantee access for needy students, while it is assumed that good performers without financial need will pursue their education without such assistance. There has also been a shift over the years towards the provision of financial aid funds to the students, rather than to the educational institutions--that is, the students who meet the government's criteria for financial need are given the assistance funds directly to be used at the university of their choice, although the application and disbursement process may be administered by the university. Providing the funds to the students directly permits the students to "shop around" among educational institutions, thus fostering the desired competition among schools on the basis of their educational programs, not their availability of aid funds. The approach also ensures that the allocation of assistance takes account of the actual financial need of the total student population, and is more efficient than an ex ante allocation to individual universities which may not be consistent with the income profile of their current enrollments. 9.05 Given the large increase in university enrollments planned in 1991- 92, it is also important that the total supply of financial aid to students keep pace. To be affordable, additional student aid will increasingly have to take the form of loans and work/study. Student loan programs can take the form of government subsidy of a share of the interest, and/or guarantee of loan repayment, with loans made either by commercial banks (on certain conditions) or by a government agency. Loan subsidies, however, require rationing to accommodate budgetary constraints. An alternative approach which has been recommended in the U.K. is income-contingent loan repayments--that is, the repayment is made a function of income achieved after graduation (Barr, 1989). Programs for work/study involve wage subsidies for part-time student employment. In the United States, for example, such programs account for much of the work force on campus in food services, libraries, etc. 9.06 It is to be expected that private voluntary sources should eventually supplant an important share of State budgetary funds for student assistance. Hungary already has the practice of the "employers' scholarship", whereby a company sponsors the education or training of an individual who promises to work for the employer. However, in an environment of rapid structural change, as presently exists in Hungary, it is risky for private business to finance future graduates in particular professional fields because of the difficulty of predicting market supply and demand for specific skills in the future. There can also be a "free rider" problem: one employer pays for the educational expenses of an individual who spends most of his career with another firm. It may therefore be preferable to encourage companies to contribute funds which can be "pooled" for education, but not specifically for their future employees. 101 While the difficult economic prospects of Hungary in the near future may make other types of private contributions to education scarce for a time, it is also worth attempting to harness the generosity of Hungarians living abroad. There are many such schemes that could be explored.' For example, a society of Hungarians in a particular city in the West could be asked to "adopt" a specific university in Hungary. Alternatively, a university in Hungary could attempt to identify highly successful past graduates or other Hungarians living abroad to develop a "hall of fame". Each person inducted into the "hall of fame" could be asked to sponsor a student--both as a mentor and financial sponsor. (The references in this section to financial aid practices in other countries have been provided by Ms. Alice Diamond, a financial aid consultant in the United States.) 102 CHAPTER 10 MEDIUM-TERM PROJECTIONS OF SOCIAL POLICY AND EXPENDITURES 10.01 From a macroeconomic perspective, the main structural issues of the social policy framework in Hungary, as discussed in the preceding chapters, include: (i) the relatively high level of national income devoted to social transfers, (ii) the high and growing share of social income in household income, (iii) the financial unsustainability of the pension program, and (iv) the heavy burden of social insurance taxation on employers. Within the individual social programs, there is need for the growth of total pension outlays to be curtailed through tightened eligibility; a reduction of expenditures through structural reform of sick pay, family allowances and child care benefits; and an increase in the level of social expenditures devoted to unemployment support and social assistance. Reference has also been made to recent studies (World Bank 1990a,b,c) which support the case for structural reform to reduce housing subsidies, and greater efficiency in the use of government expenditure allocated to health care and education. Both these previous studies and the present Report also argue for more adequate burden-sharing between the private and public sectors in the future, both through a continued significant level of copayment by beneficiaries of housing, education, and health care, and by increased private provision of these and other social services, pensions, and educational scholarships. This Chapter attempts to translate these arguments into a feasible scenario for reform in social expenditure in the context of future economic developments. 10.02 The projections below link assumptions about the major macroeconomic trends expected over the next decade with alternative patterns of social expenditure. The most important of these macroeconomic parameters are listed in Table 10.1. The values chosen for the projection period reflect assumptions about the course of the adjustment process in Hungary: after an initial period of declining or stagnant growth during the next couple of years, the performance of GDP should rebound in the second half of the decade; unemployment will correspondingly rise sharply in the next few years and then settle towards the end of the period to the kind of "steady state" rate observed in many market economies; households' share of national resources will remain steady or marginally decrease, while the share absorbed by the government sector will decline distinctly to allow a greater role for private enterprise. Although there are obviously many different macroeconomic scenarios that could be considered faasible for the next ten years, these basic parameters are not altered here in order to focus attention on the effects of changes in social policy. The scenarios are based on the projected demographic trend described in Chapter 2 (Table II.20). 10.03 The projections below are made in three main steps. First, the expenditures (by government) for individual social programs are projected according to information and assumptions about their internal dynamics, demographic factors, other economic developments, and program reforms. Second, after thus deriving "social income in cash and kind" and the subset of social insurance expenditures, the other components of household income and taxation are projected; the resulting balance of social and labor income, and 103 TabLe 10.1 Macroeconoric Parameters for Mediun-Term Projections 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 Real GDP Growth Rate (X) -4.70 -2.69 -2.09 1.21 4.87 4.72 4.48 3.44 3.25 3.79 3.00 (Average 1990-2000 = 1.75% p.a.) CPI Growth Rate (%) 30.0 28.0 26.0 24.0 22.0 20.0 18.0 16.0 14.0 12.0 10.0 (Average 1990-2000 = 20.0%) Growth of Real Disposable Income (X) 2.3 -4.2 -3.6 -0.4 3.2 3.0 4.5 3.4 3.3 3.8 3.0 (Average 1990-2000 = 1.66%) Disposable Income/GDP (%) 59.6 64.0 63.0 62.0 61.0 60.0 59.0 59.0 59.0 59.0 59.0 59.0 Consol. State Budget Exp./GDP (Y.) 54.7 52.5 49.9 48.5 46.0 45.0 43,0 40.0 39.0 38.0 37.0 36.0 General Govt. Expend./GDP (%) 60.8 58.0 58.1 56.5 55.0 53.5 52.0 50.5 49.0 47.5 46.0 45.0 Unemploy)ment (X of Labor Force) 0.6 2.0 5.0 10.0 12.0 10.0 8.0 6.0 5.0 4.0 3.0 3.0 … … a281.n298 104 total social expenditure, are adjusted through iteration until a desired structure of incomes, social insurance expenditures and contribution rates are achieved. Third, assumptions are made about private expenditure on social programs, and the total expenditures by both government and households are then derived. Details about the assumptions used and the approach to deriving results for the individual programs are provided in Annex X. The overall projections were only made to year 2000 because uncertainties beyond this period do not make further extrapolation useful, except in the case of pensions and health care which are most affected by long-term demographic trends (the growth of the elderly population). Chapter 4 discussed the pension projections to year 2051, and the assumed linkage between population structure and health expenditures is explained in Annex X. A. The Baseline Projection 10.04 The Base Case scenario attempts to illustrate the implications of continuing most of the existing programs of social insurance without structural reform. The main outcomes of this scenario are summarized in Table 10.2 (with more detail in Annex Tables X.1 and 2). As already indicated in Chapter 4, the largest single social program, pensions, would in this case remain above 9 percent of GDP through year 2000. The necessary social insurance contribution rate' to finance pensions alone would rise from 30.2 percent in 1990 to 36.4 percent in 2006 and over 47 percent from 2031 on. As a share of GDP, pension expenditure would amount to about 16 percent from year 2031 (Table 10.2, last section). 10.05 Assuming no change in the current per capita payments or eligibility requirements for sick pay, family allowances, maternity/child care benefits, and unemployment compensation, the expenditures for these programs would be determined only by changes in numbers of beneficiaries. For health care and pharmaceuticals, however, a significant reduction in per capita costs is assumed (reflected in a decline in the elasticity of growth of medical expenditure relative to GDP). Despite this change starting already from 1992, the projected growth of the elderly population (over 60) would still keep the share of health and pharmaceutical expenditures as a roughly constant share of GDP to 2000. The long term cost projection on these assumptions indicates that the implied contribution rate for health care and pharmaceuticals would 1 "Contribution rate" is the level of projected social insurance expenditure divided by the projected contribution base. The "contribution base" is defined in the pension projection model as the share of labor income in GDP (taken as a fixed 43 percent each year of the projection period) times the share of labor income which counts for social insurance purposes as the contribution base (projected as a fixed 76 percent each year). Therefore, the contribution base is estimated as (.43 x .76) or 33 percent of GDP in each projected year. This convention is used in this Chapter for determining the implied contribution rate of nonpension elements of social insurance as well; however, since the separate components of labor income are projected here, it would also be possible to compare the social insurance expenditures to any of these income components to derive the implied contribution rates. 105 Table 10.2: Sunmry of Medium Term-Projection - Base Case 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 Government Social Expenditures as % of GDP of which: Pensions 9.1 9.9 10.0 10.0 9.8 9.4 9.1 9.0 9.1 9.2 9.4 9.7 Sick Pay 1.2 1.2 1.1 1.1 1.2 1.2 1.2 1.2 1.2 1.3 1.3 1.3 Health and PhamnaceuticaLs 4.1 4.5 4.7 4.9 4.9 4.8 4.7 4.6 4.7 4.7 4.8 5.0 Maternity ard Child Care 0.9 0.9 0.9 1.0 1.0 1.0 1.0 1.0 1.1 1.1 1.1 1.2 Family Allowances 3.1 3.1 3.2 3.2 3.1 2.9 2.7 2.6 2.5 2.4 2.4 2.4 Solidarity Fund + Enployment Fund 0.1 0.5 1.2 2.7 3.4 2.7 2.1 1.6 1.3 1.0 0.8 0.8 Social Assistance (cash) 0.5 0.9 1.0 2.7 3.3 2.7 2.1 1.5 1.3 1.0 0.7 0.7 Education (incl. Scholarships) 4.5 4.7 4.7 4.7 4.7 4.7 4.7 4.7 4.7 4.7 4.7 4.7 Housing + Consumer SLbsidies 7.2 5.8 4.1 2.7 2.1 1.6 1.4 1.3 1.2 1.1 1.1 1.0 Total Social Expenditures 32.6 33.3 32.0 33.9 34.2 31.7 29.9 28.3 27.7 27.3 27.0 27.4 (as % of General Govt. Experditure) 53.6 57.4 55.1 60.0 62.2 59.3 57.4 56.1 56.6 57.6 58.6 61.0 Govt. plus Private Expenditures as % of GDP Pension(+ Priv. Net Finan. Savings) 9.2 14.1 14.0 14.0 13.8 13.3 13.2 13.3 13.6 13.9 14.3 14.7 Health 5.1 5.6 5.8 6.0 6.0 5.9 5.7 5.7 5.7 5.8 5.9 6.1 Education 5.4 5.7 5.7 5.7 5.7 5.7 5.7 5.7 5.7 5.7 5.7 5.7 Housing and Utilities 13.7 12.5 12.1 10.9 10.1 10.3 10.1 10.5 10.4 11.4 11.4 11.4 Implicit Contribution Rates for Social Insurance (as Y of Contribution Base) Solidarity Fund (uneploy. insur.) -- -- 2.0 5.6 6.9 5.6 4.4 3.2 2.6 2.1 1.6 1.6 Other Social Insurance Programs 53.0 53.0 53.0 52.8 52.4 50.8 49.8 49.5 49.9 50.8 51.7 53.4 Total Social Insurarce Programs 53.0 53.0 55.0 58.4 59.2 56.4 54.2 52.7 52.6 52.9 53.3 54.9 of which: Rate for Enployees 10.0 10.0 10.5 12.1 12.0 11.8 11.7 11.7 11.8 11.8 11.9 11.9 Rate for Enployers 43.0 43.0 44.5 46.8 48.2 47.1 46.0 44.8 44.2 43.6 43.0 43.0 Net Personal Inccie as % of GDP 62.1 67.7 69.9 72.0 72.4 69.9 67.7 66.4 66.1 65.9 65.8 66.5 as % of Net Personal Inccme: (100) (100) (100) (100) (100) (100) (100) (100) (100) (100) (100) (100) Net Earnings 59.0 59.4 60.0 56.6 55.6 56.9 58.0 59.3 59.8 60.2 60.6 60.3 Social Incane in Cash and Kind 41.0 40.6 40.0 43.4 44.4 43.1 42.0 40.7 40.2 39.8 39.4 39.7 Total Cost of Labor to Enterprises 61.1 63.3 65.7 65.8 65.5 64.5 63.4 63.1 63.1 63.0 63.0 63.1 (in terms of % of GDP) of which all Social Insurarce 14.0 14.0 14.6 15.3 15.8 15.4 15.0 14.7 14.5 14.3 14.1 14.0 Long-Term Projections Year 2006 2011 2016 2021 2026 2031 2036 2041 2046 2051 Govt. Pensions as X of GDP 11.9 13.4 14.0 14.1 14.5 15.8 16.9 16.2 15.7 15.5 Govt. Health & Pharm. as X of GDP 5.7 6.6 7.1 7.3 7.8 8.5 9.2 9.2 9.1 9.1 a411.n465 basegmr.wkI 106 rise from about 14 percent during the 1990s to 17 percent in 2006 and over 26 percent (over 9 percent of GDP) after 2031. 10.06 Thus, for all of the present programs of social insurance, the total implied contribution rate required to just balance the Social Insurance Fund (SIF) would remain between 50-53 percent annually up to 2000.2 The temporary reduction in the rate during the second half of the decade would be countered by the need for an average contribution rate for the Solidarity Fund (unemployment insurance) of 3-4 percent given the projected unemployment. It would therefore not be possible to reduce the total social insurance contribution rate, as is necessary to improve economic incentives. 10.07 Concerning the programs not included in social insurance, the projection assumes that: (i) support from the Employment Fund (e.g., retraining) amounts to half of cash compensation of the unemployed (the Solidarity Fund) each year; (ii) social assistance also keeps pace with the growth of unemployment insurance (to represent the expected increase in poverty during the toughest adjustment period); (iii) social services (in kind) grow at about the same rate as GDP; (iv) education and scholarship expenditures maintain their share of GDP; (v) other in-kind components of social income (mainly culture and recreation-related) are phased-out by 1993, i.e., passed over to the private sector; (vi) housing subsidies sharply decline in real terms over the decade, and other consumer subsidies are also eliminated by 1993.3 10.08 On these assumptions, the Base Case indicates that total social income (cash and kind) would not be reduced to any significant degree as a share of GDP, and in fact would increase in 1992-95. Due to the cut in subsidies, total social expenditure would fall slightly more, from 33.3 percent in 1990 to 27.4 percent in 2000. As a share of household income, social income would increase during the middle years of the projection and by 2000 settle to almost the same level as in 1990, about 40 percent. Assuming that employees contribute a constant 10 percent of their "main earnings" to social insurance and employers pick up the remainder of the contribution requirement, the projection shows that the social insurance taxation of employers would increase for several years rather than fall relative to the 1990 rate. The total cost of labor to employers (gross wages and social insurance contributions by employers as a share of GDP) would be basically unchanged over the decade. The main outcomes of the Base Case are illustrated in Figures 10.1-10.4 at the end of this Chapter. 2 For simplicity of presentation, the SIF is used here to combine all social insurance programs other than unemployment insurance (which is covered by the Solidarity Fund); however, the individual social insurance programs might be separated in 1992 into separate accounts, each with its earmarked share of the social insurance contribution rate. 3 The elimination of most consumer subsidies by 1993 is a feature of the Government's medium-term program as agreed in 1990 with the World Bank in the first Structural Adjustment Loan. 107 10.09 In conclusion, the Base Case demonstrates a trend of pension expenditure which would imply a sharply worsening financial burden after 2000. Even assuming some reform in health expenditure during the present decade, the projection does not indicate any possibility for the overall social insurance contribution to be reduced. There is thus little likelihood of alleviating the financial burden on employers or the overall drain of social transfers on the economy's resources, a scenario which makes it even less likely that the economy can rebound during the second half of the decade and achieve the kind of performance projected for it here. 10.10 It should be noted that if a lower average growth of CDP over the decade is assumed, most of the social programs as projected in the Base Case would create a much higher financial burden. As one illustration, assuming that the annual growth of GDP increases in a linear trend to only 2 percent in 2000 and thereafter, and unemployment peaks at 10 percent only by 2000-- assumptions which might reflect a very delayed economic adjustment--the projection model implies that the pension expenditure alone would amount to 12.1 percent of GDP in 2000. The implied contribution rate for pensions would rise steadily from 1990 to reach 36.9 percent already by 2000 and peak at 60 percent by 2036. On the same assumptions discussed above for the other social insurance programs (excluding unemployment), the total implied contribution requirement would rise to 61 percent by 1995 and 69 percent by 2000. It should be evident that the risks for the economy of delaying significant structural reforms of social expenditure programs are very great indeed. B. Alternative Scenario 10.11 The discussion of pension reform in Chapter 4 considered five possible variants based on assumptions regarding the reduction in the eligibility ratio and resulting increase in the effective retirement age. The Alternative Scenario here employs the fifth pension alternative, i.e., that with the steepest decline in the old-age eligibility ratio after 2000. This assumption permits pension expenditure to decline to about 8 percent of GDP in 1995-2000 and stabilize at about 10 percent of GDP from year 2011 (Table 10.3). The pension contribution rate would fall to about 24 percent in year 2000 compared to 30 percent in the Base Case (Annex Tables X.2 and X.4). 10.12 For the remaining social programs, the Alternative Scenario indicates the implications of the main structural reforms discussed in the preceding chapters. Concerning sick pay, it is assumed that three-fourths of the expenditure projected in the Base Case is removed from the SIF and passed on to employers. With periods of sick leave currently averaging about four weeks (30 calendar days) per employee per year, this change would roughly correspond to shifting the first three weeks of any episode of sick leave to employers and leaving the SIF to finance only longer periods. In fact, with the incentives thus changed, employers' and employees' willingness to claim sick pay should be curbed somewhat, so that the projection might overestimate the total future expenditure by employers for this benefit. 108 TabLe 10.3 Sumrary of Medium Term-Projection - ALternative Scenario 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 Goverrent SociaL Expenditures as % of CDP of which: Pensions 9.1 9.9 10.0 9.6 9.2 8.6 8.2 7.9 7.8 7.8 7.7 7.8 Sick Pay 1.2 1.2 1.1 0.3 0.3 0.3 0.3 0.3 0.3 0.3 0.3 0.3 Health and Phanirceuticals 4.1 4.5 4.7 4.9 4.9 4.8 4.7 4.6 4.7 4.7 4.8 5.0 Maternity and ChiLd Care 0.9 0.9 0.9 0.6 0.7 0.7 0.7 0.7 0.7 0.7 0.8 0.8 Family ALlowances 3.1 3.1 3.2 2.5 2.4 2.3 2.1 2.0 2.0 1.9 1.9 1.9 Solidarity Fund + Enployment Fund 0.1 0.5 1.2 3.1 3.8 3.1 2.4 1.8 1.5 1.2 0.9 0.9 Social Assistance (cash) 0.5 0.9 1.0 2.7 3.3 2.7 2.1 1.5 1.3 1.0 0.7 0.7 Education (;rct. Scholarships) 4.5 4.7 4.7 4.7 4.7 4.7 4.7 4.7 4.7 4.7 4.7 4.7 Housing + Consumer Subsidies 7.2 5.8 4.1 2.7 2.1 1.6 1.4 1.3 1.2 1.1 1.1 1.0 TotaL Social Expenditures 32.6 33.3 32.0 32.0 32.2 29.5 27.4 25.7 24.9 24.2 23.6 23.9 (as % of General Govt. Expenditure) 53.6 57.4 55.1 56.6 58.5 55.2 52.8 50.9 50.7 51.0 51.3 53.0 Govt. plus Private Expenditures as % of GDP Pension(+ Priv. Net Finan. Savirgs) 9.2 14.1 14.0 13.6 13.2 12.5 12.3 12.2 12.3 12.5 12.6 12.8 Health 5.1 5.6 5.8 6.0 6.0 5.9 5.7 5.7 5.7 5.8 5.9 6.1 Education 5.4 5.7 5.7 5.7 5.7 5.7 5.7 5.7 5.7 5.7 5.7 5.7 Housing and UtiLities 13.7 12.5 12.1 10.9 10.1 10.3 10.1 10.5 10.4 11.4 11.4 11.4 Implicit Contribution Rates for Social Insurance (as % of Contribution Base) Solidarity Fund (uremploy. insur.) -- -- 2.0 5.6 6.9 5.6 4.4 3.2 2.6 2.1 1.6 1.6 Other Social Insurance Programs 53.0 53.0 53.0 45.9 44.7 42.6 41.0 40.1 39.8 40.0 40.2 41.0 Total Social Insurance Programs 53.0 53.0 55.0 51.5 51.6 48.2 45.4 43.3 42.5 42.1 41.7 42.5 of which: Rate for Employees 10.0 10.0 10.5 13.3 14.3 15.2 16.1 17.3 17.4 17.4 17.5 17.6 Rate for Employers 43.0 43.0 44.5 37.9 37.0 32.6 29.0 25.7 24.8 24.3 23.9 24.6 Net Personal Income as X of GDP 62.1 67.7 69.9 69.8 69.7 66.6 63.8 61.9 61.4 61.0 60.6 61.1 as % of Net Personal Income: (100) (100) (100) (100) (100) (100) (100) (100) (100) (100) (100) (100) Net Earnings 59.0 59.4 60.0 57.9 56.7 58.1 59.2 60.6 61.5 62.1 62.8 62.6 Social Incore in Cash and Kind 41.0 40.6 40.0 42.1 43.3 41.9 40.8 39.4 38.5 37.9 37.2 37.4 Total Cost of Labor to Enterprises 61.1 63.3 65.7 63.7 62.7 60.6 58.7 57.8 57.7 57.7 57.7 58.1 (in terms of X of GDP) of which all Social Insurance 14.0 14.0 14.6 12.4 12.1 10.7 9.5 8.4 8.1 8.0 7.8 8.0 Long-Term Projections Year 2006 2011 2016 2021 2026 2031 2036 2041 2046 2051 Govt. Pensions as % of GDP 9.0 9.8 9.9 9.5 9.8 10.5 11.2 10.7 10.2 10.0 Govt. Health & Pharm. as X of GDP 5.7 6.6 7.1 7.3 7.8 8.5 9.2 9.2 9.1 9.1 a411.n465 altgrn1.wk1 109 10.13 For family allowances, Chapter 6 discussed a number of options for structural reform which could be applied in combination. Based on this analysis of the range of possible savings which could have been achieved as a share of family allowance expenditure in 1990, it is assumed here that the allowance expenditure projected in the Base Case is cut back 21 percent in 1991, and continues thereafter at the projected trend rate. Note that the recommended taxation of family allowances would have an even higher impact to the Central budget as an increase in revenues, but revenue effects are not projected here. Concerning maternity and child care allowances, the analysis in Chapter 7 also recommended a menu of reform options, two of which (shifting the childcare fee from earnings-related to flat-rate, and eliminating the third year of childcare leave) would have saved about 34 percent of total expenditure on child care benefits in 1989. Again taking these changes to indicate the order of magnitude of feasible reform options, the Base Case estimate for this category of expenditure is reduced by one-third in 1991, then continues at the projected trend growth rate. 10.14 The assumed growth of expenditure on health care and pharmaceuticals was taken to be the same as in the Base Case. As indicated in para. 10.05, even this significant decline in the income elasticity of health expenditures would probably require an unsustainable share of resources to be devoted to health care over the longer term, indicating the need for further curtailment of costs in this sector. 10.15 In addition to shifting most of sick pay out of the social insurance account, the Alternative Scenario assumes that the remaining maternity and child care benefits projected above would be financed by the general budget rather than by social insurance from 1992 on. Thus, the sum of projected expenditures for pensions, sick pay, health and pharmaceuticals, and the administrative costs of social insurance would require an implicit contribution rate of only about 40 percent during the second half of the decade. Adding the effect of unemployment insurance (which is assumed to be the same in the Alternative Scenario as in the Base Case), the total social insurance contribution requirement would still decline from 53 percent at present to 42 percent by 1997.4 (See Figure 10.2) This decrease is assumed to benefit employers, so that their contribution rate shows a steeper decrease, from 43 percent of the contribution base at present to 24-25 percent in the last four years projected. The period of high unemployment during 1992-95 would require an increase in the unemployment insurance contribution during these years or averaging over the decade. It is assumed that employees' social insurance contribution would rise, which would produce a better balance between the contributions of employees and employers. 10.16 Turning to the other programs outside of social insurance, the Alternative Scenario assumes that expenditure for labor market programs in- kind (Employment Fund) remains about 75 percent of income support for the unemployed, as in 1991. This results in total expenditure on employment and 4 Any additional rechanneling of components of pension or health care expenditure from social insurance to general budget financing would reduce the implicit contribution requirement even further. 110 unemployment-related services averaging 1.85 percent of GDP in 1990-2000, which compares reasonably well to the record of many OECD countries. The Alternative Scenario adopts the same projections as the Base Case for all of the remaining items of social expenditure, except for showing a slightly higher growth rate of social services to maintain their 0.5 percent share of GDP. 10.17 The Alternative Scenario results in lowering the ratios of both Social Income and Social Expenditure to GDP over the decade by about 3 additional percentage points compared to the Base Case (Figures l0.la-b). As a share of general government expenditures, total Social Expenditures in both scenarios reach their lowest point around 1996-97 and then rise again in the latter part of the decade, as it is assumed that the contraction of total government expenditure is greater in that period than the contraction of social expenditure (Figure 10.4). The Alternative Scenario's lower share in government expenditure is due entirely to its lower level of social expenditures, since there are no differences between the scenarios in the assumed size of the government sector or in assumed burden-sharing by the private sector (except for sick pay). 10.18 Both the Alternative Scenario and the Base Case assume that households continue to share in education and health expenditures during the medium term to the same degree as in 1989-90. For health, this implies that the considerable copayment which currently takes the form of illegal gratuities will become formalized in the future through official copayments and private sector financing. The current share of private expenditure in total health financing (estimated as 20 percent in 1989) is close to the average for OECD countries and maintaining this rate does not imply large scale privatization of the health sector. The projection assumes that total (government and private) expenditure on health/pharmaceuticals will increase marginally by 2000 as a share of GDP, and on education/scholarships will remain steady. These expenditure shares compare reasonably well to the average in OECD countries and the main focus of reform should be to increase the efficiency - that is, level of real service - achieved by the public expenditure. In the case of secondary and higher education, larger enrollments should be achieved with the given level of public expenditure, implying greater efficiency in the use of educational capacity. Any larger expansion in health or education services over the present decade will only be possible if additional private financing is forthcoming. 10.19 The projection of private financial savings developed for the medium-term macroeconomic framework (the basis of the structural adjustment program for the Hungarian economy) is taken here as a proxy for contributions to private pension plans, which should increase to provide a gradually larger share of total retirement income. It is also assumed that households will on average spend up to 20 percent of their Gross Earnings (about 25 percent of Net Earnings) on housing and utilities expenditure as government subsidies on housing are scaled back. These expenditure shares are based on norms in other middle-income market economies and would be achieved as both new mortgage instruments and a rental allowance system are introduced which require such a financial contribution by households. (See recommendations in World Bank, 1990c.) These assumptions regarding burden-sharing of pension and ll housing/utilities costs permit the total expenditure for these items to be maintained as a share of GDP in 1991-2000 even as the government's contribution is reduced. 10.20 Both the Alternative Scenario and the Base case make the same assumptions regarding the growth of household earned income, and both assume no change in personal income taxation (projected as a constant 12 percent of Gross Earnings). As already noted, the Alternative Scenario involves a higher social insurance contribution from employees, however. Despite this greater withdrawal from households, net earned income becomes a higher share of total Net Personal Income over the decade. The lower growth of social programs in the Alternative Scenario permits Social Income to contribute a declining share of Net Personal Income--although the decline (from about 41-42 percent in the first half of the decade to about 37-78 percent in the second half) is modest and delayed (see Figures 10.3a-b). As discussed in Chapter 2, it should be an objective of reform to reduce the social share of household income (and increase the earned share) more dramatically in order to improve work incentives. 10.21 The total cost of labor to enterprises is also projected, which for the Alternative Scenario includes the share of sick pay assumed to be transferred to employers by the SIF. In this Scenario, the total cost of labor to enterprises does decline, from 63.3 percent in 1990 to about 58 percent in the latter half of the decade. This decline is effected by the cut in employers' social insurance obligation. 10.22 In conclusion, the Alternative Scenario illustrates that a reasonable set of assumptions about reforms in the main social programs would permit, by the end of the decade, some improvement in the main structural issues listed at the beginning of this Chapter. The burden on employers to finance social insurance would be reduced quite strikingly, while the largest social insurance program, pensions, would be made financially viable over the longer term. The scenario involves structural changes in the eligibility and/or benefits of all the individual programs, although it should be noted that more than half of the projected cut in the implicit social insurance contribution requirement results from the transfer of expenditure out of this type of financing (sick pay shifted to employers and maternity/child care allowances to the general budget). 10.23 The share of GDP devoted to social expenditures and the share of social income in household income would both be reduced by a few percentage points relative to the Base Case (the "less reform" scenario). A more dramatic reduction in these shares might be necessary. The Alternative Scenario assumes that the Employment Fund and Social Assistance both increase in line with the trend of income support for unemployment, but the financial implications might be unsustainable and one or both of these programs may have to be scaled back during the peak adjustment period. There is also scope for additional cutbacks in family benefits as discussed earlier. It is important to reiterate that the projection does not include the financial implications of all of the reforms recommended in Chapters 4-8. Some of these reforms, such as taxation of family allowances, would generate additional budgetary revenues. By changing incentives, certain of the recommended reforms would 112 also reduce the demand for social expenditures (e.g., a lower level of child care benefit would reduce its use). These positive budgetary effects cannot be captured in the simple projection model used here. It is also true that the net budgetary impact of recommendations to tax pensions and exempt pension contributions from the PIT, or raise nursery fees while introducing tax incentives for private nurseries, is not included in the projections, although the likely magnitude of these effects would be relatively small compared to the reforms which have been included here. 113 Figure 101sa Social Expenditure. Social Income and Net Personal Income as s of GODP Praoentage. Bse Case s0 78- 70 . t 65 Net Personal Indgarp 60 656 60 45- 40- 36 - Ttat Sooial Expenditure 1990 1996 2000 Figure 10.1b: Social Exponditure. Social Income and Not Personal Income * s of GDP Alternative Scenario Percentago 76 70 . ...... . Net Personal Income 66 60 56 5o 46 40- 35 -T,ta Social Expenditure 20 1 2000 1990 1996 2000 114 Figure 10.3a: Coat of Labor as % of GOP and Soobal Incomo as % of NPI P.reentage BASE CASE 70 60 Total Cost of Labor to Employer to % ot GOP r;n _ _5 50 40 - Social Income In Cash a*nd Kind as S of Noat Pereenikl nooen 36 - 30 1990 1995 2000 Figure 1O.3b; Coat of Labor as % of GDP and Social Income as % of NPI Peroentaea Alternative Scenario B6 60 66 Total Cost of Labor to Employer as S of GOP 60 46 40 35 Socol lncoms in Carh S*d Krind aas e o tel Personal Inemos 30 1 2000 1990 1995 2000 115 Flure 10o.2; imtplict Contribution Rates io; Social Inuurance 70 85 55 = *0 _ 45D Alternatlt Sconarlo \ 40 , 1990 1996 2000 Figure 10.4: Total Soclal Expenditure as % of General cowrnment Exponditures 70 60 66 AIternative Scenario 1990 1995 2000 116 ANNEX I Page 1 of 2 Bibliography Adam, J (1984) "Regulation of Labour Supply in Poland, Czechoslovakia and Hungary", Soviet Studies, vol. 36, no. 1, pp.69-86. Atkinson, A B, Corlyon, J, Maynard, A, and Trinder C (1981) "Poverty in York: A Re-Analysis of Rowntree's 1950 Survey", Bulletin of Economic Research, vol. 33, pp. 59-71. Atkinson, A B and Hills, J (1989) "Social Security in Developed Countries: Are There Lessons for Developing Countries?" Welfare State Programme paper 38, STICERD, London School of Economics. Atkinson, A B and Micklewright, J, (1990) "Unemployment Compensation and Labour Market Transitions: A Critical Review", ESRC TIDI Programme Discussion Paper 143, LSE, and EUI Working Paper in Economics 90/9, EUI, Florence (forthcoming, Journal of Economic Literature). Barr, Nicholas (1989), Student Loans: The Next Steps, Aberdeen University Press. Beckerman, W, and Clark, (1982), Poverty and Social Security in Britain Since 1962, OUP, Oxford. Brinkmann, C (1988) "The Income of the Unemployed in the Federal Republic of Germany", paper presented at Expert Meeting on Unemployment Compensation, OECD, January 1988. Brown, J C, Child Benefit Options for the 1990s Save Child Benefit Campaign, London. Central Statistical Office (1984), The Impact of Policy Measures other than Family Planning Programmes on Fertility, Demographic Research Institute, Research Report 18, Budapest. Central Statistical Office, Ministry of Finance, and World Bank, "Subsidy-Income Study: The Effect of Budgetary Consumer and Housing Subsidies on the Income Distribution of the Households in Hungary", October 1989. (referenced in text as 1989 Incidence Study) Davey, K, "Local Government Reform in Hungary," Institute of Local Government Studies, University of Birmingham, November 1990 (mimeo). Esping-Andersen, G and Micklewright, J (1990) "Welfare State Models in OECD Countries: An Analysis for the Debate in Eastern Europe," UNICEF, Roundtable on Safety Nets for Children in the Central and Eastern European Countries in Transition to the Market Economy. Gordon, M (1988), Social Security Policies in Industrialised Countries: A Comparative Analysis, CUP, Cambridge. 117 Page 2 of 2 HM Treasury (1990), The Government's Expenditure Plans 1990-91 to 1992-93, Cm 1014, HMSO, London. Holzmann, R (1988), "On the Relationship Between Retirement and Labor Market Policies," in Structural Problems of Social Security Today and Tomorrow (Leuven: ACCO). ILO (1989), From Pyramid to Pillar: Population Change and Social Security in Europe, ILO, Geneva. IMF (1990), Social Security Reform in Hungary, IMF Fiscal Affairs Department. Matoricz, A (1990) "Munkaero-piaci Helyzetkep", Tarsadalmi Szemle no. 90/5, 48- 51. Micklewright, J (1990) "The Reform of Unemployment Compensation: Choices for East and West", EUI Working Paper in Economics, Florence (forthcoming, European Economic Review). Ministry of Finance (1987) Act on the Personal Income Tax, Public Finance in Hungary paper 39/B. Ministry of Finance, Central Statistical Office, and World Bank (1990) "The Hungarian Social-Policy Systems and Distribution of Incomes of Households." (referenced in text as 1990 Incidence Study) Moylan, S, Millar, J, and Davies, R (1984), For Richer. For Poorer? DHSS Cohort Study of the Unemployed, HMSO, London. OECD (1985), Social Expenditure 1960-1990: Problems of Growth and Control, Paris. ---- (1985, 1988, 1990), Employment Outlook, Paris. ----- (1988), The Future of Social Protection, Paris. Scherer, P (1990) "A Review of National Labour Market Policies in OECD Countries", Paper presented at World Bank Seminar on Economic Adjustment, 27- 28 February 1990, mimeo. Szalai, J (1989) "Poverty in Hungary during the Period of Economic Crisis", Background Paper for the 1990 IBRD World Development Report. Urban Institute (1990), "Integrating State Rental Housing with the Private Market: Designing Housing Allowances for Hungary," Washington, D.C. Varley, R (1986), The Government Household Transfer Data Base. 1960-84, OECD, Paris. World Bank (1990a), Hungary, the Transition to a Market Economy: Critical Human Resources Issues, Report No. 8665-HU. ---------- (1990b), Hungary Health Services: Issues and Options for Reform. - --------- (1990c), Housing Policy Reform in Hungary, Report No. 9031-HU. 118 ANNEX II Page 1 of 33 The Economic Context of Social Policy and Expenditures A. International Comparison of Social Expenditures 1. Size of Government. Current government expenditures in Hungary have averaged 54 percent of GDP throughout the 1980s, slightly above the average of 50 percent for the Social Welfare States in OECD. In terms of total tax revenues as a share of GDP, Hungary (with 54 percent in 1988) is matched only by Sweden and Denmark during the 1980s and is about twenty percentage points higher than the average for the Lower Income states. Government in Hungary is even somewhat larger than the average for other (formerly) socialist countries of East/Central Europe.1 2. Size of Social Expenditures. To evaluate differences among countries in size and structure of social expenditures, it is necessary to consider their demographic profiles as well (Table II.1). Given the relative size of its 0-14 year age group within the total population--a demographic feature closer to that of the Social Welfare States than to the Lower Income States--Hungary's share of education expenditure should be higher than the actual figure (and that of the Lower Income States should be even greater still). Hungary falls midway between the two comparator groups in the ratio of elderly to the total population. This ratio, which represents the main demographic source of demand for health care, would also suggest a much higher health care expenditure ratio for Hungary. Hungary's relatively high share of elderly residents helps to explain its higher level of pension expenditure relative to the Lower Income States (although policy, and not demographics, accounts for the major part of the pension growth, as explained later). 3. The "Government Household Transfers" data base of the OECD permits a more detailed intercountry comparison of the composition of social security and welfare transfers (Table II.2). Unlike the OECD "social expenditures," these data include social services in addition to health and education, as well as housing assistance transfers. In the most recent period for which comparable data exist (1980-84), Hungary's expenditure on temporary sickness and maternity benefits, relative to GDP, is shown to approach the rates for the Social Welfare States, a development partially explained by the absence of unemployment benefits during this period. Family benefits in Hungary far exceeded those in either of the two comparator groups, while welfare payments were about the same share of GDP as in the Lower Income States. 1 Poland's current government expenditure ratio matched or exceeded Hungary's in the early 1980s, but fell precipitously thereafter; the Polish tax revenue/GDP ratio has remained well below that of Hungary throughout the 1980s. Bulgaria's current expenditure ratio peaked at 53 percent, and its tax revenue ratio peaked at 50-51 percent, of GDP in the 1980s (reference to consolidated State budget). In Czechoslovakia, both tax revenue and current expenditure of the general government budget reached a maximum of 51% in the late 1980s. (Source: World Bank, most recent country economic reports) 119 Table II.1: Age Structure of OECD Population and Hungary a/ (As X of Total Population) 1980 1990 1980 1990 Age Groups Age Groups AUSTRALIA 0-14 25 22 JAPAN 0-14 24 18 65 + 10 11 65 + 9 11 Total 35 33 Total 33 30 AUSTRIA 0-14 20 18 LUXEMBOURG 0-14 18 18 65+ 16 14 65 + 13 15 Total 36 33 Total 30 33 BELGIUM 0-14 20 19 NETHERLANDS 0-14 22 18 65 + 14 14 65 + 11 13 Total 34 33 Total 34 31 CANADA 0-14 23 21 NEW ZEALAND 0-14 27 22 65 + 10 11 65 + 10 11 Total 32 32 Total 36 33 DENMARK 0-14 21 17 NORWAY 0-14 22 19 65 + 15 15 65 + 15 16 Total 35 32 Total 37 35 FINLAND 0-14 20 19 PORTUGAL 0-14 27 22 65 + 12 13 65 + 10 12 Total 32 33 Total 37 34 FRANCE 0-14 22 20 SPAIN 0-14 26 22 65 + 14 14 65 + 11 13 Total 36 34 Total 37 34 GERMANY 0-14 18 15 SWEDEN 0-14 20 17 65 + 16 15 65 + 16 18 Total 34 31 Total 36 35 GREECE 0-14 23 20 SWITZERLAND 0-14 20 17 65 + 13 12 65 + 14 15 Total 36 32 Total 34 32 ICELAND 0-14 30 20 TURKEY 0-14 39 36 65 + 10 10 65 + 5 4 Total 40 30 Total 44 40 IRELAND 0-14 31 28 UNITED KINGDOM 0-14 21 19 65 + 11 11 65 + 15 15 Total 41 39 Total 36 34 ITALY 0-14 22 18 UNITED STATES 0-14 23 22 65 + 13 14 65 + 11 12 Total 35 32 Total 34 34 HUNGARY (1986)0-14 22 65 + 12 Total 34 --------------------------------------------------------------------__-------__---------- OECD Lower Income States - Averages OECD Social Welfare States - Averages Greece, Ireland, Portugal, Spain Belgium, Denmark, Finland, France, and Turkey Netherlands, Norway and Sweden 0-14 29 26 0-14 21 18 65+ 10 10 65+ 14 15 -----------------------------------------------------------------------__----__---------- Average Total 39 36 Average Total 35 33 -----------------------------------------------------------------------__----__---------- a/ 1980 actual numbers: 1990 projected numbers. Source: OECD Demographic Data File, medium fertility variant projections. 120 Table II.2: Government Social Security and Welfare Transfers to Households (as percentage of GDP) ---------------------------------------------------------------------__------__----------------------------------------------------- 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1989 -------------------------------------------------------------------__--------__----------------------------------------------------- OECD Lower Income (unweighted average) SociaL Security & Welfare 9.1 10.4 11.2 11.6 NA Social Security 8.7 10.0 10.8 11.2 NA Temp. Sickness & Maternity 0.8 0.8 0.8 0.8 NA Pensions (Old age, disability) 5.9 6.7 7.3 7.6 NA Unemployment 1.2 1.5 1.6 1.6 NA Family Assistance 0.5 0.6 0.6 0.6 NA Other (incl. housing assist.) 0.3 0.4 0.4 0.5 NA Welfare 0.4 0.4 0.4 0.4 NA OECD SociaL Welfare (unweighted average) Social Security & Welfare 15.2 15.8 16.4 16.6 16.2 Social Security 14.5 15.1 15.6 15.9 15.4 Temp. Sickness & Maternity 1.7 1.6 1.6 1.5 1.5 Pensions (Old age, disability) 9.4 9.7 9.8 9.9 9.7 Unemployment 1.2 1.6 1.9 2.1 2.0 Family Assistance 1.7 1.7 1.7 1.7 1.6 Other (incl. housing assist.) 0.5 0.5 0.6 0.6 0.6 Welfare 0.7 0.7 0.7 0.8 0.7 Hungary Social Security & Welfare 13.3 12.3 12.3 12.6 12.9 14.2 14.7 14.2 16.1 16.0 17.1 Social Security 13.0 12.0 11.9 12.2 12.5 13.7 14.2 13.8 15.5 15.5 16.2 Temp. Sickness & Maternity 1.4 1.4 1.4 1.4 1.3 1.3 1.4 1.3 1.4 1.5 1.4 Pensions (Old age, disability) 7.8 7.8 8.1 8.4 8.6 8.9 9.1 9.0 9.2 9.1 9.9 Unemployment 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.1 0.1 0.5 Family Assistance 2.4 2.3 2.2 2.2 2.3 2.5 2.5 2.4 3.3 3.7 3.8 Other (incl. housing assist.)* 1.3 0.4 0.3 0.3 0.2 1.1 1.2 1.1 1.5 1.1 0.6 Welfare 0.3 0.3 0.3 0.4 0.4 0.5 0.5 0.5 0.8 0.5 0.9 ---------------------------------------------------------------------__------__----------------------------------------------------- For purposes of comparison, Greece, Ireland, Portugal, Spain and Turkey are classified OECD lower income countries. The OECD OECD social welfare states are Belgium, Denmark, Finland. France, the Netherlands, Norway and Sweden. In this table, data are not available for Belgium or Turkey. * Includes maintenance subsidy to the rental sector and social policy grants for housing construction. Sources: OECD - The Government Household Transfer Data Base, 1960-84 by Rita Varley. Hungary - Table 2.2. SSWELGOP.WK1 a520.1563 Table II.3a - Hungary: Structure of Social Expenditures (as X of Household Income and GDP) 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 As Percentage of Household Income Pensions 8.8 9.8 10.1 10.3 11.4 12.1 12.2 12.7 12.9 13.3 13.3 13.4 13.6 14.1 14.7 14.6 Sickness benefit 2.2 2.1 1.9 1.9 1.9 1.8 1.8 1.8 1.8 1.7 1.7 1.7 1.6 1.8 2.0 1.8 Family allowance 2.1 2.3 2.4 2.2 2.5 2.9 2.9 2.8 2.8 3.1 3.1 2.9 2.9 4.0 5.0 4.5 Pregnancy/maternity benefits 0.6 0.6 0.5 0.5 0.5 0.4 0.4 0.3 0.3 0.3 0.3 0.4 0.3 0.4 0.4 0.3 Childcare allowance 1.0 1.1 1.1 0.9 0.9 0.8 0.7 0.6 0.6 0.5 0.4 0.3 0.2 0.2 0.3 0.2 Childcar- fee 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.2 0.5 0.5 0.8 0.8 0.8 Scholarships 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 Other cash social income 1/ 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.7 0.7 0.7 0.7 0.7 0.9 0.8 1.3 o/w: regular social support 0.1 0.1 0.1 0.1 0.1 0.1 0.1 irregular social support 0.0 0.1 0.1 0.1 0.1 0.2 0.1 Employment Fund benefits in cash 6/ 0.0 0.1 0.5 Social income - cash 15.4 16.5 16.6 16.6 17.8 18.8 18.7 19.0 19.2 19.8 19.9 20.0 20.2 22.4 24.1 24.1 Total Social Income in kind 11.7 11.9 12.0 12.7 12.7 14.3 14.9 15.4 15.3 15.4 15.8 15.8 15.9 16.5 16.9 16.4 of which: Health 2/ 3.5 3.5 3.5 3.7 3.8 5.1 5.5 5.7 5.6 5.7 5.9 6.1 6.2 6.2 6.7 6.7 of which: Pharmaceutical Subsidies 1.2 1.3 1.3 1.4 1.5 1.7 1.9 2.0 2.0 1.8 1.9 Social services nurseries 0.7 0.7 0.7 0.7 0.8 0.8 0.7 0.8 0.8 0.8 0.9 0.9 0.9 1.0 0.7 0.7 Education, kindergarten 5.0 4.9 4.9 5.4 5.2 5.5 5.6 5.9 5.7 5.7 5.8 5.7 5.7 6.1 7.0 6.8 Culture, sports 3/ 1.3 1.4 1.6 1.6 1.7 1.7 1.9 1.9 2.1 2.1 2.2 2.2 2.2 2.2 1.8 1.4 Employment Fund benefits in kind 6/ 0.1 0.1 0.3 Total Social Income 27.2 28.4 28.6 29.3 30.6 33.2 33.6 34.4 34.6 35.2 35.8 35.8 36.1 38.9 41.0 40.6 Consaer Subsidies 4/ 12.7 13.5 12.9 12.5 10.3 9.8 10.3 10.5 6.4 5.5 2.9 Bousing Assistance 5/ 1.3 1.0 1.2 1.2 1.2 1.3 1.9 2.0 3.1 6.1 4.9 Total Social Expenditures 47.1 48.1 48.6 48.3 46.8 46.9 48.0 48.6 48.4 52.5 48.5 As Percentage of GDP Pensions 5.6 6.1 6.2 6.4 7.1 7.8 7.8 8.1 8.4 8.6 8.9 9.1 9.0 9.2 9.1 9.9 Sickness benefit 1.4 1.3 1.2 1.2 1.2 1.2 1.2 1.1 1.2 1.1 1.1 1.2 1.1 1.2 1.2 1.2 Family allowance 1.3 1.4 1.5 1.4 1.5 1.9 1.9 1.8 1.8 2.0 2.1 2.0 1.9 2.6 3.1 3.1 Pregnancy/maternity benefits 0.4 0.3 0.3 0.3 0.3 0.3 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.3 0.2 0.2 Childcare allowance 0.6 0.7 0.6 0.6 0.5 0.5 0.5 0.4 0.4 0.3 0.3 0.2 0.2 0.1 0.2 0.2 Childcare fee 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.2 0.3 0.3 0.5 0.5 0.5 Scholarships 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.2 0.1 0.1 0.1 0.1 0.1 Other cash social income 1/ 0.3 0.3 0.3 0.3 0.3 0.3 0.3 0.3 0.4 0.4 0.5 0.5 0.5 0.6 0.5 0.9 o/w: regular social support 0.1 0.1 0.1 0.1 0.1 0.1 0.1 irregular social support 0.0 0.0 0.1 0.1 0.1 0.1 0.1 Employment Fund benefits in cash 6/ 0.0 0.1 0.3 Social income - cash 9.8 10.2 10.3 10.2 11.0 12.1 12.0 12.1 12.5 12.8 13.3 13.6 13.3 14.6 15.0 16.4 Total Social Income in kind 7.5 7.4 7.4 7.9 7.9 9.2 9.6 9.8 10.0 10.0 10.5 10.8 10.5 10.7 10.5 11.2 of which: Health 2/ 2.2 2.2 2.2 2.3 2.3 3.3 3.5 3.6 3.6 3.7 3.9 4.1 4.1 4.1 4.1 4.6 of which: Pharmaceutical Subsidies 0.7 0.8 0.8 0.9 1.0 1.2 1.3 1.3 1.3 1.1 1.3 Social services, nurseries 0.4 0.4 0.4 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.6 0.6 0.6 0.6 0.4 0.5 Education, kindergarten 3.2 3.1 3.0 3.4 3.2 3.5 3.6 3.7 3,7 3.7 3.9 3.9 3.7 4.0 4.4 4.6 Culture, sports 3/ 0.8 0.9 1.0 1.0 1.0 1.1 1.2 1.2 1.4 1.4 1.5 1.5 1.4 1.5 1.1 1.0 Employment Fund banefits in kind 6/ 0.1 0.1 0.2 Total Social Income 17.3 17.6 17.7 18.1 18.9 21.2 21.6 21.9 22.4 22.8 23.8 24.4 23.8 25.4 25.4 27.6 Consumer Subsidies 4/ 8.1 8.7 8.2 8.1 6.7 6.5 7.0 6.9 4.2 3.4 2.0 Housing Assistance 5/ 0.8 0.7 0.8 0.8 0.8 0.9 1.3 1.3 2.0 3.8 3.4 Total Social Expenditures 30.2 30.9 30.9 31.3 30.3 31.2 32.7 32.0 31.6 32.6 33.0 Same notes on Table 2.2. 05-Feb-91 16:48:12 NEWEXP.WK1 (Percentage Distribution: Household Income & GDP) Table II.3b - Hungary: Annual Growth of Social Expenditures 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 Pensions 12.4 9.1 6.5 13.0 9.9 4.0 5.8 4.3 5.5 2.9 4.4 2.3 2.8 -1.4 1.8 Sickness benefit -8.0 -0.0 5.9 -0.2 -0.4 1.8 0.8 5.1 -3.3 -0.4 6.4 -4.0 9.2 5.0 -10.0 Family allowance 10.5 9.7 -0.1 12.8 22.1 2.1 -3.5 2.3 14.0 3.7 -2.7 -0.3 38.7 16.9 -5.9 Pregnancy/maternity benefits 2.8 -2.0 -3.0 -4.7 -5.7 -4.9 -10.4 -4.7 -0.7 -0.6 25.3 -0.5 14.9 -9.2 -8.0 Childcare allowance 9.4 5.4 -5.9 -4.8 -1.1 -9.7 -10.5 -7.4 -6.0 -22.3 -31.1 -7.6 -17.4 16.4 -6.8 Childcare fee 104.8 11.4 50.6 -4.2 1.3 Scholarships -2.3 6.1 -0.6 4.4 4.4 -4.9 13.5 -12.7 19.7 7.9 -3.6 -1.9 -23.3 -1.7 13.6 Other cash social income 1/ -9.5 6.7 2.5 2.5 1.6 8.7 10.3 29.3 8.9 0.8 0.5 7.5 23.6 -13.0 62.7 Employment Fund benefit in cash 6/ 173.1 403.1 Social income - cash 7.7 7.4 4.2 9.5 9.3 2.7 3.1 4.0 5.7 3.2 4.0 1.5 10.2 2.1 2.6 Total Social Income in kind 2.0 7.2 11.3 1.8 16.3 7.4 5.2 2.4 2.9 5.4 3.6 1.4 2.5 -2.7 -0.0 of which: Health 2/ 1.4 5.4 9.7 4.8 39.7 10.5 6.0 1.2 3.6 6.8 7.1 2.4 -0.5 1.7 3.4 of which: Pharmaceutical Subsidies 12.7 6.4 11.2 8.6 16.8 10.9 5.8 0.2 -13.1 7.9 Social services, nurseries 8.6 9.6 7.8 6.6 11.9 -12.2 12.9 4.1 6.1 5.1 4.6 3.9 7.5 -30.0 -3.8 Education, kindergarten -0.0 6.0 15.2 -1.9 8.4 6.4 5.6 0.4 2.8 4.0 1.0 0.3 6.4 10.3 -1.6 Culture, sports 3/ 11.4 18.1 8.8 5.6 6.7 14.2 2.4 11.8 2.5 6.5 4.1 -0.2 2.4 -24.0 -18.1 Employment Fund benefit in kind 6/ 43.4 102.1 Total Social Income 5.3 7.3 7.2 6.2 12.2 4.7 4.0 3.3 4.4 4.2 3.8 1.5 6.8 0.1 1.6 Consumer Subsidies 4/ 9.7 -2.5 -0.8 -15.1 -2.7 8.8 2.3 -39.1 -18.9 -45.0 Housing Assistance 5/ -17.8 24.2 1.0 0.7 13.0 49.4 7.4 51.6 85.4 -16.4 Total Social Expenditures 5.4 2.6 2.2 -0.7 2.9 6.1 1.9 -1.2 3.0 -5.4 s Same notes as Table 2.2. Values deflated by GDP deflator. 08-Jan-91 13:23:12 NEWEXP.WK1 (Real Growth Rates) 123 Table II.3c - Hungary: Structure of Social Expenditures (Percentage Distribution) ---------------------------------------------------__------------------------__-------------------------------------------------- 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 -------------------------------------------------------------------__--------__-------------------------------------------------- As Percentage of Total Social Expenditures Pensions 25.7 25.4 26.2 26.7 28.4 28.4 27.9 28.0 29.2 27.9 30.0 Sickness benefit 3.9 3.8 3.7 3.8 3.7 3.6 3.6 3.4 3.7 3.8 3.6 Family allowance 6.2 6.1 5.7 5.7 8.5 6.6 6.0 5.9 8.3 9.4 9.4 Pregnancy/maternity benefits 0.9 0.8 0.7 0.6 0.6 0.6 0.7 0.7 0.8 0.7 0.7 Childcare allowance 1.8 1.5 1.3 1.2 1.1 0.9 0.6 0.5 0.4 0.5 0.5 Childcare fee 0.0 0.0 0.0 0.0 0.0 0.5 1.0 1.0 1.6 1.5 1.6 Scholarships 0.5 0.4 0.5 0.4 0.5 0.5 0.4 0.4 0.3 0.3 0.4 Other cash social income 1/ 1.0 1.0 1.1 1.4 1.5 1.5 1.4 1.5 1.8 1.5 2.6 Employment Fund benefit in cash 6/ 0.1 0.2 0.9 Social income - cash 40.0 38.9 39.1 39.8 42.4 42.5 41.6 41.5 46.3 45.9 49.8 Total Social Income in kind 30.4 30.9 31.7 31.8 32.9 33.7 32.9 32.8 34.0 32.1 33.9 of which: Health 2/ 10.8 11.4 11.7 11.6 12.1 12.6 12.7 12.7 12.8 12.7 13.9 of which: Pharmaceutical 2.5 2.7 2.7 3.0 3.3 3.7 3.9 4.0 4.1 3.4 3.9 Social services, nurseries 1.8 1.5 1.6 1.7 1.8 1.8 1.8 1.8 2.0 1.4 1.4 Education, kindergarten 11.6 11.7 12.1 11.9 12.3 12.4 11.8 11.6 12.5 13.4 13.9 Culture, sports 3/ 3.7 4.0 4.0 4.3 4.5 4.6 4.6 4.5 4.6 3.4 3.0 Employment Fund benefit in kind 6/ 0.2 0.2 0.5 Total Social Income 70.3 69.8 70.8 71.6 75.3 76.3 74.6 74.3 80.3 78.0 83.7 Consumer Subsidies 4/ 27.0 28.1 26.7 25.9 22.1 20.9 21.5 21.5 13.3 10.5 6.1 Housing Assistance 5/ 2.7 2.1 2.6 2.5 2.6 2.8 4.0 4.2 6.4 11.5 10.2 Total Social Expenditures 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 ---------------------------------------------------------------------__------__-------------------------------------------------- Same notes as Table 2.2. 08-Jan-91 13:23:12 I4EWEXP.WK1 (Percentage Distribution: Total Social Expenditure) 124 ANNEX II Page 7 of 33 4. It should also be noted that comparisons of the relative shares of social expenditure across countries tell little about the actual quality and volume of services pro',ided. In part this is because the expenditure .data do not reflect differences in efficiency; the later chapters of this Report examine the problems of low efficiency in the Hungarian social programs which reveal that their considerable share of resources produce a lower level of services (and basic weliare) than would be otherwise expected. The comparative data on social spending do not include private expenditure, nongovernment communal expenditure (e.g., by voluntary organizations, trade unions, etc), nor fiscal support to households in the form of tax expenditures. These omissions are most significant as a source of downward bias in evaluating the overall service provision in health, education, other social services, and pensions; they would also affect mainly the assessment of the United States and some of the countries with hybrid social systems. These omissions would not change significantly the picture of the Social Welfare States, however, and so do not affect our comparison of these with Hungary. In the case of the one social program for which fairly complete expenditure data are available, health care, the inclusion of private expenditures does not change Hungary's ranking in relation to the other major country groups (Table II.4 and Fij%_re II.1).2 5. Growth of Social Income Transfers. The income elasticity of Hungary's total social expenditure has averagec well above 2 during the past decade, considerably exceeding that in most OECD countries in the early 1980s as well as in the two previous decades (Table II.5). 6. To determine the sources of this growth for the three largest social programs (education, health, and pensions), the growth of expenditure was disaggregated into that portion explained by the increase in the relevant population (demographic factor), the number of beneficiaries as a share of the eligible age group (coverage factor), and the real value of benefits per user. As indicated in Table II.6, for the 1981-89 period the demographic factor is negative for education and health, due to the actual decline in the school- age and total populations, respectively.3 Most of the increase in total education and health expenditure is attributable to growth in the average real 2 According to Esping-Andersen and Micklewright (1990), private pensions accounted for 6% of total pension expenditure in Sweden, 10% in Germany, and 20% in the U.S. in 1980. Tax expenditures for private pensions in the U.S. were valued at nearly 20% of total public pension expenditure, and tax expenditures on health also amounted to about 20% of total public health spending, in 1981. 3 It may be more appropriate to consider the elderly and youth (0-14 year old) populations together as the more relevant demographic group affecting health demand; however, this group also declined in size over the period, by 0.14% p.a.. Counting only the increase in the retirement-age population (0.58% p.a.) as the demographic factor, the implied growth of the average real health benefit would fall to 1.22 percent p.a.. 125 Table 11.4 - Comparative Health Expenditure, 1986 Total Public Public Health Expenditure Health Expenditure Health Expenditure as % of GDP as % of GDP as % of Total Australia 6.9 5.3 76.8 Austria 8.3 5.3 63.9 Belgium 7.1 5.5 77.5 Canada 8.7 6.5 74.7 Denmark 6.0 5.2 86.7 Finland 7.5 5.8 77.3 France 8.5 6.7 78.8 Germany 8.1 6.3 77.8 Greece 4.0 3.7 92.5 Iceland 7.5 6.9 92.0 Ireland 8.0 7.0 87.5 Italy 6.7 5.2 77.6 Japan 6.7 4.8 71.6 Luxembourg 6.9 6.2 89.9 Netherlands 8.3 6.6 79.5 New Zealand 7.0 6.1 87.1 Norway 6.8 6.6 97.1 Portugal 5.6 4.0 71.4 Spain 6.0 4.3 71.7 Sweden 9.1 8.3 91.2 Switzerland 7.7 5.2 67.5 Turkey 5.2 NA United Kingdom 6.1 5.3 86.9 United States 10.9 4.5 41.3 Hungary 5.1 4.1 80.4 OECD Lower 5.8 4.8 82.8 Income States OECD Social 7.6 6.4 84.2 Welfare States Sources: OECD: Supplement to the OECD Observer and OECD Social Data Bank. Hungary: National health expenditure - IMF: Social Security Reform in Hungary, September, 1990. Government health expenditure - CSO Statistical Yearbook, Social Security Administration, other government officials, and Bank Staff estimates. Notes: 1. OECD Lower Income: Greece, Ireland, Portugal, Spain and Turkey. 2. OECD Social Welfare: Belgium, Denmark, Finland, France, Netherlands, Norway and Sweden. 8 January 1991 126 Figure 11.1 National Health Expenditure as Percentage of GDP Percentage of GDP . . . . . , . . . . .''' I l~~~~~~~~~~~~~ . . . J. . . . .-.- ................ l l iE.. E . ...i. .EE. .E . iEi )i. , . .E::i. .L.E . ... l l i ~~~~~~~~. . . . .. i. . 7;ii,Sl. 0 E . . . .......i L Hungary E EOED Lower Income _EOED Social Welfare Source: Annex Table 11.4. 127 Table II.5 - Elasticities with respect to GDP ----------------------------------------------------------------------__-----__-------------- Total Health Total Social Expenditures Pensions Expenditures 1980-84 1979-84/85 1980-85 Australia 1.0 1.0 1.3 Austria 0.5 1.2 1.8 Belgium 0.9 1.4 1.6 Canada 1.1 1.5 1.5 Denmark -0.3 1.0 0.8 Finland 0.5 1.1 0.4 France 0.4 1.2 2.1 Germany 0.1 0.9 0.6 Greece 3.4 1.7 5.9 Ireland -1.5 1.5 1.4 Italy 1.1 1.3 1.4 Japan 1.1 1.8, 0.8 Netherlands -1.3 0.9 2.2 New Zealand 1.1 0.4 Norway 0.5 0.9 1.1 Portugal 1.3 Spain -0.4 1.4 0.7 Sweden 0.0 1.2 0.1 Switzerland 1.0 0.9 United Kingdom 0.4 1.2 1.5 United States 0.9 1.2 1.2 OECD Lower Income 0.5 1.4 2.6 OECD Social Welfare 0.1 1.1 1.2 Hungary 1979-85 3.8 1980-84 2.3 1980-85 3.1 1981-89 2.5 2.4 2.9 ---------------------------------------------------------------------__------__-------------- Sources: Health expenditures elasticity from OECD "Financing and Delivering Health Care", 1987. Pensions elasticity from OECD "Reforming Public Pensions", 1988. Total social expenditures growth from OECD "Future of Social Protection," 1988. Real GDP Growth from IMF International Financial Statistics. Data for Hungary derived from basic social expenditure table (Table 2.2). OECDELAS.WK1 128 Table II.6 Components of Growth of Social Expenditure (1981 - 1989) (percent per annum) Demography Coverage Average Total Elasticities Expenditure Real Value Growth wrt GDP Item of Benefit Education -0.46 0.21 3.90 3.65 2.62 Pensions 0.58 1.31 1.14 3.30 2.37 Health -0.15 1.68 1.95 3.48 2.50 Total 3.49 2.50 Memorandum Items: GDP Growth Rate (1981-89) 1.39 Income Elasticity (1981-89) 2.50 For 1981-89: Education: Demographic Factor - Increase in age group 0-24 Coverage Factor - Rate of change in school enrollment ratio Pensions: Demographic Factor - Increase in the population of: Men - Age 60+ Women - Age 55+ Coverage Factor - Rate of change in share of retirement age group drawing pensions Health: Demographic Factor - Increase in total population Coverage Factor - Rate of change in actual cases of medical treatment as a share of total population Notes: For Education, Health - "Average real value of benefit" is derived as a residual from total growth. For Pensions - The real value of benefit is observed (value of average pension); the three factors do not add to total due to an unspecified error term. m:\hun\socpol\popn3.doc 129 ANNEX II Page 12 of 33 benefits per capita.' In the case of pensions, although the retirement age group grew over the period, the increase in coverage (pensioners as a share of the retirement age group) accounted for a greater share of total expenditure growth than either the growth of the age group or of real benefits.5 7. Financing of Social Income Transfers. Up to 1987, the social insurance fund (SIF) in Hungary financed old-age, disability, and survivors' pensions, sick pay, family allowances, maternity and childcare allowances, and some other minor benefits (e.g., funeral grants). In 1988, pharmaceutical subsidies were moved out of the central government budget into the SIF, and in March 1990 health care services were added to the SIF in exchange for family allowances, which were moved into the central government budget. Social insurance contributions therefore finance items accounting for 60 percent of social income, with the remainder covered by general tax revenues. 8. As a result of the recent rate increases and despite Hungary's relatively high tax-to-GDP ratio, total social insurance contributions in Hungary amounted to a greater share of total tax revenue in 1989 (32 percent) than in the majority of OECD countries during the 1980s (and well above the average for the Social Welfare States of 25 percent). These contributions also absorb a slightly higher share of GDP in Hungary (15 percent) than the average for OECD or the Social Welfare States (about 10-12 percent).6 9. Social Incomes in relation to Total Household Income The structure of labor compensation in Hungary is compared to that of OECD countries in Table II.8. For each country, the lower line of data indicates the direct compensation of labor (gross wages and salaries); the upper line includes employers' social insurance or private pension contributions on behalf of workers, i.e., representing the full cost of labor to employers. 10. We have also tried to look at the importance of social transfers relative to total household incomes in Hungary as compared to OECD countries. For this comparison to be correct, it is necessary to abstract from biases caused by differences in the sectoral composition of employment in various countries; hence, the following analysis is made of only of "average 4 Since the average real value of health and education benefits per capita cannot be estimated directly, due both to the absence of a specific price deflator for these sectors and problems in defining the denominator (number of persons actually using the sector services), this third factor is calculated in Table II.6 as a residual. 5 Using a different methodology to estimate the contribution of the demographic structure, eligibility, and real benefits to growth in the share of pension expenditure to GDP, two-thirds of this growth is found to be attributable to increases in eligibility. (See Table IV.1) Based on data in OECD, "Revenue Statistics of OECD Member Countries", 1965-87. 130 Table II.7 Social Insurance Contribution Rates at the Income Level of an APW, 1987 (as a percentage of gross earnings) ------------------------------------------------------- Employers Employees ---------------------------------------------_.--------- Australia 0.0 1.3 Austria 22.6 16.4 Belgium 44.2 12.1 Canada 5.1 4.1 Denmark 3.1 2.0 Finland 6.1 3.4 France 16.7 Germany 17.1 17.1 Greece 21.8 13.3 Iceland 2.0 2.0 Ireland 12.3 7.7 Italy 43.5 8.9 Japan 7.5 7.0 Luxembourg 14.7 12.2 Netherlands 23.9 25.6 New Zealand 12.3 0.0 Norway 15.8 10.9 Portugal 24.5 11.0 Spain 30.9 6.0 Sweden 31.3 0.0 Switzerland 10.3 10.3 Turkey 20.5 14.0 United Kingdom 10.5 9.0 United States 12.4 7.1 OECD Lower Income 22.0 10.4 States (unweighted) OECD Social Welfare 20.7 10.1 States (unweighted) Hungary 10 - 40% of wage 3 - 15% of wages bill (average 33% (average 9% economy-wide) economy-wide) Since 1988: 43% of wage bill 10% of wages ------------------------------------------------------- APW means average production worker. Sources: OECD The Tax/Benefit Position of Production Workers, 1984-86. Figures for Australia, France (employers'), Iceland, New Zealand, and Turkey from US Dept of Human Services, "Social Security Programs Throughout the World - 1987"; rates are share of payroll. APW.WK1 02/05/91 131 Table 11.8: Compensation of Employees as Percentage of GDP Page 1 of 2. 1980 1981 1982 1983 1984 1985 1986 1987 1988 AUSTRALIA Wages & Contributions 53.0 54.6 55.7 52.1 51.6 51.0 51.0 49.7 49.3 Wages -- -- -- - -- -- -- -- -- AUSTRIA Wages & Contributions 54.9 55.8 54.4 53.5 53.0 53.2 53.8 53.6 52.3 Wages 45.6 46.2 44.9 44.0 43.3 43.4 44.0 43.8 42.6 BELGIUM Wages & Contributions 59.0 58.9 57.4 56.6 55.9 54.9 54.9 54.0 52.5 Wages 50.6 50.7 49.7 48.7 47.5 46.0 45.9 44.7 43.3 CANADA Wages & Contributions 55.7 56.0 56.9 55.1 54.1 54.3 54.9 54.8 54.7 Wages 51.0 50.9 51.6 49.8 48.9 49.0 49.5 49.4 49.2 DENMARK Wages & Contributions 56.9 56.4 56.0 55.3 54.1 53.8 53.4 55.9 55.8 Wages -- -- -- -- -- - -- -- - FINLAND Wages & Contributions 54.5 55.6 54.8 54.3 53.7 55.0 55.3 54.9 53.8 Wages 44.4 45.4 44.9 44.7 44.3 44.9 45.0 44.7 43.8 FRANCE Wages & Contributions 56.0 56.6 56.6 56.3 55.5 54.7 53.3 52.9 51.9 Wages 41.4 41.8 41.6 41.0 40.2 39.5 38.4 37.9 37.2 GERMANY Wages & Contributions 57.1 57.3 56.5 55.0 54.3 54.1 53.9 54.0 53.4 Wages 46.7 46.7 46.0 44.5 43.8 43.6 43.4 43.4 42.9 GREECE Wages & Contributions 39.2 40.6 41.5 41.8 41.9 43.7 41.4 39.2 40.9 Wages -- -- -- -- -- -- -- - -- ICELAND Wages & Contributions 49.9 49.5 49.4 44.9 43.9 47.4 47.0 53.3 -- Wages -- -- -- -- -- -- -- -- -. IRELAND Wages & Contributions 59.7 58.3 56.4 55.8 54.7 54.1 54.2 53.6 Wages -- -- -- - - -- -- -- -- ITALY Wages & Contributions 47.5 48.3 47.9 47.4 46.1 46.0 45.0 45.2 45.4 Wages -- -- -- -- -- -- -- -- -- JAPAN Wages & Contributions 54.3 54.8 55.3 56.0 55.5 54.7 55.0 55.0 54.8 Wages 48.3 48.1 48.4 48.9 48.3 47.4 47.5 47.6 47.1 LUXEMBOURG Wages & Contributions 64.1 66.0 63.0 61.1 59.5 59.2 60.5 62.5 62.6 Wages -- -- -- - - - - -- -- -- -- For notes, see next page. 132 Table II.8: Compensation of Employees as Percentage of GDP Page 2 of 2. -----------------------------------------------------------------------__----__------------------------ 1980 1981 1982 1983 1984 1985 1986 1987 1988 -------------------------------------------------------------------__--------__------------------------ NETHERLANDS Wages & Contributions 58.8 57.1 56.3 55.1 52.5 51.9 52.5 53.8 52.9 Wages 45.2 44.0 43.5 42.0 40.1 39.7 40.3 41.4 40.9 NEW ZEALAND Wages & Contributions 56.8 56.5 55.1 50.7 48.8 49.4 48.8 48.8 47.2 Wages -- -- - - -- - - -- NORWAY Wages & Contributions 51.0 50.1 50.6 49.2 47.8 47.9 52.8 54.5 55.5 Wages 43.5 42.7 43.1 42.1 40.9 41.0 45.2 46.6 47.1 PORTUGAL Wages & Contributions 51.1 52.6 50.6 51.0 49.3 47.5 45.4 46.1 45.5 Wages -- -- -- -- -- -- -- -- SPAIN Wages & Contributions 51.3 51.4 56.5 50.2 47.4 46.3 45.8 -- -- Wages 39.6 39.4 38.6 38.2 36.2 35.0 34.4 -- -- SWEDEN Wages & Contributions 64.i 64.0 61.8 59.7 58.4 58.4 58.5 58.6 58.6 Wages 46.8 46.2 44.7 43.1 42.4 42.5 42.6 43.0 42.7 SWITZERLAND Wages & Contributions 60.3 60.4 61.1 61.5 61.2 60.7 60.3 60.4 60.7 Wages 52.6 52.7 53.4 53.6 53.2 52.8 52.4 52.6 52.8 UNITED KINGDOM Wages & Contributions 59.7 59.0 57.1 56.0 55.7 54.9 55.2 54.3 -- Wages 51.6 50.4 49.1 48.0 48.0 47.6 48.1 47.6 -- UNITED STATES Wages & Contributions 61.2 60.4 61.4 60.6 59.8 60.0 60.4 60.6 -- Wages 51.1 50.3 50.9 50.1 49.5 49.9 50.2 50.5 -- HUNGARY Hungary, Gross Income 49.6 49.1 48.8 49.9 52.5 54.2 55.7 53.3 51.1 less SS Contribution 43.5 43.0 42.1 42.6 42.6 43.7 44.9 43.4 40.6 Unweighted Averages OECD LOWER INCOME STATES Wages & Contributions 50.3 50.7 49.8 49.7 48.3 47.9 46.7 34.7 21.6 Wages -- -- OECD SOCIAL WELFARE STATES Wages & Contributions 57.2 57.0 56.2 55.2 54.0 53.8 54.4 54.9 54.4 Wages 45.3 45.1 44.6 43.6 42.6 42.3 42.9 43.0 42.5 -------------------------------------------------------------------__--------__------------------------ For OECD: "Wages & Contributions" equals (i) all payments by producers of wages and salaries to their employees in kind and in cash, and (ii) contributions in respect of their employees to social security, private pension, casualty insurance, life insurance and similar schemes. Lower line ("Wages") equals (i) only. For Hungary: Upper line equals gross labor income in cash and kind plus employers' and self-employed social security contributions. Sources: OECD National Accounts 1976-88, Detailed Tables, Volume II. Top line taken either from Table 13 (Cost Components of Value-Added by Kind of Activity), Total Compensation of Employees or Table 8 (Accounts for Households and Private Unincorporated Enterprises), Income and Outlay Account, Compensation of Employees. Lower line from Table 8 where available ("Wages and Salaries"). Hungary - IMF-Recent Economic Developments and CSO, Statistical Yearbook. 01/30/91 COMPENSl.WK1 133 ANNEX II Page 16 of 33 production workers" (APW) in each country.7 An OECD study permits examination of the net tax and benefit position of such workers, considering as benefits only the cash transfers universally available to them (i.e., family allowances). (See Table II.9). The comparison reveals that the sum of take- home pay and cash transfers (equivalent to disposable income) was 106-109 percent of gross earnings in Hungary in 1988-89, while only one other OECD country had a value above 100 percent. This imbalance is mainly the result of Hungary's relatively high family allowances, although the rate of personal income tax paid by this category of worker in Hungary (about 11 percent) is also significantly below the OECD average (14 percent in 1987) (Table II.10).8 B. Composition of Household Incomes - Income Distribution 11. -The 1980s witnessed significant changes in the composition of labor income. Table II.11 shows that between 1978 and 1987, earnings from the primary employment in the socialist sector declined from 50 percent of total household income to 41 percent, while earnings from secondary employment in this sector rose from 7 percent to 11 percent and private sector earnings from 3 percent to 6 percent. As evident from Table II.12, real wages in the socialist sector declined sharply during the 1980s. Households maintained the growth of real per capita labor income through increased reliance on two fulltime earners (women's labor participation rate rose from 68 percent to 74 percent between 1976 and 1988) and an increased practice of secondary employment alongside the main job. "Second economy" employment gained impetus around 1981 when private economic activity was liberalized, and the share of such earnings peaked in 1987 before introduction of the personal income tax. 12. Income Distribution. Per capita personal income (from all sources combined) became more equally distributed during the two decades between 1962- 82, with the main determinant being the low dispersion of socialist sector wages. Between 1982-87, however, this trend was markedly reversed by the spread of supplementary employment, particularly in the private sector, which provided relatively more earning opportunity to individuals at the upper income deciles. In 1987, the lowest fifth of the income distribution received 10.5 percent of income and the top fifth 34.7 percent, in contrast to 11.3 percent and 32.3 percent, respectively, in 1982 (Table I1.13). The introduction of the PIT in 1988 moderated the dispersion of income slightly, 7 Defined as adult (male and female) full-time production workers (from among all wage and salary earners) in the manufacturing sector (classified as Division 3 of Industrial Standard Industrial Classification code). Such workers constituted 28% of all active earners in Hungary in 1988, as compared to an average of 25% for all OECD countries in 1986. 8 In the comparison for single APWs, take-home pay (without family allowances) was 78 percent of gross earnings in 1988-89 for Hungary, versus an average of 72 percent in 1987 for the countries listed in Table II.9. It should be noted that the unit used in the Table (one-earner couples) is relatively rare in Hungary. 134 TabLe 11.9: TAKE-HOME PAY PLUS CASH TRANSFERS Two-child families 1/ Country ------------------------------------------------------ 1984 1985 1986 1987 1988 1989 Australia 86.7 85.5 84.5 84.1 Austria 90.4 91.4 90.0 92.7 Belgium 82.2 81.5 82.3 82.7 Canada 90.2 89.0 87.8 86.9 Denmark 64.8 64.1 63.3 67.8 FinLand 78.5 77.4 78.4 77.7 France 92.5 92.3 91.6 90.7 Germany 77.4 76.9 79.1 78.8 Greece 102.4 101.5 100.7 101.6 Ireland 79.3 78.6 79.3 78.0 Italy 86.2 81.9 84.7 78.5 Japan 90.4 90.2 90.0 90.3 Luxembourg 94.2 94.5 95.8 97.2 Netherlands 71.1 72.8 73.5 72.8 New Zealand 83.9 84.5 83.3 85.3 Norway 83.1 83.3 82.4 82.9 Portugal 86.4 87.8 91.5 90.2 Spain 86.2 86.7 84.8 84.4 Sweden 72.6 75.4 74.2 74.8 Switzerland 87.9 89.4 89.9 90.0 Turkey 66.3 68.9 68.7 n.a. United Kingdom 81.6 81.5 81.7 82.1 United States 78.1 77.7 80.4 79.5 Hungary n.a. 114.0 n.a. n.a. 105.7 108.9 OECD Lower Income States 84.1 84.7 85.0 88.6 OECD Social WeLfare States 77.8 78.1 78.0 78.5 Note: Take home-pay refers to gross earning minus income tax and employees' social security contributions. The rate of income tax used in this calculation does not take into account the effects of non-standard tax reLiefs. Cash transfers are transfers universally paid in respect to a wife and two dependent chiLdren between five and twelve years of age ("family allowances"). The total numerator above represents "disposable income". 1/ The unit is a one-earner couple with two children, earning an amount equal to average earnings of production workers in the manufacturing sector. Source: OECD, "The Tax/ Benefit Position of Production Workers, 1984-87" (Paris: OECD, 1988), Table 4, and Bank Staff calculations. 135 Table 11.10: PERSONAL INCOME TAX AT THE INCOME LEVEL OF AN APW 1/ Two-child families 2/ Country 1984 1985 1986 1987 1988 1989 Austratia 15.8 16.8 17.4 17.6 Austria 7.0 7.6 8.3 6.4 Belgium 16.0 16.6 16.1 14.6 Canada 9.3 10.3 11.4 12.2 Denmark 33.8 34.3 35.0 35.7 FinLand 24.6 25.3 24.5 25.1 Germany 10.7 10.9 8.3 8.6 Greece 1.7 - 3.4 2.5 Ireland 15.6 16.1 16.9 17.8 Italy 14.0 16.2 14.6 16.1 Japan 2.6 2.8 3.0 2.7 Luxembourg 2.6 2.2 2.1 1.0 Netherlands 9.4 8.4 8.8 9.1 New Zealand 21.8 24.8 24.9 23.6 Norway 15.5 15.0 15.1 15.2 Portugal 6.0 4.6 2.5 4.0 Spain 8.9 7.8 9.6 10.0 Sweden 34.2 33.9 34.5 35.0 Switzerland 6.2 6.4 6.9 6.5 Turkey 25.3 22.7 22.0 n.a. United Kingdom 18.1 17.9 17.4 16.6 United States 15.2 15.3 12.4 13.3 OECD Lower Income States (unweighted average) 11.5 12.8 10.9 8.6 OECD Social Welfare States (unweighted average) 22.3 22.3 22.3 22.5 Hungary 3/ --- --- --- --- 11.2 11.7 Note: The figures shown in this table do not take into account the effects of non-standard expense-related reliefs, which are found in all countries' income tax systems. In some countries the inclusion of these reliefs would substantially reduce the average rates of income tax shown in this table. 1/ Central, state and local government income tax combined. 2/ The unit is a one-earner couple with two children, earning an amount equal to average earnings of production workers in the manufacturing sector. 3. Based on gross average monthly earnings in socialist sector industry (state-owned and cooperative). Personal income tax was introduced in Hungary in 1988. Calculation of tax takes account of Ft. 12,000 p.a. allowance for employees. Source: OECD, "The Tax Benefit Position of Production Workers, 1984-87" (Paris: OECD, 1988), Table 1, and Bank staff calculations. 136 Table 11.11: CHANGE IN COMPOSITION OF INCOMES OF THE POPULATION (in percent) Earnings from state Incomes Income j TotaL j Social Other Total firms or institutions from self from income incomes j incomes income and from cooperatives employment Ihouse- I from l(in cash I inc. I of --------------------------- and from I hold L labor and in I incomes I popu- Year IMain earnings| ALL other |(smaLL scalel plots I kind) I from I lation I (from full I earning I private I I I wealth Itime employ- I like lenterprises I I I I I ment) I incomes I I I I I Net Value 1971 51.7 5.8 2.1 14.1 73.7 23.5 2.8 100.0 1978 50.3 6.9 3.2 9.9 70.3 29.3 0.4 100.0 1981 44.0 9.5 3.9 9.6. 67.0 32.3 0.7 100.0 1987 40.4 10.5 6.3 8.8 66.0 34.1 -0.1 100.0 Gross Value 1987 41.3 10.5 6.5 8.2 66.5 31.7 1.8 100.0 1988 42.0 10.0 5.7 7.1 64.8 31.5 3.7 100.0 1989 39.4 9.8 6.1 6.5 61.8 33.0 5.2 100.0 Source: Ministry of Finance, Central Statistics Office, and World Bank, "The Hungarian Social Policy Systems and Distribution of Incomes and Households," Budapest 1990, Table 3/a. Note: "Social Incomes" in this table is lower than in Table 11.3a because the latter incLudes unemployment support in kind and pharmaceutical subsidies. 137 Table 11.12: THE GROWTH OF LABOR INCOME AND OF MAIN EARNINGS (in percent) Year Real vaLue of Real Value of Main Earnings income from work (Real Wages) (per capita population) (per capita earners) 1973 = 100 1978 115.0 117.4 1987 119.8 110.6 1988 112.2 105.2 1989 110.2 106.1 1978 = 100 1981 98.1 97.8 1987 104.2 94.3 1988 97.6 89.7 1989 95.8 90.5 Source: Ministry of Finance, Central Statistics Office, and World Bank, "The Hungarian Social Policy Systems Distribution of Incomes and Households," Budapest 1990, Table 3.1/c. 138 ANNEX II Page 21 of 33 as only earnings from main employment were grossed-up.9 In 1988-89, it was the middle income group--those with incomes at or a little higher than the average--who experienced the greatest relative decline in their incomes. 13. Changes in the distribution of income are of course affected by differences in the incidence of various sources of income, transfers, and taxes among the population (Table II.14). Earnings from "second economy" activity--household agricultural plots, tourism (e.g., bed and breakfast provision), other supplementary employment that is registered for statistical and tax purposes--accounted for almost 18 percent of total gross personal income in 1989. Access to such income from small agricultural plots is fairly evenly distributed across income quantiles, with about 55-60 percent of all households involved, while the frequency of informal tourism and other second economy activities increases steadily with each income quantile. On average, about 3 percent of all households participate in bed and breakfast services and about one-third in other second economy activity. Regarding "third economy" earnings, defined as invisible income of an illegal nature or not reported for taxation, reliable statistics are unavailable even from the household income surveys. The total income from such activity has been estimated by experts as possible Ft. 50-70 billion or more (4-6 percent of gross personal income in 1989). (Incidence Study, 1990, Table 2.6/a) 14. Total taxation is concentrated rather heavily on the top income decile, reflecting the progressivity of the personal income tax. The average rate of personal income tax is 9-13 percent of gross per capita personal income at the top two income quintiles and 3-5 percent at the bottom two, in contrast to the fixed 10 percent social insurance premium, which accounts for almost an equivalent share of per capita personal income at the same two extremes (6 percent and 5 percent, respectively). (Incidence Studv, 1990, Table 2.4/a) 15. As another indication of differences in the dispersion of various income sources, the Incidence Study reveals that average per capita net personal income in the first and tenth deciles of the distribution of net personal income differs by a factor of 4.4 in 1989--that is, the average net personal income of individuals in the poorest tenth of the population (Ft. 42,600) was 4.4 times less than that of the richest tenth (Ft. 186,800). The average value of net earnings from main employment at these extreme points in the net income distribution differs by only a factor of 1.5. The discrepancy between 4.4 and 1.5 reflects the inequality produced by supplementary If Table II.13 were calculated in real terms rather than in nominal, the equalizing trend of incomes after 1987 would not appear, because sharp reductions in consumer subsidies at that time disproportionately burdened lower income households. The basic goods consumer price index in 1988 increased 20.1% for active households with low income, 17.2% for those with medium income, and 16.4% for high income. (CSO, Statistical Yearbook. 1989) Increases in family allowances and pensions, although giving relatively higher compensation for large families and older pensions with the lower average per capita incomes, did not fully compensate for the price increases. 139 Table 11.13: THE DISPERSION OF INCOME (per capita personal income distribution) 1962 1972 1982 1987 1989* The share of deciLe in the income: 1st decile 3.6 4.0 4.9 4.5 4.7 2nd decile 5.6 5.9 6.4 6.0 6.2 9th decile 14.1 14.0 13.7 13.8 13.8 10th deciLe 20.7 19.7 18.6 20.9 20.7 The 10th decile as percentage of the 1st Total population 575 493 379 464 440 * Data from micro-simulation estimates; extrapolation made on basis of the 1987 Income survey records and macro data. Source: Ministry of Finance, Central Statistics Office, and WorLd Bank, "The Hungarian Social Policy Systems and Distribution of Incomes and Households," Budapest 1990, Table 4/a. 140 Table 11.14: SHARE OF POPULATION DECILES FROM DIFFERENT KINDS OF INCOME AND TAXES 1989 Kinds of Income 1 2 3 4 5 6 7 8 9 10 ITotal I deciles I SECTION 1: CASH FLOwS Primary Incomes of active earners 3.6 5.3 6.3 7.1 7.8 9.0 10.3 12.1 14.1 24.3 100 (Main earnings, sick pay) Primary Incomes of inactive earners 6.2 8.4 9.1 10.3 11.4 11.2 10.8 11.0 11.8 9.8 100 (Pension, childcare allowances, unempLoyment benefits) Secondary supplementary labor incomes 3.1 4.1 5.4 5.9 7.4 8.8 10.9 13.0 17.2 24.5 100 Secondary social incomes 15.6 13.7 12.7 11.0 10.1 9.5 8.7 7.5 6.1 5.2 100 (Family allowance, student aid, other cash aid) Other income 2.0 3.6 4.4 5.1 6.3 7.9 9.9 13.2 18.3 2944 100 Gross personal income 4.5 5.9 6.8 7.5 8.3 9.2 10.3 11.9 14.1 21.5 100 Taxes 3.1 4.6 5.4 6.2 7.4 8.8 10.4 12.7 15.6 25.8 100 (Pension premiumn, personal income tax, other) Net personal income 4.7 6.2 7.0 7.8 8.5 9.3 10.3 11.7 13.8 20.7 100 SECTION II: FLOWS IN KIND Social income in kind 10.9 11.0 10.4 9.9 10.3 10.0 9.5 9.1 9.4 9.6 100 Subsidies 7.1 7.6 8.0 8.1 9.0 9.2 10.7 10.2 14.0 16.1 100 SociaL income and subsidies 9.2 9.5 9.3 9.1 9.7 9.6 10.1 6.6 11.4 12.5 100 All income and subsidies 5.8 7.0 7.6 8.1 8.8 9.4 10.3 11.1 13.3 18.6 100 Source: Ministry of Finance, Central Statistics Office, and WorLd Bank, "The Hungarian Social Policy Systems and Distribution of Incomes and Households," Budapest 1990, Tables 1.2.1/b and 1.2.2/b. 141 ANNEX II Page 24 of 33 earnings. The dispersion of pensions across the highest and lowest deciles of the net income distribution is even lower (1.3:1). This narrow range results from the policy since 1986 of making a full inflation adjustment only to the pensions of retirees over 70 years of age. (See Table II.15) Thus pensions which corresponded to the two lowest pension levels in 1980 have retained or slightly gained in real value by 1989, while pensions which were in the highest benefit category in 1980 have declined in real terms by 28-36 percent over the decade, depending on the age of the recipient.'" 16. Differences across households in average earnings and average pensions received play a relatively modest role in determining differences of income per capita. It is illustrative, for example, that earnings at about the minimum wage level occur with almost equal frequency between the second and eighth per capita income deciles. The dispersion of incomes per capita is explained mainly by differences in the number of active earners per household (less than 1.5 in households at the lower 15 percent of the per capita income distribution versus 1.75 for the population overall),, and the larger number of children in the lower income quantiles (average household size is 3.5 persons in the bottom tenth of the income distribution and 2.2 in the top tenth). (See Table II.16.) C. The Measurement and Description of Poverty Measurement and Definition of the Poverty level. 17. The Central Statistical Office (CSO) of Hungary has been estimating a "subsistence minimum" income level and a "social minimum" income level annually since 1982. Both estimates are made for a number of social groups (active and inactive households, urban and rural) and varied according to the household size and age and number of dependents. The existing estimates through 1989 were based on a calculation method applied in 1984 and the results were indexed annually. A new method is now being undertaken and will provide calculations for 1989 which can be compared with those made on the old basis. 18. In both methods, the calculation starts with a nutritionally determined diet for six age groups which is priced using national price data and which makes no distinction between active and inactive households. The diet is somewhat different in the (higher) social minimum and this is the principal distinction between the two. Under the old method, regression analysis determined a linear relationship between food expenditure and income (net of housing costs) for all households in the Household Budget Survey (HBS)"; the minima were then calculated by grossing up the nutritional diet expenditure, finding the income level at which one typically attains the minimum food expenditure level. The average costs of housing maintenance, based on the reported expenditures of the respective household groups in the '° See Incidence Study, 1990, Tables 1.2.1/a. and Sections 2.1 and 2.2. "1 Survey taken biennially of 12,000 households (latest in 1989). 142 TabLe 11.15: REAL VALUE OF PENSIONS ESTABLISHED IN 1980 Starting sum 1985 1986 1987 1988 1989 of pensions --------------------------------------------- in 1980 1980 = 100% Those who were 70 or less in 1986 1500 Ft 106.4 107.0 107.3 109.0 114.3 2000 Ft 97.2 96.8 95.7 95.1 97.1 3000 Ft 84.0 82.9 80.9 78.4 77.6 4000 Ft 80.7 78.8 75.9 72.2 68.5 5000 Ft 79.3 77.3 73.9 69.3 63.8 Those who were 70 or more in 1986 1500 Ft 106.4 107.0 109.2 113.7 123.8 2000 Ft 97.2 96.8 97.8 101.3 104.2 3000 Ft 84.0 82.9 83.3 85.8 83.5 4000 Ft 80.7 78.8 79.0 81.1 75.9 5000 Ft 79.3 77.3 77.3 79.2 72.2 Source: Ministry of Finance, Central Statistics Office, and World Bank, "The Hungarian Social Policy Systems and Distribution of Incomes of Households," Budapest 1990, TabLe 3.2/b. 143 Table 11.16: THE COMPOSITION OF MEMBERS AND OF INCOME OF HOUSEHOLDS WITH DIFFERENT NUMBER OF CHILDREN* Households with 0 1 | 2 1 3+ children Number of persons per 100 households Employees 174 180 164 151 Self-employed 7 10 10 10 On child care allowance/fee 0 14 16 22 Pensioners 60 22 15 12 Children, adult dependees, family helpers 20 113 213 357 Total 261 339 418 552 Average main earnings, in thousand Ft Gross income of employees 122 126 131 120 Gross income of the self-employed 275 305 328 297 Average amount of income per households, in thousand Ft Gross earning of empl. or self-empl. 241 266 255 220 Main income of inactive earners** 40 20 17 17 Income from supplementary work 51 52 53 50 Grants for students, other benefit 2 3 3 8 Other income (without family allow.) 31 26 22 18 Gross income (without family allow.) 365 367 350 313 Deductions: taxes, deprivals 63 77 74 60 transfers within hholds. -3 +3 +5 +4 Net income (without family allow.) 299 252 281 257 Family allowance 0 18 44 77 Net total personal income 299 311 325 334 Income per capita, in thousand Ft Net total personal income 115 92 78 60 * Without one-person households ** Child care allowance/fee, pension, unemployment benefit Source: Ministry of Finance, Central Statistics Office, and World Bank, "The Hungarian Social Policy Systems and Distribution of Incomes and Households," Budapest 1990, Table 2.3/a. 144 ANNEX II Page 27 of 33 survey, are then added-in to arrive at the final income minima. Neither the former nor the new method of calculation includes the costs of acquiring housing in the minimum income requirement. The minimum income requirement for the total population is taken as the weighted average of the subgroup minima. 19. Under the new method, a sub-sample of about 1500-2000 households in the survey representing the various social groups is identified which have food expenditure about equivalent to the cost of the nutritional diet. The minima are then taken as these households' total consumption expenditures (including maintenance costs for housing). 20. The subsistence minimum income level based on the former methodology represented about 50 percent of average income, and the social minimum was about 20 percent higher than the subsistence. These proportions are expected to hold for the new calculation as well. Starting in 1991, the subsistence minimum (considered as the poverty line for social policy purposes) is to be revised quarterly. Detailed Description of Poverty Group in 1989. 21. The 1989 Household Budget Survey provides an opportunity to examine in some detail the sample population below the poverty line (the 1989 subsistence income level) in comparison to the sample population at the average income level."2 The per capita subsistence income level for active households in 1989 was Ft. 3940 monthly, 51 percent of the average per capita income for active households in the Survey. The poverty line for inactive households was Ft. 3640 monthly, or 53.5 percent of the average income for the inactive households sampled.13 About ten percent of the HBS households were reported under the subsistence income level. This would overestimate slightly the proportion of poverty in the total population insofar as high income households are underrepresented in the HBS; on the other hand, since the HBS relies on maintenance of an expenditures diary, it also misses the poorest and most disenfranchised population (homeless, the very old or sick, etc.). 22. The HBS reveals that the risk of being poor is about the same for active and inactive households; active households represent about 80 percent of the total population and about the same share of households under the poverty level. On the other hand, the risk of poverty varies with the number of dependents in the household and with age. The greatest incidence of 12 This section is based on Chapter III of the 1990 Incidence Study. Note that the 1989 subsistence income level used here is the original 1984 calculation of minimum consumption indexed to 1989 by the consumer price index. The new subsistence income level being calculated presently by the new method described above may produce some different conclusions, because of changes in the structure of consumption of low income groups between 1984-89. 13 The subsistence income level as a share of average income varied between 62-64% for active households in the Household Budget Surveys conducted between 1978-87, and from 60-63% for inactive households. 145 ANNEX Ii Page 28 of 33 poverty is found in those households with three or more children, and among households of single pensioners (in the latter case, typically women depending only on a widow's pension). For the economically active population, poverty is more likely to be temporary and associated with some event in the family (e.g., divorce, death of a wage earner). Active earners with low educational attainment are also more likely to be poor (the share of adult earners who completed less than eight years of primary school is 20 percent among the poor, versus about 10 percent of the labor force at large). Poverty is also a permanent status among the oldest pensioners (those over 70 years of age), who are typically not able to supplement their pension with earnings; the situation is worst for those receiving above the minimum pension who have not been receiving full indexation for inflation (see above para. 15 regarding the decline in real pension values). 23. In households under the poverty line, the average per capita available monthly income is Ft. 3040, only 40 percent of the respective income for households in the "average income" group (Ft. 7750). The main reasons for this difference are, first, that there are only 150 active earners per 100 households below the poverty line, as compared to 183 in the average income group; poor women are thus more often either dependent or on maternity allowance. Secondly, households under the poverty line have three times more dependents than average households. It is also the case that the income of the active earners under the poverty line is lower than the average. However, in inactive households, the average pension of the poor is not much less than that of the nonpoor (Ft. 4667/month versus Ft. 5417/month); the much larger differences in per capita incomes between these groups is explained by the fact that there are 32 adult dependents per 100 pensioners in households below the poverty line while there are only three in the average inactive household. (See Table II.17). 24. The household survey data permit estimation of the "poverty gap", the total amount of money that would be needed to raise the poor to just above the poverty level. The data from the CSO's 1990 Incidence Study gives a figure of Ft 3,040 per month as per capita income in households beneath the minimum subsistence line in 1989. The line itself in 1989 can be taken as about Ft 3,840 Ft per capita (weighted average of active and inactive level). Thus, the average shortfall is Ft 800 per month or Ft 9,600 per year. With about 1 in 10 of the population below the line this implies a poverty gap of about Ft 10 billion in 1989 (although excluded from the calculation are the institutional population and the homeless) or about 0.6 percent of GDP. 25. Housing conditions. The Incidence Study examined the living conditions of the poor in terms of their access to housing, consumer durables, and general level of consumption. No major differences are apparent between the poor and the average households in the type of tenure (owner-occupied or State-owned housing), except for inactive households, where those of average income are more likely to live in State-owned flats. The density of housing is much greater in the poor households (with 1.87 persons/room) than in average income households (1.19 persons/room), largely reflecting the larger household size in the poverty group. The main distinction between the housing conditions of the two income groups appears in the quality of units--the poor 146 TabLe 11.17: FACTORS DETERMINING INCOME OF POOR AND NONPOOR HOUSEHOLDS, 1989 Average income HousehoLds under households the poverty line In 100 HousehoLds: the number of maLe active earners 93 93 the number of female active earners 90 57 Total 183 150 number of inactive adults 33 39 number of dependants 83 239 Total 299 428 Average yearly gross earning per household (1000 Ft.) 134.1 76.3 Average pension of inactives, Ft./mo. 5417 4667 Average per capita net income (Ft./month) 7750 3040 in inactive households 6800 3080 Source- Ministry of Finance, Central Statistics Office, and World Bank, "The Hungarian Social Policy Systems and Distribution of Incomes and Households," Budapest 1990, Table 1.3/a. 147 Table 11.18: INDICATORS OF HOUSING OF POOR AND NON-POOR HOUSEHOLDS, 1989 Active Households Inactive Households Under Under Average Poverty Average Poverty Income Line Income Line Dwelting size (sq. m) 74.4 68.4 67.9 59.2 Room Number (in percent) No room 0.2 0.6 0 0.3 1 room 6.9 14.2 21.9 25.8 2 rooms 45.5 46.7 51.9 43.6 3 rooms 37.5 33.2 23.3 26.2 4 and more rooms 9.9 5.3 2.9 3.5 Proportion of state owned rented dwellings 17.6 19.0 28.0 19.1 Portion of dwellings both 88.2 64.1 81.6 40.9 with bathroom and WC Without bathroom and WC 6.5 30.6 13.2 48.1 With running water 95.4 73.8 92.0 63.1 With telephone 18.6 5.3 19.4 7.1 With traditional heating 19.7 46.3 34.7 69.1 Source: Ministry of Finance, Central Statistics Office, and World Bank, "The Hungarian Social Policy Systems and Distribution of Incomes of Households," Budapest, 1990, Table 1.4.1/a. 148 ANNEX II Page 31 of 33 households have far fewer units equipped with running water, bathroom, or central heating. The fewest amenities exist in the homes of the poor inactive households: almost half have no bathroom facilities, 37 percent have no running water and two-thirds have only traditional heating (Table II.18). 26. Other studies which have examined the distribution of housing in Hungary according to the occupation of the head of the household have had broadly similar findings. Research conducted in the 1970s and early 1980s generally found that households in the State rental sector had higher incomes and socioeconomic status than households in the owner-occupied sector. The microcensus in 1984 found no major difference between the rental and nonrental sectors in the occupational distribution, although retired and white collar households are more likely to rent and agricultural workers to own; this pattern also reflects the prevalence of State rental housing in urban areas. Within the rental sector, however, the inequalities are more striking. Renter households of unskilled, semi-skilled, skilled workers, and agricultural workers have the highest housing density, while the share of housing units "without comfort" (amenities) is higher than the average in households of semi-skilled, unskilled, agricultural, and retired workers. Managers, intellectuals, and retired white collar workers are more likely to live in the largest flats (over 2.5 rooms), which indicates that inequities in density (floor space per capita) are not explained entirely by household size."14 27. Consumption. Households under the poverty line were reported to be almost as equipped with the major consumer durables--refrigerator, washing machines, radio, and television--as the average income households. The per capita consumption expenditures of the poor households were on average about 50 percent of those of the average income households. The smallest difference appeared in expenditure on construction and real estate, where poor active households spent virtually the same amount as their counterparts with average incomes. Expenditure on food, alcohol, and smoking materials by households under the poverty line was also relatively high (60 percent of the average). D. Demographic Prospects 28. Table II.19 summarizes the main parameters used for five alternative scenarios of population growth and structure in 1990-2051. Scenario II, the "medium" set of assumptions, is used in this Report's projections of social policy expenditures in Chapter 10. 14 Jozsef Hegedus and Ivan Tosics, "The Hungarian State-Rental Sector: its Development and Present Problems," Background paper for World Bank Housing Mission to Hungary, May 3-18, 1990. 149 Table 11.19: HUNGARY - DEMOGRAPHIC ASSUMPTIONS, 1986-2051 1/ (Scenarios IA, IB, II, IIIA, IIIB) 1986 2021 2051 Gross fertility rate Scenarios IA and IB (High) 2.1 2.1 2.1 Scenario 11 (Medium) 1.9 1.9 1.9 Scenario IIIA and 1118 (Low) 1.8 1.8 1.8 Male Female Male Female Male Female Life expectancy (in years)2/ Scenarios IA and IIIA At age: 0 (High) 65.1 73.1 72.6 77.4 77.1 82.9 Scenario II (Medium) At age: 0 65.1 73.1 70.8 77.5 73.2 79.3 20 50.8 58.6 55.9 62.7 58.3 64.5 40 26.9 34.3 31.5 38.2 33.8 39.8 60 10.3 16.0 13.6 19.2 15.3 20.6 Scenarios iB and IIIB (Low) At age: 0 65.1 73.1 69.2 75.4 69.2 75.4 1/ Net migration is assumed to be negligible in all scenarios. 2/ Data on life expectancy refer to the 1980-84 period. Source: Research Institute for Population Sciences, Central Statistical Office; and IMF staff estimates. Table taken from IMF, "Social Security Reform in Hungary," October 12, 1990, Tables 38 and 39. Table II.20: Hungary: Age Distribution of Total Population Projection (Scenario II), 1986-2051 Medium Scenario (In thousands) Age Bracket 1986 1989 1990 1991 1996 2000 2006 2011 2016 2021 2026 2031 2036 2041 2046 2051 0 - 4 646 621 621 618 654 699 689 629 601 614 638 625 598 582 585 588 5 - 9 821 699 662 644 617 641 705 688 628 600 614 637 624 597 582 584 10 - 14 812 882 863 820 643 619 652 704 687 628 599 613 637 624 597 581 15 - 19 719 740 770 810 818 659 615 (,I 703 686 627 599 613 636 624 596 20 - 24 638 678 698 715 807 857 640 613 649 702 685 625 598 611 635 623 25 - 29 755 657 644 634 711 763 811 637 611 646 699 683 624 596 610 634 30 - 34 904 864 803 749 629 689 797 806 633 608 643 696 680 622 594 608 35 - 39 781 838 873 893 739 631 699 789 799 628 603 639 692 676 618 591 40 - 44 699 724 733 766 875 777 611 688 779 789 621 597 633 686 671 613 45 - 49 641 667 686 677 743 832 706 597 674 764 775 611 588 624 677 662 50 - 54 639 609 605 612 647 681 819 682 578 655 743 756 597 575 611 663 55 - 59 645 624 614 598 574 616 674 778 651 555 628 716 730 577 557 592 60 - 64 616 591 590 585 543 518 561 625 728 613 522 594 680 694 551 531 65 - 69 369 532 535 536 511 489 468 504 568 667 560 481 551 630 646 512 70 - 74 411 273 269 300 434 421 400 398 435 498 574 489 423 483 556 570 75 - 79 305 318 318 294 218 312 307 295 297 326 385 457 393 337 387 446 80 - 84 158 168 174 179 171 115 185 181 173 175 210 260 312 261 225 259 85 + 82 88 89 90 102 108 89 105 108 106 111 142 178 194 182 163 Total 10,640 10,572 10,546 10,521 10,438 10,428 10,427 10,370 10,301 10,260 10,235 10,222 10,151 10,006 9,906 9,815 Source: International Monetary Fund, Social Security Reform in Hungary, October 1990, Table 40. 151 ANNEX III Page 1 of 3 A Methodological Note on the Analysis of Incidence of Social Income The distributional analysis of social income in the Incidence Studies of 1989 and 1990 was based on a statistical allocation of social expenditures according to the characteristics and reported consumption of the population as indicated in the Household Budget Survey. The allocation of cash transfers was based on reported receipt of income sources, adjusted where necessary by aggregate data on the characteristics of actual beneficiaries. Subsidy expenditures were allocated according to reported consumption and expenditure on the subsidized goods and services by households. In the case of social income in kind, higher education benefits were attributed to various types of households according to the known distribution of enrollments, values, for hospital care according to the known patient profile, etc. Regarding the components of social income for which socioeconomic and demographic data on actual beneficiaries were unavailable, the researchers imputed a distribution of benefits based on related attributes of the survey population; e.g., outpatient care was attributed according to reported indicators of the health status of households, such as expenditures on medications, amount of sick leave taken, etc. (See Appendix to 1990 Incidence Study for a detailed description of methodology.) A household survey designed to measure consumption of and demand for various social services could produce more reliable evidence of the incidence of benefits in kind. 152 Table 111.1: SHARE OF LOWER AND UPPER DECILES FROM INCOME, TAXES AND TRANSFERS, 1989 (in percent) 1 j 2 9 10 I | Total decile j decile decile j decile Main earnings 3.6 5.3 14.2 24.5 100 Sick Payment 5.1 6.2 12.8 19.9 100 Primary incomes of active earners 3.6 5.3 14.1 24.3 100 Pension 5.3 7.5 12.4 10.2 100 Child care alt./fee 19.1 20.8 3.7 4.6 100 Unempl. benefit 11.1 12.4 12.9 1.1 100 Primary incomes of of inactive earners 6.2 8.4 11.8 9.8 100 Secondary Supplementary labor incomes 3.1 4.1 17.2 24.5 100 Family allowance 16.4 14.2 5.6 4.7 100 Grants for students 4.8 7.5 12.8 9.9 100 Other benefit, aid 13.9 12.4 7.4 6.9 100 Secondary social incomes 15.6 13.7 6.1 5.2 100 other income 2.0 3.6 18.3 29.4 100 Gross personal income 4.5 5.9 14.1 21.5 100 Pension premiums 3.5 5.3 15.3 21.4 100 Personal income tax 1.9 3.5 16.4 30.9 100 Other taxes, duties 6.2 7.1 13.2 17.5 100 Taxes 3.1 4.6 15.6 25.8 100 Household transfer received + 6.5 7.1 14.1 19.4 100 Household transfer given - 7.0 7.3 13.6 19.7 100 Net Personal Income 4.7 6.2 13.8 20.7 100 Source: Ministry of Finance, Central Statistics Office, and World Bank, "The Hungarian Social Policy Systems and Distribution of Incomes and Households," Budapest 1990, Table 1.2.1/f. 153 Table 111.2: SOCIAL INCOMES IN KIND AND SUBSIDIES RANKED BY THEIR SHARE IN THE LOWER AND IN THE UPPER DECILE, 1989 (in percent) i 9 l 10 Totat decile I decile I I decile I decile I T 1 Social items 40.7 37.0 0.0 0.0 100 2 Nurseries 25.6 9.1 1.5 2.9 100 3 Kindergartens 19.5 16.3 5.2 5.4 100 4 Primary schools 15.9 14.7 6.8 4.6 100 5 School books 12.8 12.5 7.5 5.1 100 6 Secondary schools 8.6 9.7 8.5 4.3 100 7 Water, sewage for owner occ. dwellings 10.0 10.6 9.4 9.8 100 8 owner housing inv. 16.3 10.7 10.4 15.6 100 9 MiLk 8.7 9.0 10.5 10.6 100 10 Other health 9.3 9.4 10.4 11.7 100 11 Medicine 8.7 9.3 11.2 11.4 100 12 HospitaLs 8.2 9.2 10.2 11.4 100 13 Social income in kind and subsidies 9.2 9.5 11.4 12.5 100 14 Transportation 7.9 8.0 12.4 10.2 100 15 Dairy products 6.8 8.1 11.8 13.5 100 16 Heating 6.3 7.6 12.5 13.0 100 17 Depr. of state dwellings 5.7 7.6 14.4 14.6 100 18 Rent 5.7 7.6 14.4 15.2 100 19 Water, sewage for state dwellings 5.7 7.6 14.4 14.6 100 20 Culture, sport 6.5 7.4 12.8 16.0 100 21 Canteen bonuses 5.0 6.8 14.1 14.7 100 22 Higher education 2.7 8.9 10.6 13.2 100 23 Other trainings 4.3 6.5 12.6 15.7 100 24 Personal (net) income 4.7 6.2 13.9 20.5 100 25 Mortgage payment 4.8 6.0 17.4 21.5 100 26 Vacation 2.9 5.6 14.0 16.1 100 27 Theater, etc. 3.8 5.3 15.7 20.4 100 Source: Ministry of Finance, Central Statistics Office, and World Bank, "The Hungarian Social Policy Systems and Distribution of Incomes and Households," Budapest 1990, Table 1.2.2./I. 154 ANNEX IV Page 1 of 6 Social Insurance Pensions 1. Pension expenditure and eligibility trends. Table IV.A examines quantitatively the main determinants of past pension expenditure growth (i.e. population aging, rising eligibility, and increased pension levels compared to income of the working population). The Table reveals that over the last three decades almost two thirds of the dramatically increased expenditure share is due to the eligibility ratio (i.e. number of beneficiaries compared to population aged 55 and over), whereas population aging and higher pension levels counted in roughly equal proportion for the remaining increase. The stark impact of beneficiaries on the expenditure trend was even aggravated since 1980, a period during which the rise in the eligibility rate accounts for almost 70 percent of the increased expenditure level. An even higher expenditure trend was prevented during that period by a fall in the pension level compared to the GDP per employed as a result of rising price level and insufficient indexation of pension benefits. 2. Financial scenarios and retirement reform options. The long term projection model used to quantify the financial operations of the pension program captures all the main characteristics of the Hungarian pension system, but due to data and computational constraints has to rely on aggregate relations. The resulting projections can provide only information about magnitudes, which, however, are similar to results for other countries and based on more sophisticated models. 3. The assumptions of the baseline scenario are summarized in Tables IV.2 and IV.3, with the latter Table also outlining the assumptions of five alternative scenarios. In order to convert changes in the eligibility ratio into the effective retirement age for old-age pension, two major assumptions are made: (1) The average retirement age for disabled remains constant; i.e. at the given age, any reduction in the disability eligibility ratio is achieved by a lower inflow into the disability program. (2) Any change in the old age eligibility ratio changes the years of average retirement, which, when subtracted from the average years of death (calculated at age 55) results in the average retirement age.' Furthermore, it is assumed that any change in the length of retirement effects the average years of employment and hence the labor force participation rate. 4. The projections demonstrate that to keep the implicit contribution rate roughly constant over the projection period, a substantial change in the effective retirement age for old age pensioners has to take place. The required large increase in the retirement age serves to compensate for four main effects: (i), the rise in life expectancy at retirement by some five Since with rising retirement age the life expectancy at retirement is also increased, this procedure underestimates somewhat the corresponding years in retirement. On the other hand, however, the use of the bench-mark age of 55 for the disability pension leads to some overestimation. 155 Table IV.1: HUNGARY - DECOMPOSITION OF PENSION EXPENDITURE DEVELOPMENTS, 1960-90 (Ratio of final year as proportion of ratio in base year; percentage contribution to change in expenditure ratio is shown in parenthesis) Due to changes in: Period 1/ Change in -------------------- Expenditure Dependency EligibiLity Transfer Inactive ratio 2/ ratio 3/ ratio 4/ ratio 5/ ratio 6/ 1960-90 3.88 1.27 2.35 1.25 1.03 (100.0) (17.7) (63.1) (16.7) (2.5) 1970-90 2.51 1.07 1.63 1.37 1.06 (100.0) (7.3) (52.9) (33.8) (6.1) 1980-90 1.26 1.06 1.17 0.97 1.04 (100.0) (25.8) (69.3) (-12.7) (17.7) 1960-75 2.23 1.19 1.76 1.09 0.97 (100.0) (22.1) (70.9) (10.6) (3.5) 1975-90 1.74 1.07 1.33 1.15 1.06 (100.0) (11.4) (51.9) (25.6) (11.2) 1/ Preliminary estimates for 1990. 2/ Pension expenditure to GDP ratio. 3/ Population 55+ to population 15-54. 4/ Beneficiaries to population 55+. 5/ Average pension level to GDP per employed. 6/ Population 15-54 to number of employed. Sources: Hungarian Authorities, and IMF staff estimates. 156 Table IV.2 - Hungary: Assumptions for Baseline Policy Scenario 1990 2000 2021 2051 Eligibility Ratios Old age 0.58 0.65 0.65 0.65 (Number of pensioners/Pop 55 +) Disability 0.22 0.25 0.25 0.25 Widows 0.16 0.18 0.15 0.10 Other 0.02 0.01 0.00 0.00 Replacement Rates Old age 0.93 0.93 0.93 0.93 (Average new benefits/ Disability 0.86 0.86 0.86 0.86 pension base) Widows 0.65 0.65 0.65 0.65 Other 0.45 0.45 0.45 0.45 New Entrants Ratios Old age 0.080 0.070 0.065 0.060 (New entrants/ Disability 0.120 0.110 0.100 0.090 total pensioners) Widows 0.080 0.080 0.075 0.070 Other 0.040 0.035 0.030 0.020 Employment rate 0.81 0.72 0.78 0.78 Unemployment rate 0.02 0.03 0.03 0.03 Labor force participation rate 2/ 0.83 0.80 0.80 0.80 Real GDP per employed growth rate -0.065 0.04 0.04 0.04 Shares of Labor income/GDP 0.430 0.430 0.430 0.430 Contribution base/labor income 0.760 0.760 0.760 0.760 Consumer price index 0.300 0.100 0.020 0.020 Pension indexation factor 0.900 1.000 1.000 1.000 1/ Linear interpolation of parameter values between the anchor years, 1990, 2000, 2021 and 2051, except for unemployment rate and GDP growth, where values for intervening years 1991-99 are exogenous (See Table 10.1). 2/ Active labor force to population 15 to 54 years of age. Source: Hungarian Authorities, and Bank Staff assumptions. m:\hun\socpol\tabiiib2.doc 157 Table IV.3: Eligibility Ratio and Effective Retirement Age 1990 2000 2021 2051 Life expectancy at age 60 1/ male 10.30 11.87 13.60 15.30 female 16.00 17.52 19.20 20.60 average, unweighted 13.15 14.70 16.40 17.95 Life expectancy at age 55 18.05 19.60 21.30 22.85 Average year of death 73.05 74.60 76.30 77.85 Baseline scenario Eligibility ratio Total 80.0 90.0 90.0 90.0 Old-age 58.0 65.0 65.0 65.0 Disability 22.0 25.0 25.0 25.0 Retirement age Old-age 59.7 57.6 57.9 58.1 Change (+increase) -2.0 -1.8 -1.6 Alternative scenario I Eligibility ratio Total 80.0 80.0 80.0 80.0 Old-age 58.0 60.0 62.0 62.0 Disability 22.0 20.0 18.0 18.0 Retirement age Old-age 59.7 59.9 60.2 60.6 Change (+increase) 0.3 0.6 ' 0.9 Alternative scenario II Eligibility ratio Total 80.0 75.0 70.0 70.0 Old-age 58.0 55.0 52.0 52.0 Disability 22.0 20.0 18.0 18.0 Retirement age Old-age 59.7 61.2 62.8 63.4 Change (+increase) 1.5 3.2 3.7 Alternative scenario III Eligibility ratio Total 80.0 70.0 60.0 60.0 Old-age 58.0 50.0 42.0 42.0 Disability 22.0 20.0 18.0 18.0 Retirement age Old-age 59.7 62.4 65.4 66.2 Change (+increase) 2.7 5.8 6.5 Alternative scenario IV Eligibility ratio Total 80.0 70.0 60.0 55.0 Old-age 58.0 50.0 42.0 37.0 Disability 22.0 20.0 18.0 18.0 Retirement age Old-age 59.7 62.4 65.4 67.6 Change (+increase) 2.7 5.8 7.9 Alternative scenario V Eligibility ratio Total 80.0 70.0 52.0 50.0 Old-age 58.0 50.0 34.0 32.0 Disability 22.0 20.0 18.0 18.0 Retirement age Old-age 59.7 62.4 67.5 69.0 Change (+increase) 2.7 7.9 9.3 1/ Values for 1986. Sources: Research Institute for Population Sciences, Central Statistical Office and Bank staff estimates and assumptions. m:\hun\socpol\tabiiib3.doc 158 ANNEX IV Page 5 of 6 years, leading, at constant effective retirement age, to corresponding lengthening of the average period for which a pension is granted; (ii) the increase in the effective replacement rates, at constant nominal replacement rates. The assumed reduction in consumer price level change, declining from 30 percent in 1990 to 10 percent in 2000 and 2 percent in 2021, increases the initial pension level compared to the last labor income since the non- indexation of prior earnings becomes less effective;2 (iii) the price indexation of all pension benefits from 2000 onward, and (iv) the assumed transitory large increase in unemployment, leveling off to an unemployment rate of 3 percent in later years, which result in a low number of contributors but not necessarily lower number of retirees or lower pension level. Tax treatment 5. It is recommended to eliminate the inconsistent tax treatment of contributions and benefits through an inclusion of benefits into the personal income tax base (for retirees) while excluding the contributions of the population from the tax base (for employed). The following tentative considerations provide preliminary orders of magnitude of the main expenditure and revenue effects involved. 6. In 1990, the average tax rate under the PIT is expected to amount. to 17 percent, the average marginal tax rate to some 30 percent.3 Out of 2.5 million pensioners, currently only some 0.5 million pay PIT (those with additional earnings and above the allowance of Ft 120.000 per year). If the currently exempted pension benefits were to be fully subject to PIT, the average tax rate on pensions (without additional earnings) should be around 5 percent; taking account of additional earnings, the effective marginal tax rate on pensions could be around 10 percent. Against this background, and assuming a constant implicit pension contribution rate of 30 percent, the following expenditure and revenue effects can be estimated. 7. In the short run, if the current stock of pension benefits is to be compensated for the average tax effect on pensions only (i.e. 5 percent), pension expenditure and the implicit contribution rate have to increase by the same percentage, amounting to a rise by some 1.5 percentage points for the latter. The corresponding tax revenues from the full inclusion of benefits into the PIT base, however, would increase by the average marginal tax rate on 2 To illustrate the magnitudes, under the baseline scenario the effective replacement rate (new average old age pension as a percentage of the average contribution base of that year) increases from 60 percent in 1991, to 70 percent in 2000, and to 83 percent from 2021 onward. The nominal replacement rate (new pension as a percentage of the pension base which is calculated as the average of the contribution base of the three prior years) remains constant at 93 percent throughout the projection period. 3 The percentage of grossing up of the family allowances estimated by the Hungarian authorities, if they were to be included into the PIT base in order to achieve fiscal neutrality. 159 ANNEX IV Page 6 of 6 pensions (i.e. 10 percent), equivalent to some 3 percentage points of the implicit contribution rate. The deduction of the contributions of the population (10 percent) from the PIT base would lead to a fall in PIT revenues, amounting to some 3 percentage points of the implicit contribution rate (10 percent contribution rate with an average marginal tax rate of 30 percent). Consequently, in the short run, the positive and negative revenue effects with regard to PIT may be expected to cancel out (probably on the low side), and the net expenditure effect on the pension program amounts to some 1.5 percentage points of the implicit contribution rate. 8. In the long run, when all retirees are in receipt of pension benefits based on gross earnings, pension expenditure will have increased by the percentage of the relevant average tax rate on earnings prior to retirement, say 17 percent, if we assume the same average tax rate as for the total population. This increase corresponds to a rise by some 5 percentage points of the implicit contribution rate. As regards the PIT revenue effects, the revenue loss from excluding contributions from the tax base are the same as in the short run (i.e. 3 percentage points of the implicit contribution rate). The revenue gains from the taxation of benefits are likely to be somewhat higher in the long run, since a higher benefit level will increase the average tax rate (hence some 3+ percentage points may be expected). But again, the PIT revenue effects may be expected to largely cancel out (probably this time on the high side). The net expenditure effect on the pension program, however, of some 5 percentage points of the implicit contribution rate is substantial and suggests some correcting adjustments in the pension formula will be necessary, particularly in view of the other expected financial problems of the SIF. 160 ANNEX V Page 1 of 5 Unemplovment Support 1. Provisions of the unemployment benefit scheme in operation prior to March 1, 1991 and of the unemployment insurance in effect from that date are provided in Tables V.1 and 2. The paragraphs below provided a detailed discussion of the issues regarding administration, costing, and monitoring of the unemployment support system. 2. Administration and Job Matching Since 1989, there has been a substantial investment made in computerizing both claims for benefit and notification of job vacancies in the regional Employment Offices. There are at present about 180 local offices and 20 regional offices dealing directly with the unemployed. Some Ft 0.3 billion was spent on hardware in 1989 in setting up a modern PC-based networked system in each employment office system (apparently, a large office would normally have a 386 server with about half a dozen other PCs in the network). The aim is to develop telecommunications links between each office system, first at the level of the old counties and then at national level (at present information exchange and national collection of data is via floppy discs). There are at least two concerns with this system. 3. Firstly, while the computerized claim and matching system represents a considerable achievement, it is not clear that its capability is being adequately realized at present. This applies to the use of data in monitoring the performance of the benefit system (discussed below); it also applies to the use of the system for matching of unemployed persons to vacancies. When a person registers as unemployed the details of his/her skills and the type of work being sought are entered into the computer system and then matched with details of vacancies held in the system at that time. The Mission was told in the office which was visited in Budapest that this matching is only done when the client is in the office at the initial visit; there did not appear to be any attempt to contact suitable applicants in the client stock when an employer registers a new vacancy. 4. Secondly, the speed with which claims are processed needs to be carefully appraised. Eligibility for unemployment benefit is established usually from the claimant's Work Book, a record of a person's employment history in which employers enter the length of service in each job. This is given to a person leaving employment by the employer and he or she is asked for it when claiming benefit. Details of earnings do not appear to be entered in the Work Book however, and the last employer is asked to verify the claimant's statement of these for the purpose of calculating the earnings- related benefit level (which is calculated on the basis of last employment only). The office which the Mission visited did not have hard figures for the average time taken to establish the validity of a claim and the level of any benefit to be paid, but thought that 1-3 months would be the normal gap between claim and payment. This represents a considerable delay. Conflicting reports were given to the Mission on the place at which the actual calculation of benefit takes place, whether this is in the employment office where the 161 ANNEX V Page 2 of 5 claim is made or rather at the county capital offices after the periodic delivery of floppy discs from the employment offices. If at the latter, this may be a contributory factor to delays in benefit payments. 5. Costing Expenditure on Unemployment Benefit The Mission was provided by the Ministry of Labor in October 1990 with projections of the income and costs of the proposed Unemployment Insurance scheme. The projections assume that 80 percent of the registered unemployed stock will receive benefit until the stock exceeds about 160,000, at which point coverage will fall slightly to 75 percent. The assumption of a near constant proportion of the unemployed stock receiving unemployment insurance is highly unrealistic (an assumption of constant coverage of the inflow would be more realistic). As unemployment increases and affects those with good employment histories, coverage may at first rise but a sustained increase in unemployment rapidly brings exhaustion of entitlements through long-term and repeat unemployment, resulting in falling benefit coverage. The experience of OECD countries bears this out.' 6. A more rigorous model of benefit prediction is required. This should be based not on an assumption about the numbers of unemployed in the stock and benefit coverage of the stock (which are outcomes) but rather on the flows in and out of unemployment, i.e. the processes which determine the stock figures. Such a model could be developed in stages. In its most simple form, assumptions could be made about the total inflow into unemployment in each month in the projection period, about the average duration of unemployment experienced by each month's inflow and hence the proportion of each month's inflow leaving unemployment in any month (the reciprocal of the average duration of unemployment). This would generate the unemployed stock. The assumed benefit coverage would now relate to the inflow, rather than the stock, and could initially be taken as a constant. Coupled with an assumption about the number of months of unemployment that would be covered by unemployment insurance, this would determine the coverage of the stock in each month; this would fall if the average unemployment duration of each month's inflow was assumed to rise. The model would allow naturally for a distinction between the levels of benefit paid in the first and second halves of the entitlement period of the new scheme. (More developed versions of the model could make assumption about the numbers of individuals in the inflow in each 1 In Germany, the proportion of the registered unemployed stock receiving unemployment insurance (UI) fell from 51% in 1980 to 36% in 1986, during which time unemployment rose from 3.3% to 8.0%. In Britain, coverage of UI declined from 45% in 1980 to 30% in 1986 as unemployment rose from 6.1% to 11.8%. In both countries, however, UI coverage initially rose during the period (Brinkmann, 1988, Department of Health and Social Security unpublished data, and OECD, 1988, Table 1.7). The coverage by UI of the inflow in Germany showed much less variation than that of the stock, the coefficients of variation for the period 1977-86 being 0.04 and 0.15 respectively (figures not available for Britain). 162 ANNEX V Page 3 of 5 of the categories of benefit entitlement.) The assumptions required in this sort of model could in part be based on data on payment of unemployment benefit currently collected by the National Labor Market Center. 7. Monitoring the Benefit System. The National Labor Market Center collects a wealth of information on spells of unemployment benefit receipt, all in computerized form. The process of job matching and benefit calculation requires details of personal identity number, sex, marital status, work history, previous salary, occupation and skills; these are recorded for each claim for unemployment benefit together with the level of benefit payment and a computer code indicating the reason for the eventual completion of the claim (new job, retirement, death etc.). 8. These data represent a rich national database which should be heavily used for monitoring the incidence and structure of unemployment and the performance of the unemployment benefit system. Emerging problems can be monitored; for example, the recording of the personal identity number will allow the amount of recurrent unemployment to be measured from the data. The Ministry of Labor should aim to publish at regular intervals (e.g. quarterly) basic statistics on the operation of the unemployment benefit system. These should include (i) the proportion of both the registered unemployed stock and inflow in receipt of unemployment insurance, together with a breakdown of the different reasons for non-receipt, and (ii) the proportion of all weeks of registered unemployment covered by unemployment insurance. These statistics on the registered unemployed could then be combined for analysis with sample survey microdata on other socioeconomic characteristics of this group and of the nonregistered unemployed. 163 ANNEX V Page 4 of 5 Table V.1: Unemployment Benefit Scheme in Place from 1/1989 to 3/1/91 Eligibility (a) Unemployment Benefit: 18 months employment in the 3 years prior to claim (b) Temporary Allowance: Expired entitlement to Unemployment Benefit Duration of Benefit (a) Unemployment Benefit: 12 months in a 3 year period (b) Temporary Allowance: 12 months Benefit Rate (a) Unemployment Benefit: (i) 70% of previous gross earnings for first 6 months, then 60%. These rates are 65% and 55% in the event of voluntary quitting when notice is given and 60% and 50% if no notice is given. (ii) minimum benefit is 80% of the minimum wage and maximum benefit is three times the minimum wage. (b) Temporary Allowance: (i) 75% of previous unemployment benefit i.e. 45% of previous gross earnings (less for voluntary quitters) (ii) maximum of twice the minimum wage Deductions from Benefit Both benefits are subject to income tax and a 5% social insurance deduction (rather than the 10% paid while in employment). Financing Paid from the Employment Fund which receives transfers direct from the State Budget. 164 ANNEX V Page 5 of 5 Table V.2: Unemployment Benefit Scheme Effective 3/1/91 Eligibility (a) Unemployment Insurance: 12 months insured employment in the 4 years prior to claim (b) Career Beginners Benefit: (i) Diploma obtained from a secondary or higher education or training institution within the 24 months prior to claim. (ii) registered unemployed for at least 3 months Duration of Benefit (a) Unemployment Insurance Period of Insured Employment Period of Benefit in the previous 4 years Phase I Phase II 360-479 days 90 days 90 days 480-599 .. 120 120 600-719 .. 150 150 720-839 .. 180 180 840-959 .. 210 210 960-1079 .. 240 240 1080-1199 .. 270 .. 270 1200-1319 .. 300 .. 300 1320-1439 .. 330 .. 330 1440 + 360 .. 360 (b) Career Beginners Benefit: 6 months Benefit Rate (a) Unemployment Insurance: (i) For first half of the entitlement period ("Phase I"), 70% of gross earnings during last 12 months of employment. 50% of same earnings base in the second half ("Phase II"). (ii) minimum benefit is the prevailing minimum wage (or previous wage with the last employer if lower than the minimum wage); maximum benefit is three times the minimum wage. (b) Career Beginners Benefit: 75% of the prevailing minimum wage. Deductions from Benefit Both benefits are subject to income tax and a 5% social insurance deduction (rather than the 10% paid while in employment). Financing Paid from the Solidarity Fund financed by employer and employee contribution. 165 ANNEX VI Page 1 of 7 Support for the Family 1. Perspectives on the Importance of Family Allowances in Hungary. The Family Allowance is one of the most important cash benefits currently provided in Hungary, as indicated in Table VI.I. The table also shows that the real value of the family allowance has increased sharply since 1987.' 2. The importance of family allowance in Hungary may be further illustrated by looking at the situation in other industrialized countries. Table VI.2 shows family allowances in Eastern Europe in 1980 to be much more generous in relation to wages than in Western Europe. Hungary at this time had the most generous scheme of all and this generosity has since increased still further. 3. Allowing for the greater variety of instruments used by OECD countries to provide support for the family can significantly alter the comparison. For example, expenditure on universal child benefit in the UK in 1988-89 amounted to 1 percent of GDP and the payment for a two-child family was only 6 percent of average gross full-time earnings, figures well below those for Hungary. However, a number of means-tested benefits in the UK for those on low incomes provide additional support for families. There is a means-tested benefit explicitly for low-income families where the head is in work, and the comprehensive social assistance and housing assistance programmes provide a higher level of benefit for those with children. Such schemes are much less important in Hungary. Overall, cash benefit expenditure on UK families (excluding maternity pay) is said to be some 2 percent of GDP.2 In addition, married couples receive a substantially more favorable treatment in the income tax system, something which does not occur in Hungary. 'One reason for this rise in real value was the introduction of personal income tax in 1988, this being a way of compensating for the lack of any grossing-up of any second economy incomes (Ministry of Finance, 1987, p.7). A second reason was compensation for the removal of price subsidies. 2Source: HM Treasury (1990) Table 14.6. 166 ANNEX VI Page 2 of 7 Table VI.1: Family Allowance (FAM): Measures of Importance. 1987-90 Year 1987 1988 1989 1990 Number of Families 1,357 1,357 1,368 1,452 Receiving FAM ('0OOs) FAM for Two Children as % 22.5 29.2 33.0 n.a. of Average Gross Earnings FAM for Two Children as % 25.5 37.1 42.6 n.a. of Average Net Earnings FAM Expenditure as % 3.4 4.8 6.0 4.6 of Total Household Cash Income FAM Expenditure as % 14.2 17.9 20.5 18.8 of Total Expenditure on Social Income in Cash FAM Expenditure as % of 43.9 48.5 52.4 47.5 Total Non-Pension Expenditure on Social Income in Cash FAM Expenditure as % of GDP 1.9 2.6 3.1 3.1 Sources: line 1: CSO, Statistical Yearbook, 1989, Table 20.6. lines 2 and 3: CSO, Statistical Pocket Book, 1989 p.31, 1988 p.43 and Statistical Yearbook, 1988, p.433 (1989 and 1988 figures are based on averages of figures for "wages" and "earnings"; 1987 gross earnings taken from 1987 Income Survey, 1987 net earnings calculated from 1988 figure and net earnings index. FAM in each year taken as average of January and August rates). lines 4-7: Table 2.2. 167 ANNEX VI Page 3 of 7 Table VI.2: Family Allowances for Two Children as % of Average Remuneration. 1980 Per Cent Eastern Europe Hungary 24.9 Bulgaria 22.2 Czechoslovakia 20.0 Poland 19.6 Romania 17.0 Western Europe Austria 16.9 Belgium 10.7 Netherlands 9.0 Sweden 8.7 UK 8.2 Switzerland 6.9 West Germany 6.6 France 6.5 Norway 6.4 Italy 5.4 Denmark 3.0 Source: ILO (1989) Table 8. 4. Differentiating Payments by Family Size. Table VI.3 illustrates the current schedule of family allowances, indicating the extent of differentiation of payment per child by the number of children in the family. Where both parents are present, the family allowance for the second child is one third higher than the single child rate, whereas the total support given for three children implies an addition of only 4 per cent in the rate for the third child over that for the second. The total child support for the third child may be increased by the special tax allowance for large families but the value of this at the lowest marginal rate of tax represents only another 6 per cent of the second child family allowance (and zero per cent for the non tax- payer). Family allowance payments for subsequent children are at a rate which is actually some 10 per cent lower than for the third child. 5. The treatment of single parent families in Hungary should also be noted; these families account for some 18 percent of cases where allowances are paid in respect of two or less children. Single parents receive higher family allowance for the first child but the increase paid for the second child is only 13 per cent and the rate declines for the third child. In fact, 168 ANNEX VI Page 4 of 7 single parents with three children receive the same total child support as families with both parents present. 6. Taxation of Family Allowance. In considering the implications of subjecting family allowance to taxation, the first issue is the revenue impact. Since the PIT is a progressive tax, with both a tax free allowance and a number of marginal rates, the revenue from treating family allowance as the wife's income could be expected to be significantly lower than in the case where the husband paid the tax. 7. Some indication of the quantitative importance of this difference may be gained from looking at the distribution of marginal tax rates faced by men and women in their main employment. Table VI.4 gives estimates using the most recent available published data (for 1988) on the distribution of earnings in Hungary and applying the tax system prevailing at that time. Table VI.5 provides the same, based on preliminary results of the 1990 Income Survey. The distributions are given separately for men and women but it should be noted that the data do not separately identify those who are married. 8. Nearly a quarter of women are estimated to have paid no tax on their main earnings in 1988 and another 30 percent to have faced only the lowest marginal rate. This compares to figures of 8 percent and 18 percent respectively for men. The median marginal rate is estimated to have been 30 percent for working men but only 20 percent for working women, and in addition there are those women not at work. 169 ANNEX VI Page 5 of 7 Table VI,3: Rates of Family Allowance and Number of Recipients Family Allowance - Forints per month per child January 1990 August 1990 Both Parents Present First Child 1,770 1,870 Second Child 2,370 2,470 Third Child 2,460 2,560 Fourth plus Child 2,200 2,300 Both Parents Present One Child (where previously 2+) 2,070 2,170 Single Parent First Child 2,070 2,170 Second Child 2,330 2,430 Third Child 2,200 2,300 Number of Families Receiving Family Size July 1989 August 1990 One Child 171,000 305,000 One Child (where previously 2+) 248,000 251,000 One Child (single parent) 147,000 143,000 Two Children 554,000 546,000 Two Children (single parent) 63,000 63,000 Three Children 114,000 ) 144,000 Four or more Children 30,000 Notes: The family allowance is received until the age of 16 (20 in the case of children continuing secondary education or training, no age limit for chronically ill or handicapped children). All allowances are Ft 100 per month higher up to the age of three (and 20 percent higher for the chronically ill or handicapped). Source: CSO Statistical Pocket Book, 1989, p.53, and Ministry of Welfare. 170 ANNEX VI Page 6 of 7 9. It is of course the progression in the PIT which provides the rationale for taxing family allowance. A proportional tax with a single rate levied on all income would imply an equal reduction in its value to all recipients; in this situation taxing the allowance would be no different in its effect to merely reducing the level of benefit paid (and the reduction through the tax system would be administratively much more complicated). With this in mind, it is important to note that the structure of tax rates has been greatly simplified since the introduction of the PIT in 1988. This is shown by a comparison of the previous table with Table VI.5. In 1990, the 30 percent marginal tax rate applied to 71 percent of men in their full-time jobs and 60 percent of women. 10. A second point to be made is that the tax payable on family allowance would not necessarily be at the same marginal rate as that applied to earnings. Given the high level of the family allowance in Hungary, it would clearly be possible for the benefit to push the recipient into a higher tax bracket. In fact, the payment for two children in August 1990 of Ft 4,340 per month exceeds the range of earnings taxed at the marginal rate of 15 percent, implying that at least part of the allowance would always be taxed at 30 percent for a recipient paying the 15 percent rate on his or her earnings. The greater the number of children, the higher the probability of this happening. This would further increase the application of the 30 percent band which, as noted, already applied to the majority of employees in 1990.3 11. The potential tax revenue obtainable through taxation of the family allowance can be roughly estimated as follows. If, at the one extreme, the total family allowances in 1990 (Ft. 63 billion) were received by fathers and taxed with their income, the vast majority would probably be subject to the 30 percent marginal rate (even for men whose other income placed them in the 15 percent tax rate). Thus (Ft. 63 x .30) or Ft. 19 billion might be the high end of the range of resulting revenue. If, on the other hand, mothers received the allowance and, according to the 1990 earnings distribution, it is assumed that one-third of them would pay the 15 percent marginal rate and two- thirds the 30 percent marginal rate, the potential revenue would be Ft. 15 billion. Since more women are nonparticipants in the labor force than men, this lower bound would be somewhat overestimated as it assumes that all mothers have some earned income. (Note, however, that women on maternity allowance or child care fee, who are not counted in the labor force, are subject to income tax on these benefits.) 3It might be thought that the answer to this problem would be to tax family allowance in any month at the same marginal rate as main earnings (or unemployment benefit). However, this would produce a most undesirable poverty trap in which an individual's total income could fall if an increase in earnings resulted in a movement into a higher rate band. 171 ANNEX VI Page 7 of 7 Table VI.4: The Distribution of 1988 Marginal PIT Tax Rates on Earnings of Persons in Full-Time Jobs Tax Bracket (Ft) Marginal Tax % Men % Women Monthly Earnings Rate (%) Paying Paying up to 5,000 0 7.8 24.3 5,000 - 6,833 20 17.7 29.6 6,833 - 8,500 25 20.2 18.7 8,500 - 11,000 30 23.4 14.6 11,000 - 13,500 35 13.7 6.9 13,500 - 16,000 39 6.3 2.6 16,000 - 21,000 44 6.5 2.3 21,000 and above 48 to 60 4.4 1.1 Total 100.0 100.0 Note: The additional monthly Ft 1,000 tax allowance for persons in employment has been applied. The calculations take no account of the third-child tax allowance and refer only to the marginal tax rates applicable to the full-time job. Source: CSO Statistical Yearbook. 1988, Table 4.9 (linear interpolation within ranges), and Ministry of Finance (1987) p.36. Table VI.5: The Distribution of 1990 Marginal PIT Tax Rates on Earnings of Persons in Full-Time Jobs Tax Bracket (Ft) Marginal Tax % Men % Women Monthly Earnings Rate (%) Paying Paying up to 5,583 0 2.8 5.2 5,583 - 8,500 15 15.6 30.1 8,500 - 26,000 30 71.0 60.0 26,000 - 42,666 40 7.7 3.6 over 42,666 50 2.9 1.1 100.0 100.0 Note: See Note above. Source: Tax schedule supplied by Ministry of Finance. Earnings distribution from 1990 Income Survey (linear interpolation within ranges). 172 ANNEX VII Page 1 of 4 Maternity and Child Care Support 1. Comparison of benefits in Hungary and in OECD Countries. It has been noted that the practice in European countries of allowing extended periods of leave after the birth of a child was pioneered in Hungary (ILO, 1989). However, schemes in Western Europe, and throughout the OECD area, are markedly less generous than the present Hungarian system. This is brought out by Table VII.l, which distinguishes between maternity and parental leave (the latter usually following immediately after maternity leave with either parent being eligible). 2. Most OECD countries average around 12-20 weeks of paid earnings- related maternity leave (over the whole pre- and post-natal period). This is often at less than full earnings, although account needs to be taken of the base period used in the benefit assessment and any indexing provisions, which are not shown in the table; 90 percent of the last monthly wage may provide higher benefit than 100 percent of the previous year's monthly average. This may be compared with the 24 weeks at 100 percent of earnings in Hungary. A number of OECD countries then have no further parental leave and where benefit is paid it is usually a flat-rate payment, as was the situation in Hungary prior to the introduction of the earnings-related child care fee in 1985. The table also underlines the important distinction between paid versus unpaid leave. In a number of countries, the mother may have the right to a period of unpaid leave after which she may re-enter her previous job; this may be seen as a part of employment protection legislation. The payment of benefits during leave may be seen as a related but separate issue. 173 ANNEX VII Page 2 of 4 Table VII.1: Maternity and Parental Leave in OECD Countries Maternity Leave Parental Leave Maximum Gross Maximum Gross Duration Replacement Duration Replacement (weeks) Rate (X) (weeks) Rate (%) Australia 52 No Benefit - Austria 16 100 Up to age 1 flat-rate Belgium 14 80-100 Canada 18 60 Denmark 28 90 10 flat-rate Finland 18 80 28 (then unpaid 80 up to age 3) France 16-28 84 Up to age 3 flat-rate (3rd child) Germany 14 100 up to 18 months flat-rate Greece 14 100 up to 30 months no benefit Iceland 13 flat-rate Ireland 14 60 Italy 20 80 Up to age 3 no benefit Japan 14 60 Netherlands 16 100 parents may 6 months work part-time New Zealand 14 no benefit - Norway 28 100 Up to age 1 flat-rate Portugal 13 100 Up to age 3 flat-rate Spain 16 75 3 years no benefit Sweden 12 90 up to 18 months 90 then flat-rate UK 18 90 then flat-rate - US Source: OECD, Employment Outlook, 1990, Table 5.8 (detailed footnotes in original source). 174 ANNEX VII Page 3 of 4 Table VII.2 Child Care Allowance and Fee: Recipients and Expenditure 1987 1988 1989 1990 Number of Persons Receiving Child Care Allowance (flat-rate) 113,360 80,780 86,920 na Number of Persons Receiving Child Care Fee (earnings-related) 117,350 159,300 157,760 na Expenditure on Child Care Allowance 2.0 1.9 2.7 3.2 (Ft. billion) Expenditure on Child Care Fee 4.1 7.1 8.3 10.7 (Ft. billion) Expenditure on Child Care Fee as share of Total Expenditure on Fee and Allowance (%) 67.2 78.9 75.5 77.0 Expenditure on Child Care Allowance and Child Care Fee as % of GDP 0.50 0.64 0.64 0.68 Source: Table 2.2 and CSO, Statistical Yearbook, 1989, Table 20.7. 175 ANNEX VII Page 4 of 4 Table VIII.3: Nursery Usage (children aged 0-3) 1987 1988 1989 Children Enrolled 51,788 44,362 42,870 Children Enrolled as % of Children of Nursery Age 13.7 11.9 11.7 Average Number of Children Enrolled as % of Available Places 62.0 60.6 60.2 Source: CSO Statistical Yearbooks, 1988 and 1989, Table 21.19 Table VII.4 Kindergarten Usage (children aged 3-6) 1987 1988 1989 Children Attending 398,325 393,735 392,273 Children Attending as % of those Entitled 86.8 86.0 85.7 Kindergarten Pupils as % 100 Kindergarten Places 97.7 97.8 100.4, Source: CSO, Statistical Yearbooks, 1988 and 1989, Table 24.1 176 ANNEX VIII Page 1 of 5 Social Assistance and Social Services The System of Aid for Adults 1. Regular welfare aid may be granted to persons over age 18 whose income does not reach the maximum amount that can be given for such aid (equivalent to the minimum widow's pension, Ft. 4400/month in December 1990), or who do not dispose of any income, have no relatives able and obliged to support them and have lost at least 67 percent of their capacity for work, or who are 10 years older than the retirement age. Persons without any income may receive the full sum of such aid; those with some income may receive the difference between their income and the set maximum. Supplementary aid may be provided for dependents of the individual determined to be eligible for the regular welfare aid. The local council determines the entitlement to regular welfare aid and makes the payment. 2. Irregular or emergency social aid may be paid to persons over age 18 whose livelihood is threatened and who experience a temporary financial distress which cannot be alleviated by other means. This type of aid may be awarded to an individual a maximum of six times in a year and the sum on each occasion may not exceed the upper limit of the sum that can be paid in regular welfare aid, except in exceptional cases as determined by the chairman of the local council. The awards are set and paid by the local councils. 3. For the following types of allowance, entitlement is determined by the Councils but payment is effected by the SIF: regular social aid for the blind, personal allowance for the blind, central social aid and cash allowance for war invalids and their dependents, family aid for national servicemen. System of child and youth welfare aid 4. Under Decree no. 51/1986.(XI.26.) MT, two different forms of aid may by applied within the framework of protective and welfare measures by the Court of Guardians: a) In cases where the parent is temporarily unable to provide for the child (food, clothing,-school equipment, nursery fee,, kindergarten fee, etc.), emergency aid may be awarded. The amount of the aid is defined by central government decrees from 1987 and 1990, which specify the minimum amount of aid (equivalent to half of the monthly family allowance payable to a single parent per child for two children), and the maximum (payment on each occasion may not exceed three times this minimum sum, and the amount paid in a calendar year may not normally exceed ten times that sum). The Court is not required to call the parent to account for the use of the aid. In 1989, Ft. 0.6 bn. was paid out nationally for such emergency aid to about 350 thousand children (average payment Ft. 1785). 177 ANNEX VIII Page 2 of 5 b) The Decree cited above also provides for regular child-raising aid. The conditions for this aid are (i) that the per capita income calculated on the basis of the income of the parents who are otherwise fit to raise the child, and of other dependents in the family whom they are obliged to maintain, is, through causes beyond the family's control, less than the prevailing minimum pension paid on the recipient's own right; and (ii) the development of the minor is threatened by these financial problems. If the per capita income exceeds the minimum pension by up to 20 percent but the parent is a single parent, is raising three or more children or a physically handicapped child, the aid may be awarded on an exceptional basis. The regulation sets only the minimum level of aid, as one-half of the monthly family allowance per child. The Court of Guardians reexamines the entitlement every three years, and the parent or guardian is obliged to report any substantial change in income to the Court within 15 days. In the case of such change in circumstances, the amount of regular child-raising aid may be modified or the aid terminated. In 1989, Ft. 0.6 bn. was paid for such assistance to 80 thousand minors (average payment Ft. 666/month). Table VIII.1 Social Assistance: Expenditure and Numbers in Receipt Year Stock of Number of Annual Expenditure Average Per Capita ---- Recipients of Exceptional on Regular and Regular Monthly Regular Support Payments Exceptional Support Payment as % of Ft. bn % of GDP Subsistence Min. 1985 48,671 446,235 1.8 0.14 60.8 1986 48,703 485,480 1.5 0.14 62.3 1987 48,070 563,952 1.8 0.15 63.8 1988 47,002 545,485 2.1 0.15 64.7 1989 47,406 781,122 3.0 0.17 63.4 1990 48,000 700,000* 3.6 0.18 *(420,000 persons) Recipients of Cases of Regular Child- Emergency Aid Raising Aid for Children 1989 79,725 348,781 1.3 0.08 (# families 36,660 153,600) Note: The subsistence minimum level used in column 4 is the overall figure for both active and inactive households Source: CSO, Statistical Yearbook, 1989, Tables 19.16 and 20.12 and information provided by Ministry of Welfare. 178 ANNEX VIII Page 3 of 5 Table VIII.2 Social Services: Recipients of Selected Programmes Year Recipients of Recipients of Persons Members of Free or Fuel Assistance Receiving Clubs for Subsidized Domestic the Aged Meals Care 1985 32,321 24,729 49,834 27,608 1986 39,448 27,925 56,065 29,064 1987 50,772 26,733 64,104 30,417 1988 64,192 29,075 68,102 33,464 1989 82,606 33,083 82,225 35,010 Source: CSO, Statistical Yearbook, 1989, Tables 20.11 and 20.12 Table VIII.3 Households and Families in 1984 1. Not Family-type Households Number ('000.s) 1.1 Single Persons 759.4 1.2 Several Persons 131.9 2. Households with families 2.1 Nuclear families, possibly 2222.8 including grown-up children 2.2 Nuclear families plus relatives 251.8 or other unrelated persons 2.3 One-Parent nuclear families, 309.4 possibly including grown-up children 2.4 One-Parent families plus relatives or other unrelated persons 41.2 2.5 Two families 111.2 2.6 Three or more families 3.5 TOTAL 3831.2 Note: The figure in the text for the minimum proportion of multi-family households is calculated as lines 1.2, 2.2, 2.4, 2.5 and 2.6 as a percentage of the total. Source: Information supplied by CSO, based on 2% Microcensus. 179 ANNEX VIII Page 4 of 5 Table VIII.4 Subsistence Minima in 1989 (Per Capita Monthly Amounts) Forints Households with active earners 3,940 One parent with one child 4,310 Married couple without child 4,540 in towns 5,080 in villages 4,190 Married couple with one child 4,230 in towns 4,540 in villages 3,800 Married couple with two children 3,760 in towns 4,120 in villages 3,310 Married couple with three children 3,310 Households with no active earners 3,640 Single persons 4,100 in towns 4,330 in villages 3,600 Married couples 3,460 in towns 3,760 in villages 3,210 Active and Inactive Households together 3,890 Source: CSO, Statistical Pocket Book, 1989, p.39. 180 ANNEX VIII Page 5 of 5 Table VIII.5 Schedule of CoRayment for Social Services for the Elderly. 1990 Pension Ranges (Ft. per month) Over less than 4600- 5100- 5600- 6100- 6600 and 70 years old: 4600 5100 5600 6100 6600 over Less than less than 4849- 5350- 5851- 6350- 6851 and 70 years old: 4848 5349 5850 6349 6850 over Share of indicative service charge imposed (percent) 0 20 40 60 80 100-200 Indicative Service Charges in 1990: for Full time Professional Caretaker Ft. 40/hour for Volunteer Caretaker Ft. 20/hour for Meal Catering Ft. 23/meal for Residential Nursing Home Ft. 2070/month Source: Sliding scale for service charges is set by Ministry of Welfare; service charges are set by Ministry of Finance. 181 ANNEX IX Page 1 of 2 The Financing of Local Government 1. In 1990, Hungary enacted a major decentralization of government with the passage of the Act on Local Government (Act No. LXV of 1990). This act restored autonomy to the local self governments, giving them legal identity and abolishing the hierarchical relationship between them and higher levels of government (county and central). The self governments have been given wide- ranging functions, including responsibility for social services, primary and preprimary education, and basic health care. Ensuring minimum income support (social assistance) is expected to remain the responsibility of local governments as well (although not enumerated among these in the law), as it was before the act was passed. Following the local elections in September/October 1990, 3100 towns and villages ("settlements") affirmed themselves as self governing units headed by councils.' 2. As of 1990, budgetary grants from the central government to local councils are of two types: normative grants based on capitation formulas, and special purpose grants, i.e. for specific functions or investments. The State formula grants are capitation grants based mainly on age group composition of the local population, and enrollments in educational institutions and social institutions (e.g. residential homes for elderly and handicapped). In 1990, the councils were also given 100 percent of the (planned) PIT revenues collected two years previously (the 1988 returns); as of 1991, half of PIT revenues will automatically be returned to the councils and the rest of the revenues will be shared through the formula grant system. Up to 1990, the central government also provided ex post deficit financing of councils; this is to be discontinued for all practical purposes in 1991. 3. In 1990, the revenue of local councils was composed as follows: State formula grants (28%), State special purpose grants (5%), State deficit financing (7%), PIT revenues (27%), social insurance fund (for operation of health services) (16%), other local revenue (e.g. from operation of council- owned enterprises) (14%), and local taxes (3%). 4. The planned revenues for 1991 consist of: State formula grants (44%), State special purpose grants (5%), State deficit financing (1%), PIT revenues (14%), social insurance fund (17%), other local revenue (11%), local taxes (7%), loans (1%). I The local government reform law is described in detail in a background paper to this study ("Local Government Reform in Hungary" by Kenneth Davey, November 1990). 182 ANNEX IX Page 2 of 2 PROPOSED COMPONENTS OF THE FORMULA GRANT, 1991 1. Lump Sum Grant to each Town/Village FT 2 million (but only for population above 200; below 200 the grant is FT 10,000 per capita) 2. Per capita grant to Towns and Villages FT 3,500 per capita 3. Per capita grant to Counties/Capital City FT 1,200 per capita 4. Per capita grant for population of ages O to 18 and over 60 FT 2,850 per capita 5. Per capita grant for people in care: children in care FT 175,000 per capita homes for the elderly or mentally handicapped FT 150,000 per capita 6. Per capita grant for primary schools FT 31,000 per capita 7. Per capita grant for independent music school FT 20,000 per capita 8. Per capita grant for schools for the mentally handicapped FT 58,000 per capita 9. Per capita grant for high schools FT 45,000 per capita 10. Per capita grant for vocational schools FT 55,000 per capita 11. Per capita grant for boarding schools FT 54,000 per capita 12. Grant for theatres and open air performances FT 450,000 per capita NB. In the case of pupils undertaking primary, high or vocational school courses by correspondence or evening classes, one third of the full time per capita grant is payable. 183 ANNEX X Page 1 of 8 Derivation of Medium-Term Projections Social Expenditures 1. For the projection of pensions, health care and pharmaceuticals, sick pay, maternity and child care benefits, "other" social insurance expenditures, unemployment compensation, and family allowances, the basic methodology used is that described in IMF, Social Security Reform in Hungary, October 12, 1990, Appendix II. The "contribution base" for social insurance was estimated as the product of labor income/GDP (projected as a fixed 43 percent) and the contribution base/labor income (projected as a fixed 76 percent). 2. Pension. See also Annex Tables IV.2 and 3 of the present Report for the specific parameters used for the Baseline and Alternative scenarios. 3. Health Care. The parameters used for projecting health care expenditure in both the Base Case and Alternative Scenario are as follows: Year 1990 2000 2021 2051 Real Medical Benefits per capita (growth rate) 0.00 0.06 0.05 0.04 (mark-up to GDP/employed) 0.065 0.02 0.01 0.00 Real GDP per employed growth rate -0.065 0.04 0.04 0.04 Medical expenditure/GDP elasticity 1.40 1.03 1.22 1.13 (The values of medical expenditure parameters for intermediate years derived by linear interpolation.) 4. The health care expenditures under the SIF were projected as the 1989 per capita expenditure times the population aged 60 and below (weighted by a factor of 1) and aged over 60 (weighted by a factor of 4). The weighted per capita expenditure was assumed to move in parallel with the GDP growth rate indicated above, plus the expenditure mark-up, and was inflated by the CPI. Expenditures on pharmaceuticals were weighted by a factor of four for the population aged over 55 and adjusted by a compound medical price index. Private health expenditures were projected as a fixed 23 percent of the government health and pharmaceutical expenditures based on estimates from the 1989 Incidence Study. 5. Sick Pay. Annual expenditure on sick pay calculated as fixed percentage (4.5 percent) of the product of the number of employees times the monthly contribution base of the previous period. For the Alternative Scenario, one-fourth of this projected expenditure was included in the SIF from 1992 on (intended to represent only episodes of sick leave longer than 30 days) and the remaining share of projected expenditure was included among employers' total labor costs. 6. Maternity and Child Care. Benefits calculated as a fixed proportion of the contribution base times the projected 0-4 year old age group; For the Alternative Scenario, this projection was reduced by one- 184 ANNEX X Paze 2 of 8 third from 1992 on; this assumption represents the estimated impact of the recommendations to shift from earnings-related to flat-rate child care fee, and eliminate the third year of paid child care leave. For this scenario, this expenditure item was assumed to be transferred out of social insurance to the general budget. 7. Other Social Insurance Expenditure. Representing administrative and other costs, projected as a fixed 1.7 percent of total social insurance expenditure in each year. 8. Family Allowance. Although family allowance was paid from the SIF during the first three months of 1990, the model shows this benefit expenditure for 1990-on outside the SIF. To project family allowances, the implicit per capita expenditure in 1989 was multiplied by the numbers in the 0-15 age group and the CPI inflation index in each year. For the Alternative Scenario, this projection was reduced by 21 percent from 1992 on, which represents the magnitude of cost savings from elimination of the one-child allowance in two-parent families. 9. Social Insurance Revenues. "Other" revenues projected as a fixed proportion (0.7 percent) of the contribution revenue. The implicit social insurance contribution rate in each projection year (1992 on) was calculated as the expenditures of the SIF divided by the contribution base (in other words, the rate which would exactly balance the SIF). For the Base Case, the employees' contribution to the SIF and to the Solidarity Fund was calculated as 10 percent and 1 percent, respectively, of employee Main Earnings; for the Alternative Scenario, employees' SIF contribution was raised gradually to 15 percent of Main Earnings by 1996. Employers' contribution was calculated as the difference between total SIF expenditure and the sum of employees' contribution and Other Revenues. Thus, the total implicit contribution rate is higher than the sum of employees' and employers' rates by the amount of Other Revenues. 10. Unemployment/Employment Benefits. For 1989-90, both cash benefits and in-kind benefits were covered by the Employment Fund. From 1991, cash benefits (unemployment insurance as well as benefits to career starters and older workers) are included under the Solidarity Fund. To project cash benefits from 1992, it was assumed that the average unemployed worker receives a replacement rate of two-thirds of the contribution base. The number of unemployed was determined by the assumed unemployment rate and the estimated number of employees, with the further assumption that 75 percent of the unemployed stock in any year obtain unemployment benefits. (As discussed in Annex V, the rate of coverage of the unemployed stock is more likely to decline somewhat each year during periods of growing unemployment.) From 1991, the Employment Fund is assumed to cover only benefits in kind (the "proactive" element). The Base Case assumes that these are equivalent to half of the cash benefits in each year; the Alternative Scenario assumes that this share remains at 75 percent. All cash benefits are assumed to be financed by unemployment insurance contributions. 185 ANNEX X Page 3 of 8 11. Scholarships. Assumed to grow in line with inflation from 1992. 12. Social Assistance (in cash). Figure for 1990 is a residual of total social income in cash, therefore includes some unspecified expenditure other than social assistance. Estimate for 1991 based on calculations in Chapter 8 (assumed expenditure equivalent to 1 percent of CDP); for later years, assumed to increase in parallel with Solidarity Fund expenditures. 13. Education. Projected so as to maintain a constant share of GDP from 1990-2000. Since the population aged 5-24 is projected to decrease by 0.7 percent each year of the period, the total education expenditure would therefore grow by about 2.6 percent per capita per year, allowing for improvements in enrollment coverage and quality. Private education expenditures were projected as a fixed 21 percent of government education expenditures based on evidence from 1989 Incidence Study. 14. Social Services in Kind. Projected so as to maintain a constant share of GDP from 1990-2000. It can be expected that private contributions would increasingly support social services, but the private financing is not projected here. 1S. Other Social Income in Kind. These items (mainly recreational and cultural activities, as well as depreciation of State rental housing) are assumed to be phased out of public social expenditure by 1994. 16. Housing Subsidies. This is projected in two components: the interest subsidy on the outstanding balance of pre-1989 mortgage loans is assumed to be cut by about one-third in 1992 and thereafter remain at the 1992 nominal level (phased out in real terms); and subsidies to new homeowners are assumed to remain constant at the real 1991 level (no volume increase). No rental subsidies are assumed after 1990, and any future housing allowance scheme is assumed to be self-financing (financed by internal cross-subsidies among renters). Private expenditure on housing and utilities was projected to increase gradually from 13 percent of gross earnings (17 percent of net earnings) in 1989 (based on expenditure survey results reported in Incidence Study, 1989) to a target level of 20 percent by 1998. 17. Consumer Subsidies (including utilities). These are assumed to be phased out evenly by 1993. Household Income Aggregates 18. Earnings. Based on the 1990 Incidence Study, household income from 1989 was broken down into Gross Earnings (sum of Main Earnings, Secondary Labor Income, Labor Income in Kind, Other (nonlabor) Income), from which are deducted Personal Income Taxation, Other Taxes and Duties, and Social Insurance premium (including unemployment insurance premium). The sum of Net Earnings thus derived and Social Income in Cash and Kind equals Net Personal Income, which is conceptually the same as Disposable Income (although the base year (1989) value of Net Personal Income used in the projection model is slightly higher than that for Disposable Income in the national accounts). In 186 ANNEX X Page 4 of 8 both the Base Case and Alternative Scenarios, the income components were projected from 1992 to move in parallel with the assumed growth of Real Disposable Income, with the following adjustments: annual growth of Main Earnings is 0.5 percent higher, Secondary Labor Income is 1.0 percent higher, Labor Income in kind is 3 percent lower, and Other Income is 4 percent higher. The values for 1990-91 are based on estimates from government officials in November 1990. In estimating the costs of labor to enterprises, Gross Earnings less Other Income are taken as the total payments to labor. 19. Taxes. Personal Income Tax is projected to remain 12 percent of the estimated sum of Gross Earnings, based on 1989 actual receipts. Other Taxes and Duties are projected as a constant 25 percent of personal income tax receipts. 187 Table X.1 Base Case - Projection of Social Expenditures 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 SOCIAL POLICY EXPENDITURES AS PERCENT OF GDP Expenditure by Govermnent Pensions 9.1 9.9 10.0 10.0 9.8 9.4 9.1 9.0 9.1 9.2 9.4 9.7 Sick Pay 1.2 1.2 1.1 1.1 1.2 1.2 1.2 1.2 1.2 1.3 1.3 1.3 Health 3.0 3.3 3.4 3.5 3.5 3.4 3.3 3.3 3.3 3.4 3.4 3.6 Pharmaceuticals 1.1 1.3 1.3 1.4 1.4 1.3 1.3 1.3 1.3 1.3 1.4 1.4 Matemity & Childrare 0.9 0.9 0.9 1.0 1.0 1.0 1.0 1.0 1.1 1.1 1.1 1.2 Other Social Insurance Expenditure 0.3 0.3 0.3 0.3 0.3 0.3 0.3 0.3 0.3 0.3 0.3 0.3 Family Allowances 3.1 3.1 3.2 3.2 3.1 2.9 2.7 2.6 2.5 2.4 2.4 2.4 Solidarity Fund 0.1 0.3 0.7 1.8 2.2 1.8 1.4 1.0 0.9 0.7 0.5 0.5 Enployment Fund 0.1 0.2 0.5 0.9 1.1 0.9 0.7 0.5 0.4 0.3 0.3 0.3 Scholarships 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 Social Assistance (cash) 0.5 0.9 1.0 2.7 3.3 2.7 2.1 1.5 1.3 1.0 0.7 0.7 Education 4.4 4.6 4.6 4.6 4.6 4.6 4.6 4.6 4.6 4.6 4.6 4.6 Social Services in Kind 0.4 0.5 0.5 0.5 0.5 0.5 0.4 0.4 0.4 0.4 0.4 0.4 Other Social Inccne in Kind 1.1 1.1 0.4 0.1 0.1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Total Social Inrone Cash & Kind 25.4 27.5 27.9 31.3 32.1 30.1 28.4 27.0 26.5 26.2 25.9 26.4 Housirg Subsidies 4.3 3.9 2.9 2.0 1.8 1.6 1.4 1.3 1.2 1.1 1.1 1.0 Consumer Subsidies 2.9 2.0 1.2 0.6 0.3 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Total Social Expenditure 32.6 33.3 32.0 33.9 34.2 31.7 29.9 28.3 27.7 27.3 27.0 27.4 as % of General Govt. Expenditure 53.6 57.4 55.1 60.0 62.2 59.3 57.4 56.1 56.6 57.6 58.6 61.0 % of Total Social Expenditure financed by: Social Insurance Contributions 52.2 52.0 56.3 51.8 51.3 55.3 58.8 61.9 63.3 64.2 65.1 64.0 Other General Goverraient Revenues 47.8 48.0 43.7 48.2 48.7 44.7 41.2 38.1 36.7 35.8 34.9 ,36.0 Social-Type Expenditure by Households Savings (Net Financial) (as % of GDP) 0.1 4.2 4.0 4.0 4.0 3.9 4.1 4.3 4.5 4.7 4.9 5.0 Health Expenditure (as % of Govt. Health Exp.) 23.2 23.0 23.0 23.0 23.0 23.0 23.0 23.0 23.0 23.0 23.0 23.0 Education Expenditure (as X of Govt. Education Exp.) 21.2 21.0 21.0 21.0 21.0 21.0 21.0 2180 21.0 21.0 21.0 21.0 Housing and Utilities (as X of Gross Earnings) 13.0 13.0 15.0 16.0 16.0 17.0 17.0 18.0 18.0 20.0 20.0 20.0 Govt. & Private Social Expenditures as % of GDP Pensions 9.2 14.1 14.0 14.0 13.8 13.3 13.2 13.3 13.6 13.9 14.3 14.7 Health & Pharmaceuticals 5.1 5.6 5.8 6.0 6.0 5.9 5.7 5.7 5.7 5.8 5.9 6.1 Education & Scholarships 5.4 5.7 5.7 5.7 5.7 5.7 5.7 5.7 5.7 5.7 5.7 5.7 Housing & Utilities 13.7 12.5 12.1 10.9 10.1 10.3 10.1 10.5 10.4 11.4 11.4 11.4 a299.n353 188 Table X.2: Base Case - Projection of Household Income Aggregates 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 lmplicit Contribution Rates for Social Insurance (Projected Expenditure as X Contribution Base) 1/ Pensions 27.6 30.2 30.5 30.7 30.0 28.7 27.9 27.6 27.8 28.3 28.8 29.7 Sick Pay 3.8 3.6 3.5 3.5 3.5 3.6 3.7 3.7 3.8 3.9 3.9 4.0 Health -- 10.0 10.4 10.7 10.7 10.4 10.2 10.2 10.2 10.4 10.6 10.9 Phamaceuticals 3.5 3.9 4.1 4.2 4.2 4.1 4.1 4.0 4.0 4.1 4.2 4.3 Matemity & Child Care 2.8 2.8 2.8 2.9 3.0 3.1 3.1 3.2 3.2 3.3 3.4 3.6 Family Allowance 9.6 -- -- -- -- -- -- -- -- -- -- -- Other Soc. Insur. (adninistration) 1.3 0.8 0.9 0.9 0.9 0.8 0.8 0.8 0.8 0.8 0.9 0.9 Solidarity Fund -- -- 2.0 5.6 6.9 5.6 4.4 3.2 2.6 2.1 1.6 1.6 Total Contribution Rate 53.0 53.0 55.0 58.4 59.2 56.4 54.2 52.7 52.6 52.9 53.3 54.9 Components of Household lncome A. Social Income in Cash ard Kird 41.0 40.6 40.0 43.4 44.4 43.1 42.0 40.7 40.2 39.8 39.4 39.7 (as % of Net Personal Income) B. Gross Earnings 75.3 74.9 76.4 73.2 71.9 73.6 75.0 76.7 77.4 78.0 78.5 78.0 (as % of Net Personal Income) less Taxes and Duties, of Which: Personal Income Tax (as X of Gross Earnings) 12.0 11.5 12.4 12.2 12.2 12.2 12.2 12.2 12.2 12.2 12.2 12.2 Soc. Insur. Contrib. (esployee) (as % of Contribution Base) 10.0 10.0 10.0 11.5 11.4 11.3 11.1 11.2 11.2 11.3 11.3 11.4 Unemployment Insur. Levy (employee) (as % of Contribution Base) 0.0 0.0 0.5 0.6 0.6 0.6 0.6 0.6 0.6 0.6 0.6 0.6 C. Net Earnings 59.0 59.4 60.0 56.6 55.6 56.9 58.0 59.3 59.8 60.2 60.6 60.3 (as X of Net Personal Income) D. Net Personal Incom. (A + C) 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 (as % of GDP) 62.1 67.7 69.9 72.0 72.4 69.9 67.7 66.4 66.1 65.9 65.8 66.5 Cost of Labor to Enterprises (in terms of % of GDP) Gross Payments to Labor 47.1 49.3 51.2 50.5 49.8 49.1 48.4 48.5 48.6 48.8 48.9 49.0 (Main Earnings + SecoErary Labor Inccma + Labor Inccme in Kind) Social Insur. Contrib. (epLoyer) 14.0 14.0 14.1 13.7 13.7 13.8 13.8 13.8 13.8 13.8 13.7 13.7 (as % of Contribution Base) 43.0 43.0 43.0 41.8 41.9 42.1 42.2 42.2 42.1 42.1 42.0 42.0 LUnerploymnt Insur. Levy (e,ployer) 0.0 0.0 0.5 1.6 2.1 1.6 1.2 0.9 0.7 0.5 0.3 0.3 (as X of Contribution Base) 0.0 0.0 1.5 5.0 6.3 5.0 3.8 2.7 2.1 1.5 1.0 1.0 Sick Pay 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Total Cost 61.1 63.3 65.7 65.8 65.5 64.5 63.4 63.1 63.1 63.0 63.0 63.1 a355.n410 189 Table X.3 Alternative Scenario - Projection of Social Expenditures 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 SOCIAL POLICY EXPENDITURES AS PERCENT OF GDP Expenditure by Goverrnent Pensions 9.1 9.9 10.0 9.6 9.2 8.6 8.2 7.9 7.8 7.8 7.7 7.8 Sick Pay 1.2 1.2 1.1 0.3 0.3 0.3 0.3 0.3 0.3 0.3 0.3 0.3 Health 3.0 3.3 3.4 3.5 3.5 3.4 3.3 3.3 3.3 3.4 3.4 3.6 Pharmaceuticals 1.1 1.3 1.3 1.4 1.4 1.3 1.3 1.3 1.3 1.3 1.4 1.4 Maternity & Childcare 0.9 0.9 0.9 0.6 0.7 0.7 0.7 0.7 0.7 0.7 0.8 0.8 Other Social Insurance Expenditure 0.3 0.3 0.3 0.3 0.3 0.3 0.2 0.2 0.2 0.2 0.2 0.3 Family Allowances 3.1 3.1 3.2 2.5 2.4 2.3 2.1 2.0 2.0 1.9 1.9 1.9 Solidarity Furd 0.1 0.3 0.7 1.8 2.2 1.8 1.4 1.0 0.9 0.7 0.5 0.5 Enploynent Fund 0.1 0.2 0.5 1.3 1.6 1.3 1.0 0.7 0.6 0.5 0.4 0.4 SchoLarships 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 Social Assistance (cash) 0.5 0.9 1.0 2.7 3.3 2.7 2.1 1.5 1.3 1.0 0.7 0.7 Education 4.4 4.6 4.6 4.6 4.6 4.6 4.6 4.6 4.6 4.6 4.6 4.6 Social Services in Kind 0.4 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 Other SociaL Income in Kind 1.1 1.1 0.4 0.1 0.1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Total, Social Incone Cash & Kind 25.4 27.5 27.9 29.4 30.1 27.9 26.0 24.4 23.7 23.1 22.6 22.8 Housing Subsidies 4.3 3.9 2.9 2.0 1.8 1.6 1.4 1.3 1.2 1.1 1.1 1.0 Consumer Subsidies 2.9 2.0 1.2 0.6 0.3 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Total, Social Expenditure 32.6 33.3 32.0 32.0 32.2 29.5 27.4 25.7 24.9 24.2 23.6 23.9 as % of General Govt. Expenditure 53.6 57.4 55.1 56.6 58.5 55.2 52.8 50.9 50.7 51.0 51.3 53.0 X of Total Social Expenditure financed by: Social Insurance Contributions 52.2 52.0 56.3 46.9 45.4 47.1 48.9 51.0 52.3 53.9 55.6 56.1 Other General Government Revenues 47.8 48.0 43.7 53.1 54.6 52.9 51.1 49.0 47.7 46.1 44.4 43.9 Social-Type Expenditure by Households Savings (Net Financial) (as % of GDP) 0.1 4.2 4.0 4.0 4.0 3.9 4.1 4.3 4.5 4.7 4.9 5.0 Health Expenditure (as % of Govt. Health Exp.) 23.2 23.0 23.0 23.0 23.0 23.0 23.0 23.0 23.0 23.0 23.0 23.0 Education Expenditure (as % of Govt. Edication Exp.) 21.2 21.0 21.0 21.0 21.0 21.0 21.0 21.0 21.0 21.0 21.0 21.0 Housing and Utilities (as X of Gross Earnings) 13.0 13.0 15.0 16.0 16.0 17.0 17.0 18.0 18.0 20.0 20.0 20.0 Govt. & Private Social Experditures as X of GDP Pensions 9.2 14.1 14.0 13.6 13.2 12.5 12.3 12.2 12.3 12.5 12.6 12.8 Health & Pharnaceuticals 5.1 5.6 5.8 6.0 6.0 5.9 5.7 5.7 5.7 5.8 5.9 6.1 Education & Scholarships 5.4 5.7 5.7 5.7 5.7 5.7 5.7 5.7 5.7 5.7 5.7 5.7 Housing & Utilities 13.7 12.5 12.1 10.9 10.1 10.3 10.1 10.5 10.4 11.4 11.4 11.4 a299.n353 190 Table X.4 Alternative Scenario - Projection of Household Income Aggregates 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 Implicit Contribution Rates for Social Insurance (Projected Expenditure as % Contribution Base) 1/ Pensions 27.6 30.2 30.5 29.3 28.1 26.3 25.1 24.2 23.9 23.8 23.7 24.0 Sick Pay 3.8 3.6 3.5 0.9 0.9 0.9 0.9 0.9 0.9 1.0 1.0 1.0 Health -- 10.0 10.4 10.7 10.7 10.4 10.2 10.2 10.2 10.4 10.6 10.9 PharmaceuticaLs 3.5 3.9 4.1 4.2 4.2 4.1 4.1 4.0 4.0 4.1 4.2 4.3 Maternity & Child Care 2.8 2.8 2.8 -- -- -- -- -- -- -- -- Family Allowance 9.6 -- -- -- -- -- -- -- -- -- -- Other Soc. Insur. (adninistration) 1.3 0.8 0.9 0.8 0.8 0.8 0.7 0.7 0.7 0.7 0.8 0.8 Solidarity Furd -- -- 2.0 5.6 6.9 5.6 4.4 3.2 2.6 2.1 1.6 1.6 Total Contribution Rate 53.0 53.0 55.0 51.5 51.6 48.2 45.4 43.3 42.5 42.1 41.7 42.5 Compcnents of Household Inccme A. Social Income in Cash and Kind 41.0 40.6 40.0 42.1 43.3 41.9 40.8 39.4 38.5 37.9 37.2 37.4 (as X of Net Personal Income) B. Gross Earnings 75.3 74.9 76.4 75.6 74.8 77.3 79.6 82.3 83.4 84.3 85.2 84.9 (as % of Net Personal Income) less Taxes and Duties, of which: Personal Incame Tax (as % of Gross Earnings) 12.0 11.5 12.4 12.2 12.2 12.2 12.2 12.2 12.2 12.2 12.2 12.2 Soc. Insur. Contrib. (efployee) (as % of Contribution Base) 10.0 10.0 10.0 12.7 13.7 14.6 15.6 16.7 16.8 16.9 17.0 17.0 Una1mplo)rent Irsur. levy (eployee) (as % of Contribution Base) 0.0 0.0 0.5 0.6 0.6 0.6 0.6 0.6 0.6 0.6 0.6 0.6 C. Net Earnings 59.0 59.4 60.0 57.9 56.7 58.1 59.2 60.6 61.5 62.1 62.8 62.6 (as % of Net Personal Income) 0. Net Personal Income (A + C) 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 (as % of GDP) 62.1 67.7 69.9 69.8 69.7 66.6 63.8 61.9 61.4 61.0 60.6 61.1 Cost of Labor to Enterprises (in terms of % of GDP) Gross Parments to Labor 47.1 49.3 51.2 50.5 49.8 49.1 48.4 48.5 48.6 48.8 48.9 49.0 (Main Earnings + Secondary Labor Income + Labor Income in Kind) Social Insur. Contrib. (effployer) 14.0 14.0 14.1 10.7 10.0 9.0 8.2 7.5 7.4 7.4 7.5 7.7 (as % of Contribution Base) 43.0 43.0 43.0 32.9 30.7 27.6 25.1 23.0 22.7 22.8 22.9 23.6 Uneiptomyment Insur. Levy (employer) 0.0 0.0 0.5 1.6 2.1 1.6 1.2 0.9 0.7 0.5 0.3 0.3 (as % of Contribution Base) 0.0 0.0 1.5 5.0 6.3 5.0 3.8 2.7 2.1 1.5 1.0 1.0 Sick Pay 0.0 0.0 0.0 0.9 0.9 0.9 0.9 0.9 0.9 0.9 1.0 1.0 Total Cost 61.1 63.3 65.7 63.7 62.7 60.6 58.7 57.8 57.7 57.7 57.7 58.1 a355.n410 191 GLOSSARY OF TERMS accrual rate - the percentage point increase of the pension replacement rate with each additional year of insured employment. block grant - budgetary transfer to an authority (e.g., from central to local government) which is earmarked for certain purposes. contribution rate - in this study, refers to the social insurance tax paid by employers (or employees) as a percentage of the contribution base, which is the wage bill (earnings). The implicit contribution rate is the tax rate that would be necessary to cover estimated social insurance expenditures. contributory benefits - benefits received by virtue of the recipient having made earnings-related payments or contributions into a social insurance fund (see social insurance). coRayment - (or user fee) partial payment by the user or consumer of a good or service which is also financed in part by other (usually public) funds. degressivitv - in Hungary, the declining-scale formula used to determine the share of the individual's prior earnings base to be taken into account in calculation of the pension. earnings-related - benefit calculated as a percentage of previous earnings. effective retirement age - the average age of retirement according to actual practice; reflects the individual's response to regulations regarding the legal minimum retirement age and incentives to remain in the labor force. flat-rate - a payment (or tax) which is specified in absolute amount, rather than as a percentage of some other value such as previous earnings. in kind - benefit or income received in the form of goods or services (e.g. health care, food supplies or food vouchers, etc.). main employment or earnings - in Hungary, defined as the primary job and source of income, identified as such in the individual's "work book" (employment record maintained for social insurance purposes). means-testing- - determination of individual's eligibility for benefits which are provided only when income, and sometimes also assets, fall below a certain ceiling (also called income-testing). minimum retirement age - the minimum legal age of eligibility to receive a public pension. net personal income - the sum of all gross income, both earned (received either in cash or in kind from employment or from investments) and unearned (social income in cash), net of taxes and other withdrawals and net of transfers among households. normative - based on objective, formally-specified criteria or rules, e.g. normative formulas for central government subsidies. 192 Rrimarv income of active earners - in Hungary, defined as main earnings and sick pay. primary income of inactive earners - in Hungary, defined as pensions, child care fee and allowance, and unemployment compensation. progressive - structure of policies whereby the percentage rate of transfers received (or taxes paid) increases as income increases. regressive - structure of policies whereby the percentage rate of transfers (or taxes) is inversely related to income levels, e.g., the percentage of subsidy received (or tax paid) is lower as income rises. replacement rate - the ratio of benefit to income when in work. Distinctions may be made between the gross (gross benefit/gross work income) or net reRlacement rate (net benefit/net work income); or between statutory (nominal) replacement rate (the ratio as declared in the benefit regulations) and the effective replacement rate, which is the actual ratio in practice considering the effects of taxation, inflation, etc. retirement income testing - for persons otherwise eligible for a public pension, their right to receive a given amount of pension at a given time is subject to a determination of how much employment income they continue to receive. secondary (or supplementary) labor income - in Hungary, defined as earnings which are from other than "main employment". secondary social incomes - in Hungary, defined as family allowance, student scholarships, and other cash benefits such as social assistance. service period - periods counted towards eligibility for social insurance benefits. Normally, periods of service should be the same as periods of employment during which social insurance contributions are paid; however, some periods of noncontribution (e.g. time spent in the military) may be counted as service periods. sickness benefit (or sick pay) - income replacement benefit to individuals during periods of temporary sickness or injury. social assistance - cash benefits usually financed out of general taxation and paid to needy individuals on the basis of a means test (sometimes called welfare or public assistance). social expenditure - in the present study, defined as the sum of social income plus consumer and housing subsidies. (In the OECD data base, social expenditure is defined more narrowly as the sum of public old-age, survivors', and disability pensions, unemployment compensation, sick pay, family and maternity/child care benefits, social assistance and social/welfare services, health services and pharmaceuticals, and education services. Except for the omission of cultural/sports services and unemployment services, this classification is roughly equivalent to "social income" in Hungarian terms.) 193 social income - in Hungary, definied as the sum of benefits paid to individuals in cash (old age, survivors', and disability pensions, unemployment compensation, sick pay, family allowances, maternity/child care benefits, social assistance, and student scholarships) as well as the value of services consumed by individuals in kind (health services and pharmaceutical subsidies, education services, social services, subsidized cultural and sports activities, and services for the unemployed). social insurance - cash benefits awarded on the basis of past contributions; by definition, social insurance benefits should be actuarially-based (taking account of risk factors) and contributions, by employers and/or employees, should be mandatory. standard retirement age - the age at which full pension benefits can be received. subsistence minimum income - income level for individual/family estimated to be required to maintain a level of consumption sufficient for a healthy existence. This estimated income level represents the poverty line. A social minimum income level is also calculated to represent a somewhat higher standard of consumption. targeting - the awarding of social benefits to only certain groups in the population, or only on certain prespecified criteria. Means-testing is one method of targeting. tax allowance - benefit provided through a specified reduction in the taxable income base. tax credit - benefit provided through a specified reduction in the amount of tax payable. A refundable tax credit may be converted to a cash payment to individuals whose initial tax liability is less than the amount of the tax credit. tax expenditure - a benefit defined in terms of a tax concession, e.g. a tax allowance or tax credit. The cost of such benefit to government is the revenue foregone. transfer - refers to any public expenditure outlay for the benefit of individuals. universal benefit - a benefit delivered without ex ante restrictions on eligibility. There are few purely universal benefits. E.g., a family allowance is both targeted (only to families), but also universal in that all families receive it on the basis of residence (or citizenship) without preconditions regarding income, social insurance status, etc. A universal benefit may still be income-tested ex post by being subject to a progressive personal income tax. Distributors of World Bank Publications ARGENTINA FINLAND MEXICO SOUTH AFRICA. BOTSWANA C.io- Hiscb. SRL Aktin Kkarpps INFOTEC Foras9s tJbL- Gar1e-i Gure rr P.O. Box 128 Aparsdo Po.la 22-560 Oxfod Uni-endty Pre Floridd 165, 4th Flor-Ok. 453/465 SF-00101 14060 ThaIpas MdicoD.F. Suthoera Afrii 1333 Bueoeo Aiue Heleinki 10 P.O. Box 1141 MOROCCO CopeTomn 8X0 AUSTRALIA, PAPUA NEW GUINEA, FRANCE Sxifte d'Erudor Ma3kding Moro e FIJL SOLOMON ISLANDS, Word B.&k Pibrhao 12 rueMazart,Bd. d'Asf. ForooMbR-ii oridrs VANUATU, AND WESTERN SAMOA 66, avenue d'lar Caroblb. fi-taar6asl Sbbeamipt-o Smrvice DA. Bok. & Journs 75116 Paris P.O. Box 41095 648Wh,ehorreRo-d NETHERLANDS C-gholl Mitdrn 3132 GERMANY hot-IPubhkh.sti b-. Joh.nmobug 2D24 Vid ris UNO.VfrIto P.Q Box 14 Popprdcrfe Aldee55 7240 BA LoDro SPAIN AUSTRIA D-53)D Blls I Mud.&-Pmerraa Ubrm4 SA. Gerndd sod Co. NEW ZEALAND Csottll 37 Ceabor 31 GREECE Hills libray and Ilzforotio. S.evace 2811I Madrid A-ll0 Wirn KEMB Pri-at Bag 24, Ippodamou Safed Plati Plarlir-s Ne-v Mar}et Lbred3 IrrERrsion AEDOS BAHRAIN Athbs-11635 Adxcklrrmd CoooeldeCenCL 391 B3hssin Rrtrrh aod Coneultnny 051B)9 Ba. ma Atroamtts Ld. GUATEMALA NIGERIA P.O. Box 2103 Lbiedi Piedr. Senln Univrrity Prora limitid SRI LANKA AND THE MALDIVES Moramo Tovv 317 Sa. CGae 7-55 ThreeCrowss Baildingjericao Lake Hl{oreBoolahop Zas I Pri-ate Mail Bag S1QS P.O. Box 244 BANGLADESH G-astoala City lb dan 100, Si, Chittloxpspala A. Mi nroldodsriODvelIqet GordineMorvath3 Aedoisr Soritty (MIDAS) HONG IONG, MACAO NORWAY Cadombo 2 Hose S, Road 16 Asia Lid. Narvertn Iolmod...oo C-nte Dh-mmoxoh R/A-e 48-4S Wyodham Street Book Depolromot SWEDEN DhskA 1209 Wuindog Coerler P.O. Bo" 6125 Ettorslad For atslc atha 2nd Foor N-0602 Orb 6 Fritoro FParbafo"etsget Rrd,Gff CraCtea Hrmg Kong Reg- agtl 12, Boo 16356 Mai Road OMAN S-10327StoWddtd Moidee Co*t INDIA MEMRBhoeerHioS- Nokhsli - 3l0 Allied Pbhe Private Ltd. P.O Box 161, Bee Airpot F-o -pri 751 Mosot Ro3d Moorat Wtnn -WiBiame AB 76, KDA. Avenue Md -.6000O2 Bo 3X0D4 KuPlrm PAISTAN 5-104 25 Stt dth.ol Bo,rd oJe Mirs Book Agcy BELGLJM IEjN.HeediaMag 65,hlalh-4aUid-Axm SWITZERLAND) J e3DeL3nooy Ballad Eatl P.O. Bo. No. 729 F-osiagtilk Av. du REd 202 Bombay - 40D 08 Lahore3 Libririe Payot 1060 Bserr 6, rue GToSrun 13/14AdaAli Road PERU Caepo-.le381 CANADA NevDelbi-l10112 Edtorial Deaollo SA CH 1211 Geva 11 LeDifhrx, Aptado 3824 C.P. 85, ISM reAmpi&e 17 Chittznnjan Arenu Li.o Fo or Bo-cdarville, Q.e&b Cal-ttl - 700 072 ibnralie Payot 14B SE6 P1111,P1lES Seervie deAbtnntmentr. Joy.d-r HodeBrildiog lrnont dl B;k Cototr C;se poet31e3312 CHINA 5th Mo Rod Goodhiosg-r Fifhh FlR, Mhip Life Building CH 11D2 La-ee Choo Fitondal & Ecomaeic Baegoloe.-D560 119 Ayala A-nue, Maka Publi hggHo aie Me;TOMManji TANZANIA 8,1 Da Po9 Dmoglie 3-5-1129 Kdg1aa Crwo Road Oxford Univermty Pre3s Blous Hyderobiad - 09 027 POLAND P.O. Box 5299 ORPAN DaresSalaam COLOMBIA P.arthano Flwk 2nd Floor Pol-a KultyN Noi Infolc Ltd. Nea-flokt3ar-& Nwrrngp-ra OD-l Warr THAILAND Apart.doA.ro 34270 Ahsredabad-3900C9 Cotal Departoet StIre Bogota D.E. PORTUGAL 3D6 SilOs Rad Pdalol Horse lirrtarl Po*tugal Bs-gkok COTE D'IVOIRE 16-A Aslok M ag Roe Do Carmo 70-74 Cenm d'Ebdiho et d1 Diffuion Lo - 226 OM 121O Lubon TRNDAD & TOBAGO, ANTIGUA Afric(CEDA) BARBUDA, BARBADOS, 04 BY. 541 INDONESIA SAUDI ARABIA, QATAR DOMINWCA, GRENADA, GUYANA, Abidjan 04 PIaeu Pt oirlimited Jdre BookSaIre JAMAICA, MONTSERRAT. ST. JL Sam Rahulngi 37 P.O. Box 3196 KITTS & NEVIS, ST. LUCIA, CYPRUS P.O. Box 181 Riyadh 11471 ST. VINCENT & GRENADINES MEMRB hlo nafr Setvrs Jakarba Pot Syte-a StudiesULit P.O. Box 2h98 MEMRB Ihf,natioo Sevi 19WolsStreet Nicosa ITALY Bed,ofa Cumepe Limrsa C'omni aria Son-i SPA AlAl. Stt Troidad, Wet ldies DENMARK Via Beredetto FPBtio, 120/10 Al Daho. Cotaer S.nofds.U ttus Crell2 Po32e 552 Pie Foor UNITED ARAB EMIRATES Rosa oreaAD1611 SM125 Florcnce P.O. Box 7188 MEMRB Gulf Co. DK-192SZDPrederikser6 C Riyadh P.O. Box 6097 JAPAN S-ajah DOMINCAN REPUBUC Earah- Bok S-eroice Ha0 Abdulsh AlireA aBuilding Editor Tolle, C. porA. 374 Hogo3.Choe,Bunky-k,.113 King KsdStreel UNITED KINGDOM Ralatord6 e oabd Is Cat61ia 309 Tokyn P.O. Boo 3969 Miarolfo Ltd. Apoetado Podd 219D Dr mm P.O. Box 3 Sooto Domingo KENYA Al, Hamprluie GU34 2PG Africa Book S-vice (EA.) Lid. 33, Modammed Hiaxrn Awad Street Enbgnd EL SALVADOR P.O. Bo 45245 P. Box 5978 Fu-odes Naiobi jeddalh VENEZUELA Avoid. Mooued EsiqueArajuo 03513 Lfbrien dd E Re Edifido 52SA, lec. Pio KOREA, REPUBLIC OF SINGAPORE,TAnWAN, Aptdo. 60337 Boo Spdod Pan Korea Book Copernhi MYANMAR,BRUNEI Cars 1060-A P.O. Boo 101, Kwangwhmrn hafomution Pubhcaw EGYPT, ARAB REPUBLIC OF Serol Prirat. lid. YUGOSLAVIA Al Abr 02-06 lot Fl, PYc-Fu lndrtol ogso1verna Kjiga Al Galas Seed KUWAIT Bldg. P.Q Boo 36 Caio MEMRBES lonfoor Senrves 24 New hdulttil Road TrR Repub2ike P.O. Box465 SiSWrel953 YU.LLDlSe4grde The Middle EBoa Observe 41, S1rsif Saeet MALAYSIA Cai- Urairraty od MatayaCoopelive Bokrdp, Limited P.O. Box- l27, Jtb. Pa Bar- Koal bLiamprar l..e Wa7r1d ;arlI Hea Tqu afss Enropean Office Tok7yo Office 1818 H Street, N.W. 66, avenue d'16na Kokusai Building Washington, D.C. 20433, U.S.A. 75116 Paris, France 1-1 Marunouchi 3-chome Chiyoda-ku, Tokya 100, Japan Telephone: (202) 477-1234 Telephone: (1) 40.69.30.00 Facsimile: (202) 477-6391 Facsimile: (1) 40.69.30.66 Telephone: (3) 3214-5001 Telex: wui 64145woRLDBANK Telex: 640651 Facsimile: (3) 3214- 3657 RCA 248423 WORLDBIK Telex: 26838 Cable Address: INTBAFRAD WASHINGTONDC ISBN 0-8213-2085-8