Hungary: Selected Issues

This Selected Issues paper on Hungary presents an empirical analysis of the leading indicators for inflation and models the determinants of inflation. It summarizes current knowledge about the behavior of inflation and thus underpins the subsequent discussion of possible changes to the current nominal anchor framework. The paper analyzes the growth potential and fiscal issues affecting that potential, and the external constraint. The paper suggests that some relatively stable econometric relationships can be found, despite the considerable structural and policy changes that occurred during the 1990s.


This Selected Issues paper on Hungary presents an empirical analysis of the leading indicators for inflation and models the determinants of inflation. It summarizes current knowledge about the behavior of inflation and thus underpins the subsequent discussion of possible changes to the current nominal anchor framework. The paper analyzes the growth potential and fiscal issues affecting that potential, and the external constraint. The paper suggests that some relatively stable econometric relationships can be found, despite the considerable structural and policy changes that occurred during the 1990s.

VI. Tax Reform in Hungary48

A. Introduction

140. Like most European OECD countries, Hungary places a higher burden of taxation on labor income than on capital income. By international standards, marginal labor tax rates are one of the highest of the European OECD countries, but are not atypical as compared to most Central and Eastern European countries. The main reason for the high marginal rates on labor is the high social security contribution rates levied on employers and employees to finance old-age pensions, health care, and other social transfers. During the last three years, developments in the Hungarian tax system have shown a reduction in the labor tax burden. In line with this trend, the 1999 tax reform aims at a further reduction in personal income and payroll taxes and a broadening of the base of social security contributions.

141. This chapter studies taxation issues in Hungary with a special focus on the impact of the reduction in labor taxes on economic efficiency and income distribution. To this end a simple model of the labor market is developed and calibrated with plausible parameters which are taken from the literature. The structure of the chapter is as follows. Section B provides a brief description of the tax system before the 1999 tax reform, comparing the Hungarian system with that of other countries. Section C studies the short-run and long-run effects of labor taxation from a theoretical point of view. Section D presents the tax reform incorporated in the approved 1999 budget. It tries to quantify the likely effects on tax rates, employment, and income distribution. Section E studies alternative sources of revenues that may be considered in the near future to compensate for the revenue loss of further cuts in labor taxes. Section F concludes and summarizes the results.

B. Aspects of the Hungarian Tax System

142. The principal elements of the Hungarian tax system were laid down in the 1988 tax reform which introduced the personal income tax (PIT)—with a top marginal rate of 60 percent and 11 tax brackets—and a value-added tax (VAT). In 1989, a market-oriented enterprise profit tax was introduced, although it came into effect only in 1992. Since then, no major structural reforms of the tax system have been undertaken. However, marginal tax rates, tax credits, tax exemptions and tax brackets of the PIT have changed almost on a yearly basis.49

143. Hungary’s tax system imposes a heavy burden on the economy. The overall tax burden—as measured by the ratio of tax revenue to GDP50—is above the OECD average and that of selected Central and Eastern European countries (Table 8).51 Labor taxes (including social security contributions) and VAT are the major revenue generating taxes, contributing 46 percent and 18 percent to 1996 revenues, respectively. Hungary’s tax burden declined gradually from 47 percent of GDP in 1990 to 42 percent in 1998 (Table 9), in line with the government’s commitment as agreed in the 1995 public finance reform program.52 The latter aim was accomplished through a reduction in marginal PIT rates and a cut in social security contribution rates. Moreover, the contribution base of social security declined as wages and salaries fell as percentage of GDP.53

Table 8.

Hungary: Some Comparative Revenue Ratios, 1996

(In percent of GDP)

article image
Source: OECD (1998) and staff calculations.

Unweighted average of 29 OECD countries.

Table 9.

Hungary: Consolidated General Government, 1990–99 1/

(In percent of GDP) 2/

article image
Source: Data provided by the Hungarian authorities; and Fund staff estimates.

Central budget, social security funds, extrabudgetary funds, and local governments. The data shown for consolidated revenue and expenditure exclude staff estimates of double-counting arising from transfers of funds between different levels of government. These estimates are subject to further revision. Data for 1996–99 are not strictly comparable with those for earlier years due to incomplete data for the extrabudgetary funds prior to 1996.

Official GDP data for 1994 have been adjusted to be comparable with the data for 1995 onwards.

Primary revenue is defined as total revenue minus transfer of profits from the National Bank of Hungary and interest revenues.

144. In 1998, the progressive part of the PIT had six different marginal rates of which the minimum rate is 20 percent and the top rate was 42 percent. Dividend income and capital gains on securities are taxed at a flat rate of 20 percent. The number of marginal rates was larger than what is typically observed for middle-income European countries, but lower than in most Central and Eastern European countries (see Table 10). The simple average of marginal PIT rates was 31.5 percent in 1998 which is broadly in line with the 30.5 percent weighted average.54 As average earnings were close to the lower income bound of the top PIT bracket, high marginal rates bit at a relatively low income level. Accordingly, a substantial share of the taxable base (about 33 percent) was taxed at the top rate so that changes in this rate would significantly affect economic efficiency and public revenues.

Table 10.

Hungary: Top Marginal PIT Rates and Number of Marginal Rates for Various Countries, 1997

article image
Sources: Deloitte Touche Tohmatsu International (1997) and Owens (1997).

Based on data for 1995.

United States, Japan, Germany, France, Italy, United Kingdom, and Canada.

Salary tax.

Flat rate that applies to all taxable income.

145. PIT revenues in Hungary are low by international standards. However, the top marginal PIT rate is close to the OECD average of 43 percent suggesting a relatively narrow PIT base. This can be explained by various factors. First, a part of the progressive tax base is lost due to various tax credits and specific exemptions. For example, interest income and social security benefits (provided on a net-of-tax basis) are tax exempt. Also, individuals could credit 20 percent of wages and salaries (up to Ft 4,200 per month) against the tax owed, thereby reducing the effective average tax rate.55 In 1998, 10 percent of the PIT base was lost due to tax credits. Second, high labor taxes in Hungary have adverse effects on employment and economic growth (see Section C), and may have caused a shift of economic activities to the black economy. Lackó (1995) estimated that the size of the hidden economy in Hungary was 30 percent in 1990 and that share increased by 4–5 percentage points between 1990 and 1993. This means that a significant portion of economic activity is not captured within the tax net.

146. As observed in most OECD countries, social security contributions in Hungary raise more revenues than personal income taxes mirroring high levels of social security expenditures. In 1997, 33 percent of revenues was derived from social security contributions against 15 percent from personal income taxes. The bulk of social security tax revenues was paid by employers who contributed 83 percent to social security revenues in 1998. This is reflected in the structure of contribution rates. Employers paid 43 percent of gross wages to social security of which 39 percent was earmarked for old-age pensions and health insurance and 4 percent was allocated to unemployment insurance. Employees paid much less social security contributions (11.5 percent of wages). Hungarian social security contribution rates are among the highest in the group of Central and Eastern European countries (Table 11). Nominal contribution rates have been cut in small steps from 63 percent in 1993 to 54.5 percent in 1998.

Table 11.

Hungary: Social Security Contribution Rates of Selected Countries, 19971

article image
Source: Deloitte Touche Tohrnatsu International (1997).

All contributions are made on gross salaries unless otherwise indicated.

Employers also pay a fixed contribution of Ft 3,600 per employee.

Individuals engaged in business activities make contributions on the basis of 35 percent of the previous year’s income tax base.

Employers contributions are based on gross salaries and other remuneration paid.

147. Revenues from the corporate income tax in Hungary are significantly below the OECD average (Table 8). This is not surprising given that the corporate income tax (CIT) rate is currently only 18 percent compared with an average rate of about 32 percent for OECD countries.56 The effective CIT rate in Hungary was only 15.7 percent in 1997. This can be explained by various factors. First, the CIT law reflects the goal of attracting foreign capital and stimulating investments in underdeveloped regions and sectors. Numerous special tax incentives were granted to firms in the past. Many of them have already been abolished but enterprises continue to enjoy the incentives for which they qualified. In 1997, CIT credits amounted to 33 percent of taxes payable before credits of which 61 percent was granted to joint ventures. Also, a special CIT rate of 3 percent applies to offshore companies which are engaged in trading activities or provide services to third countries. Second, due to the falling rate of inflation, the effective CIT rate also fell because depreciation allowances are defined on a historic cost basis.

148. The standard VAT rate in Hungary is 25 percent. A number of products and services are charged at 12 percent (such as basic foodstuffs and household electric power) and exports and pharmaceuticals are zero rated. Financial services, education, health services, and postal services (which constituted 5 percent of the VAT base in 1998) are exempted. Revenues from VAT (as percentage of GDP) are closely in line with international trends, but the top VAT rate is well above the OECD and European Union (EU) average.

C. The Macroeconomic Effects of Labor Taxation: Theory and Evidence

Short-run effects

149. Labor taxes cause a wedge between the after-tax income of workers and the gross labor costs paid by employers (i.e., the producer wage). Personal income taxes and employees’ social security contributions reduce the take-home pay of workers which discourages households’ labor supply. Similarly, if consumption taxes increase, households’ after-tax real wage declines, inducing them to substitute leisure for consumption. Employers’ social security contributions increase firms’ labor costs, which reduces their labor demand (or lowers the gross wage offered at a given level of employment). Accordingly, the effective tax wedge57 is the relevant concept to measure the tax burden on labor. Table 12 shows the nominal and effective tax wedge for Hungary during 1992–99. Indirect taxes form only a small part of the effective tax wedge. From 1989 to 1995, the nominal tax wedge (expressed as percentage of employers’ wage costs) increased by 15 percentage points. Since 1995, the nominal tax wedge has been declining and will reach a level of 57 percent of employers’ labor costs (70 percent of employers’ costs in effective terms) in 1999.

Table 12.

Hungary: Nominal and Effective Tax Wedges, 1989–99

article image
Source: Ministry of Finance: and Fund staff estimates

The effective PIT rate on progressively taxed incomes before tax credits.

150. To study the short-run employment effects of labor taxation, a simple model of the labor market is developed which is set out in Appendix I.58 To keep matters simple, a perfectly competitive labor market is assumed in which workers and firms take wages and prices as given.59 This reflects the flexible nature of the Hungarian labor market in which the majority of wage agreements are concluded at the plant level. Both a labor demand and labor supply schedule—featuring various tax rates and social spending—are derived from firm and household optimization behavior. The labor market model can be employed to study the effects of changes in PIT and social security rates on employment. The following equation defines the relative change in employment:


where tA, tM, tW, and tE denote the average PIT rate, the marginal PIT rate, employees’ social security contributions and employers’ social security contributions, respectively. Public spending on transfer income is denoted by G. The αis are composite coefficients of various elasticities:


where SU, SC, SI, SG, and D denote the uncompensated wage elasticity of labor supply, the compensated wage elasticity of labor supply, the income elasticity of labor supply, the public spending elasticity of labor supply, and the wage elasticity of labor demand, respectively.

151. The labor market model shows that if the government cuts marginal personal income (or social security) taxes combined with a reduction in government transfers to workers, employment is affected through three channels. First, the increase in the real consumer wage rate causes workers to supply more hours of labor—the substitution effect of a wage change. Second, higher consumer wages increase households’ income discouraging them from working additional hours—the income effect of a wage change. The net effect on employment is positive if households can easily substitute between consumption and leisure. Third, transfer income acts as a subsidy on leisure, so that a cut in government transfers will increase labor supply. If transfers are primarily provided to non-working pensioners and unemployed households, the increase in employment of a coordinated labor tax-transfer cut is smaller than when transfer income mainly accrues to working households because the third channel does not operate. Although it may seem counterintuitive, a reduction in average PIT rates depresses employment because the lower average rates raise after-tax income on intramarginal hours which induces households—through the income effect—to work less hours at the margin.

152. Statutory tax rates are not a good indication of tax incidence as the person upon whom the tax is levied may shift some (or all) of it to other parties. Tax shifting occurs if the price of what is taxed changes when the tax is imposed. Firms, for example, may be able to shift their tax burden to workers (through lower wages) or to consumers (via higher prices of final goods). The incidence of labor taxation depends on the wage elasticities of labor demand and labor supply. Employers are better able to shift the burden of taxation to employees if households’ labor supply is relatively inelastic and firms’ labor demand is relatively elastic (see the tax-shifting coefficient, ρ, in equation (2)). Estimates of wage elasticities of labor demand and labor supply are lacking for Hungary. Empirical evidence for the U.S.—which features a flexible labor market like in Hungary—demonstrates that employers are only able to shift a part of the burden of labor taxes to employees. Alesina and Perotti (1994) show that taxes are less distortionary and thus have smaller effects on employment in economies where wages are predominantly negotiated at the plant level60

Long-run effects

153. Labor taxes also affect employment through the household’s intertemporal tradeoff between consumption today and consumption in the future. Heijdra and Ligthart (1998) show, using a dynamic general equilibrium model for a closed economy, that an increase in labor income taxation causes a reduction in short-run labor supply, interest rates and consumption. Interest rates, private saving, domestic investment, and employment decline along the transition path toward the long-run equilibrium, which is characterized by a lower level of output, less employment, and a smaller capital stock.61 In the shortrun employees can shift a part of the burden of labor taxes to capital, but in the long run the incidence of labor taxes is fully borne by labor.

154. The direction of the effect of labor taxation on the level of income and the rate of economic growth is not clear cut. On the one hand, increases in taxes create distortions which may impede capital formation and economic growth, particularly at high rates of taxation.62 On the other hand, a higher level of public goods provision increases the productivity of private physical capital and human capital.63 Plosser (1992) finds that taxes on income and profits are growth depressing; an increase in the average income tax of 0.05 percent is associated with a reduction in the annual growth rate of more than 0.4 percentage points.

Various other studies—see Slemrod (1995) for an overview—have found a negative relationship but recent studies have demonstrated that this relationship is by no means robust.64

D. The Government’s Tax Reform Proposals

155. The 1999 budget introduced various new fiscal measures. This section studies these measures and analyzes quantitatively the consequences for tax rates, employment, and income distribution.

156. Box 1 summarizes the new measures most of which pertain to reforms in the area of labor taxation. The main changes made to the PIT system were: the reduction in marginal rates, the introduction of tax credits for children, and the reduction in general tax credits. In addition, social security contribution rates were cut, the social security tax base was broadened, and means testing for family allowances65 was abolished. Minor changes were made to VAT and the CIT system remained unchanged. The 1999 reform goes much further than the 1996 and 1997 reforms which mainly achieved the reduction of the top marginal PIT rate from 48 percent to 42 percent.

Effects on tax rates

157. The 1999 tax reform cuts the top marginal PIT rate from 42 percent to 40 percent. In addition, the number of tax brackets will be reduced from 6 to 3 simplifying tax administration considerably. Figure 19 shows the new and old rate structure. Weighted by the share of taxable base, the average of marginal PIT rates declines from 30.5 percent in 1998 to 28.2 percent in 1999. The latter measure neither takes into account any tax relief households enjoy from tax credits nor does it say anything about the effect on different types of households. To get an impression of the distributional consequences of the 1999 PIT reform, the effect on marginal and average PIT rates for three different groups of households are calculated.66 Also, an attempt is made to determine the effect on the effective average tax rate.

Figure 19.
Figure 19.

Marginal Personal Income Tax Rates

Citation: IMF Staff Country Reports 1999, 027; 10.5089/9781451817843.002.A006

The latter incorporates the effect of a reduction in general tax credits, the introduction of children’s tax credits, and the abolishment of means testing for family allowances in 1999. Therefore, families are also differentiated by the number of children they have.

158. Table 13 summarizes the tax rate effects of the 1999 tax reform. Column 3 presents the change in the average tax rate without any credits and allowances whereas in columns 4–6 the change in the effective average tax rate is calculated. The main beneficiaries of the 1999 tax reform are middle–and high–income households with two or more children. They enjoy the largest cut in effective average rates making the tax system less progressive for them. Due to the elimination of means testing, high–income and middle–income households with less than three children receive family allowances whereas they did not get any under the old system. Low-income households without children are hurt the most. They experience an increase in their effective average PIT rate of 6.5 percentage points.

Table 13.

Effects of the 1999 Tax Reform on PIT Rates

article image
Key: MR=marginal tax rate, AR=average tax rate, and CA=credits and allowances.

Includes general tax credits, children’s tax credits, and family allowances.

Low-income households (Ft 300,000), middle-income households (Ft 900,000), and high-income households (Ft 1,300,000).

Taxable base in the respective tax bracket as estimated by the Ministry of Finance.

Weighted by the share of taxable base in the respective tax band.

159. Two proposals pertaining to the introduction of tax credits to families with children were submitted to Parliament during preparation of the draft 1999 budget. The first proposal called for a high level of children’s credit and no increase in family allowances, which was adopted in the 1999 budget (see Box 2). An alternative proposal called for a lower level of tax credit of Ft 1,300 per child per month for families with one or two children and Ft 1,700 per child per month for families with three or more children. Lower children’s credits were complemented by an increase in family benefits of 8 percent as of May 1999.

Tax Reform in the 1999 Budget

Personal income tax:

  • Reduction in individual income tax: A new PIT schedule is introduced consisting of three tax brackets (20 percent, 30 percent, and 40 percent). The lower income bound of the top bracket is reduced from Ft 1.1 million in 1998 to Ft 1 million capturing a larger share of the taxable base in the top income band.

  • General tax credits are reduced: 10 percent of wages and salaries up to a maximum of Ft 3,000 per month may be credited if incomes are not larger than Ft 1 million. In 1998, tax credits amounted to 20 percent of wages and salaries up to a maximum of Ft 4,200 per month.

  • Introduction of children’s tax credits: Families with one or two children can credit Ft 1,700 per child per month against personal income taxes owed and with three or more children receive a tax credit of Ft 2,300 per child per month. Families with seriously disabled children may credit Ft 2,600 per child per month.

  • Reduction in investment tax credits: Rates are reduced from 30 percent to 20 percent and the amount is limited to Ft 200 thousand per year.

Social security contributions:

  • A reduction in employers’ social security contribution rates: Rates are reduced by 7 percentage points. Contributions to old-age pensions fall by 2 percentage points, social security payments earmarked for health insurance drop by 4 percentage points, and contributions to the unemployment benefit scheme decrease by 1 percentage point. To partially compensate for the loss of revenues from the rate cuts, the fixed health care contribution per employee is increased from Ft 2,100 in 1998 to Ft 3,600 in 1999.

  • An increase in employees’ social security contribution rates: The rate of contributions to old-age pensions increases by 1 percentage point.

  • Broadening of the contribution base: An 11 percent health–care contribution is levied on dividend incomes which was previously exempt. Income from company cars will be subjected to a 25 percent health-care contribution. In addition, a 3 percent contribution to the unemployment benefit scheme applies to company cars.

Consumption taxes:

  • Increase in excises: The specific rates of various excisable products increase on average by 11 percent. Cigarettes, petrol, vehicle gasoline, and fruit brandy will increase by more than the average.

  • Changes to VAT: VAT rebates (up to Ft 400 thousand) are granted for the construction or purchase of new dwellings. Textbooks for primary and secondary education, diapers and oxygen for medical treatment are taxed at a zero rate. The VAT rate on pharmaceutical raw materials which can be used in other industries will increase from zero percent to 25 percent.

160. Tables 13 and 14 show the effects of both proposals on the effective average PIT rates for various types of households. Clearly, the alternative proposal (Table 14) provides more tax relief on average than the adopted proposal. In particular, the effective tax burden of low-income households would be reduced by more than under the high tax credits scenario because they benefit from the increase in family allowances. Moreover, the alternative proposal would put less pressure on the budget. The Ministry of Finance had estimated that the high tax credits scenario would cost the budget Ft 36 billion against Ft 31.2 billion for the alternative proposal.

Table 14.

Effect of the Alternative Proposal on Effective Average PIT Rates 1/

article image

Alternative proposal compared to the past rates.

Weighted by the share of taxable base in the respective tax band.

161. The 1999 tax reform reduces employers’ social security contribution rates by 7 percentage points whereas employees’ rates increase by 1 percentage point. However, the fixed health care contribution employers must pay per employee increases by 71 percent thereby attenuating the effective reduction in contribution rates. However, the contribution base is broadened to include non–wage income. As of January 1999 dividend income and income from the use of company cars are subjected to an 11 percent health care contribution.

Employment effects

162. To estimate the employment effect of the 1999 labor tax reform, ideally one would like to employ a computable general equilibrium model of the Hungarian economy. However, this is beyond the scope of the present paper. To get a feel for the likely size of the tax impact on employment, a partial equilibrium approach is taken which uses the model developed in the previous section. Plausible values of wage elasticities of labor supply and labor demand, taken from the literature, are substituted in equations (1)–(2) together with changes in marginal and average PIT rates and changes in social security taxes.

163. Because Hungary’s labor market is quite flexible, estimates of U.S. elasticities are employed as proxies for the true values. A high and low scenario for the wage elasticity of labor demand is considered, the results of which are presented in Table 15.67 Tax shifting by employers—defined previously in (2)—is small in the low elasticity scenario, but reaches a value close to unity for a wage elasticity of labor demand of about 0.3. In that case, employers are able to shift almost the entire burden of labor taxes to employees. The average employment effect is larger in the high elasticity scenario, but is generally small. Employment of middle-income households is affected the most by the cut in labor taxes in both scenarios. The labor supply response of high-income households is the smallest as they experience a larger negative income arising from the reduction in average tax rates.

Table 15.

Employment Effects of the 1999 Reform

(In percentage points)

article image

Weighted by the share of taxable base in the respective tax band.

Revenue effects

164. Cutting personal income taxes and social security contributions does not necessarily lead to lower revenues if initially labor taxes are set beyond levels that maximize total revenues. Then, the economy operates on the downward-sloping section of the Laffer curve.68 No empirical studies have estimated the Laffer curve for Hungary, but revenue losses associated with labor tax rate cuts in previous years indicate that the Hungarian economy is likely to be on the correct side of the Laffer curve.

165. The 1999 fiscal reform aims at a reduction in the overall tax burden (as measured by the revenue to GDP ratio) by 1.6 percentage points whereas expenditures are budgeted to decline by 2 percentage points so that the fiscal balance improves slightly compared to the previous year. The Ministry of Finance has budgeted a reduction in the ratio of labor tax revenues (including social security taxes) to GDP of 0.9 percentage point as a result of the 1999 tax reform (see Table 9). However, the ratio of PIT revenues to GDP does not change despite the lower PIT rates because a part of the revenue loss due to a fall in tax rates is recouped by a decline in tax credits. Tax credits have been projected to decline by 9 percent because of the large fall in general tax credits which exceeds the increase in tax credits from the introduction of children’s credits.69

E. Coordinated Tax Reform: Offsetting the Revenue Loss

166. The Hungarian government intends to bring down the burden of labor taxes in the medium term. Given the government’s aim to close the fiscal gap there is limited latitude for revenue losses unless expenditures are further cut. In search of alternative sources of revenue to compensate for the revenue loss of labor tax cuts, three compensatory revenue sources will be studied which have received attention in the European tax reform debate in the early 1990s: (i) consumption taxes, (ii) environmental taxes, and (iii) corporate income taxes.

Consumption taxes

167. Changing the tax mix from personal income taxes toward consumption taxes was on the tax reform agenda in many OECD countries during the last decade. It was argued that this would increase work incentives as after-tax wages were expected to increase. However, the labor market model (see Appendix I) shows that consumption taxes have adverse effects on employment as they are “implicit taxes” on labor income. Higher consumption taxes decrease the real value of take-home pay which reduces the household’s willingness to supply an additional hour of labor. However, from an efficiency point of view, broadly-based consumption taxes are a better device to raise revenues. Due to their broader tax base—including transfer income and capital income in addition to net labor income—they are generally less distortionary than labor taxes. Moreover, consumption taxes are less easy to avoid and evade than personal income taxes.

168. Converting personal income taxes to general consumption taxes may have undesirable distributional consequences. Higher consumption taxes reduce pensioners real income compared to labor taxation as pensioners will benefit from the transfers but do not bear the burden of labor taxes. This conclusion depends crucially on the assumed absence of wage-linked transfers. Because VAT rates in Hungary are high by international standards and a large share of transfer income is provided to low-income households outside the labor market, switching from labor to consumption taxation is not a viable option in Hungary.70 Indeed, Hungary needs to reduce the 25 percent VAT rate so that its VAT system will not hinder Hungary’s aim of EU integration. As Vamosi–Nagy et al: (1998) have suggested, revenue losses associated with reductions in the top VAT rate could possibly be offset by an increase in the middle VAT rate of 12 percent.

Environmental taxes

169. The Hungarian Environmental Action Plan calls for the introduction of air pollution taxes in the near future which will be discussed in Parliament this year.71 Pollution taxes improve economic efficiency by charging polluters with a price for the damage they cause. Accordingly, the (too low) private marginal costs of polluting activities are raised to the level of the social marginal costs. Besides internalizing pollution externalities, pollution taxes generate government revenues as an additional benefit. Recently, economists have suggested using the revenues from environmental taxes to cut distortionary labor taxes in a revenue-neutral fashion. This may yield a “double dividend:” (i) an improvement in environmental quality, and (ii) higher employment through a less distortionary tax system (Pearce (1991), Bovenberg and de Mooij (1994), and OECD (1995)). Ligthart (1998) argues that such a double benefit can only be reaped if the burden of green taxes can be shifted to those outside the labor market such as capital owners, transfer recipients, and foreign energy producers. However, this is likely to come at the cost of a distortion in the income distribution.

170. In various European countries the double dividend hypothesis was eagerly embraced by green lobbies to push for politically unpopular environmental taxes. Resistance to green taxes is based on their adverse effect on the income distribution and the existing pattern of international competitiveness. Energy taxes are generally regressive as they disproportionally hurt low income households which spend a relatively large share of their income on energy. Moreover, energy taxes raise production costs of energy–intensive firms. This weakens their competitiveness and may induce capital flight in the long run. Therefore, the OECD “Jobs Study” (1995) proposes a coordinated introduction of energy taxes in all EU countries. In 1996, however, the Netherlands unilaterally introduced an energy tax on households and small firms.

171. Some empirical studies—based on large scale macroeconometric models—have found small but positive employment effects of green tax reforms for various European countries (OECD (1997)). Simulations with the European Community’s QUEST model for the EU as whole show that a reduction in social security contributions by 10 percent, financed by an increase in carbon taxes decreases unemployment by 0.9 percentage points.72 Other studies based on simulations with general equilibrium models have found negative employment effects (see Goulder (1995) for an overview). No systematic sensitivity analysis has been performed yet to unravel the channels underlying the differences in results.

172. Transition economies face several difficulties beyond the typical problems encountered in market economies in adopting a system of eco-taxes.73 First, administrative limitations inherited from centralized planning may curtail the effective administration and enforcement of green taxes. Second, the provision of energy below market prices contributes to wasteful energy use by households and firms. Third, many enterprises retain monopoly power which limits the incentive effects of environmental taxes because they can pass on any increase in production costs to consumers via higher prices. Moreover, many enterprises, even after privatization, lack sufficient information on the technological possibilities for pollution control.

Corporate income taxes

173. When deciding on the optimal level of CIT policy makers need to strike a balance between two opposing forces. On the one hand, they want to sustain tax revenue. On the other hand capital is highly mobile, so increasing tax rates on it may lead to capital flight and governments also try to boost investment by providing an attractive tax climate to both foreign and domestic investors. Hungary has modeled its CIT system along the second line. Consequently, the marginal excess burden of capital taxation is significantly below the marginal excess burden of labor taxation. Capital is thus likely to be “undertaxed”—particularly compared to the U.S. tax system—and may promote the substitution of capital for labor in an environment of already high unemployment.

174. Hungary promotes investment via low nominal CIT rates and various special investment incentives which erode the base of the CIT. In 1995, the general CIT rate was lowered from 36 percent to 18 percent. A tax of 20 percent applies to dividends. Profit income distributed as dividends are effectively taxed at a rate of 34.4 percent which creates an incentive to reinvest earnings as these are only taxed at 18 percent.74 Although, the effective tax on income taken out as dividend is close to the implicit rate75 on progressively taxed personal income (31 percent in 1998), the tax burden on labor significantly exceeds this value due to social security taxes. To make it less attractive for enterprises to take out income in the form of dividends rather than wage income, the dividend tax is raised to 35 percent if dividends withdrawn exceed twice the basic rate of the National Bank of Hungary.

175. Tax incentives granted to particular sectors may be justified on the grounds of domestic market failures—if the industry under consideration generates positive externality to the rest of the economy. These externalities are hard to measure and it is therefore difficult to determine how much government support a particular industry should receive. In the absence of any spillovers, however, investment incentives misallocate resources since they create effective tax rates that differ across sectors or regions. Broadway and Shah (1992) conclude in their review of studies on the impact of tax incentives in various developing countries that these incentives yielded windfall gains to investments that would have occurred anyway.

176. Broadening the base of the CIT by eliminating special tax incentives would generate additional revenues. Tax reforms in OECD countries since the mid–1980s were aimed at reducing the extent of tax incentives and broadening the CIT base. In the early 1990s, the Hungarian tax law repealed various CIT exemptions and tax incentives, but this trend was reversed in 1996 when the range of activities that qualified for tax credits was significantly expanded to exports, certain investments, and underdeveloped regions.

F. Conclusion

177. Hungary’s tax system taxes labor income heavily. Marginal labor taxes are high by international standards and the labor tax base is narrow due to various exemptions and tax credits. The main reason of the high marginal tax rates is the high social security contribution rates, particularly those levied on employers, which are needed to sustain high levels of social spending. High labor tax rates have caused a substitution of capital for labor in a setting of already high rates of unemployment.

178. Since it was laid down in 1988, the tax system of Hungary has been revised on almost a yearly basis. Tax reforms in the past three years were aimed at reducing the labor tax burden to foster employment and economic growth. This chapter has studied the reforms introduced in 1999 which consist of a further reduction in the burden of labor taxation—through a reduction in marginal rates and a broadening of the tax base—in line with the past trend and tax reforms pursued in other countries. It was shown that the 1999 reform increases after-tax income of middle–and upper–income households. The employment effects of cuts in labor taxes are likely to be relatively small, pointing at the need for further reform.

179. Unless the government undertakes a major expenditure reform, tax reform aimed at a further reduction in marginal labor tax rates will widen the fiscal gap. Increases in corporate income tax rates and the elimination of various investment incentives may (partially) offset revenue losses associated with reductions in labor taxes.


  • Alesina, A., and R. Perotti, 1994, “The Welfare State and Competitiveness,NBER Working Paper No. 4810 (Cambridge: NBER).

  • Barro, R.J., 1990, “Government Spending in a Simple Model of Endogenous Growth,Journal of Political Economy, Vol. 98, pp. S10325.

    • Search Google Scholar
    • Export Citation
  • Bovenberg, A.L., and R.A. de Mooij, 1994, “Environmental Levies and Distortionary Taxation,American Economic Review, Vol. 84, pp. 108589.

    • Search Google Scholar
    • Export Citation
  • Broadway, R., and A. Shah, 1992, “How Tax Incentives Affect Decisions to Invest in Developing Countries,World Bank Policy Research Working Paper No. 1011 (Washington: World Bank).

    • Search Google Scholar
    • Export Citation
  • Deloitte Touche Tohmatsu International, 1997, Taxation in Eastern Europe (New York).

  • Easterly, W., and S. Rebelo, 1993, “Fiscal Policy and Economic Growth: An Empirical Investigation,Journal of Monetary Economics, Vol. 32, pp. 187212.

    • Search Google Scholar
    • Export Citation
  • Goulder, L.H., 1995, “Environmental Taxes and the Double Dividend: A Reader’s Guide,International Tax and Public Finance, Vol. 2, pp. 24551.

    • Search Google Scholar
    • Export Citation
  • Hamermesh, 1993, Labor Demand (Princeton: Princeton University Press).

  • Hausman, J.A., 1985, “Taxes and Labor Supply,in A.J. Auerbach and M. Feldstein (eds.), Handbook of Public Economics, Vol. I (North-Holland: Elsevier Science Publishers).

    • Search Google Scholar
    • Export Citation
  • Heijdra, B.J., and J.E. Ligthart, 1998, “The Dynamic Macroeconomic Effects of Tax Policy in an Overlapping Generations Model,IMF Working Paper 98/182 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Lackó, M., 1995, “The Hungarian Hidden Economy in International Comparison: Estimation Method Based on Household Electricity Consumption and Currency Ratio,Working Paper No. 25, Institute of Economics (Budapest: Hungarian Academy of Sciences).

    • Search Google Scholar
    • Export Citation
  • Ligthart, J.E., and F. Van der Ploeg, 1995, “Environmental Policy, Tax Incidence, and the Cost of Public Funds,Nota di Lavoro No. 4.95 (Milano: Fondazione Eni Enrico Mattei).

    • Search Google Scholar
    • Export Citation
  • Ligthart, J.E., 1998, “The Macroeconomic Effects of Environmental Taxes—A Closer Look at the Feasibility of ‘Win-Win’ Outcomes,IMF Working Paper 98/75 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Leibfritz, W., Thorton, J., and A. Bibbee, 1997, “Taxation and Economic Performance,OECD Working Paper No. 176 (Paris: OECD).

  • Newbery, D.M.G., and T. Révész, 1997, “The Burden and Disincentive Effects of Hungarian Personal Taxes, 1988–96,CEPR Discussion Paper no. 1570 (London: CEPR).

    • Search Google Scholar
    • Export Citation
  • OECD, 1994a, “Taxation and the Environment in European Economies in Transition,OECD Working Paper No. 29 (Paris: OECD).

  • OECD, 1994b, Taxation and Household Saving (Paris: OECD).

  • OECD, 1995, The OECD Jobs Study: Taxation, Employment, and Unemployment (Paris: OECD).

  • OECD, 1997, Environmental Taxes and Green Tax Reform (Paris: OECD).

  • OECD, 1998a, Revenue Statistics, 1965–97 (Paris: OECD).

  • OECD, 1998b, Statistics on Eco-Taxes: Progress Report, DAFFE/CFA/CT(98)19 (Paris: OECD).

  • Owens, J., 1997, “Tax Reform for the 21st Century,Tax Notes International, Vol. 14, February 17, pp. 58395.

  • Pearce, D.W., 1991, “The Role of Carbon Taxes in Adjusting to Global Warming,Economic Journal, Vol. 101, pp. 93848.

  • Plosser, C.I., 1992, “The Search for Growth” in Policies for Long-Run Economic Growth (Kansas City: Federal Reserve Bank of Kansas City).

    • Search Google Scholar
    • Export Citation
  • Ruggiero, E., 1998, “Reform of Public Finances, 1994–97,in Cottarelli, C. et al. Hungary: Economic Policies for Sustainable Growth, Occasional Paper No. 159 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Slemrod, J., 1995, “What Do Cross–Country Studies Teach about Government Involvement, Prosperity, and Economic Growth?Brookings Paper on Economic Activity, No. 2, pp. 371431.

    • Search Google Scholar
    • Export Citation
  • Vamosi–Nagy, S., I. Kocsis, and L.A. Sanchez, 1998, “Tax Policy Reforms in Hungary,in L.J. Bokros and J.J. Dethier, Public Finance Reform during the Transition—The Experience of Hungary (Washington: The World Bank)

    • Search Google Scholar
    • Export Citation

APPENDIX I: The Employment Effects of Labor Taxation: A Theoretical Model

1. In this Appendix a simple analytical model is developed to study the employment effects of government spending, labor taxation, and consumption taxation. In addition, the model touches upon the issue of tax incidence. It is shown under what conditions the incidence of labor and consumption taxation falls on employees. The model assumes a perfectly competitive labor market that fully clears. Below, the three sectors of the economy—firms, households, and the government—are described.

A. The Analytical Model

2. Consider a representative firm which supplies a homogeneous good, Y, which is produced under constant returns to scale. The production function is given by Y=aF(L), with properties FL>0 and FLL<0 where L denotes employment and a represents economy-wide productivity. The firm maximizes profits, II, under perfect competition by choosing employment:


where p is the price of final output, w is the gross wage rate and tE is the employer’s tax on labor (i.e., social security contributions). Profit maximization yields the following optimality condition: aFL(L)=w(1+tE)/p=wp which says that the marginal productivity of labor should equal the real producer wage (inclusive of tax). Rewriting this expression yields the firm’s demand for labor: Ld=ld(wP,a). To study the comparative static effects of changes in tax rates, the model is linearized around an initial position. Denote a relative change by a “~”, e.g., w~=dw/w, except t~E=dtE/(1+tE). The equation describing the relative change in labor demand is given by:


where ∈d denotes the wage elasticity of labor demand. A rise in the employer’s tax on labor shifts the demand curve for labor to the left thereby reducing employment at a given (tax exclusive) real wage. A favorable productivity shock is equivalent to a reduction in employers’ payroll taxes, shifting the labor demand curve to the right, thereby increasing real wages for a given level of employment.

3. The representative household derives utility, U from consumption, C, and leisure, V ≡ 1-L:76


Households maximize utility subject to the household budget constraint: p(1+tC)C=w(1-tA)L+G, where G denotes lump-sum transfers from the government (which is assumed to be tax exempt), tC is a consumption tax, and tA is the average labor tax levied on employees’ labor income.77 No separate social security tax on employees is distinguished because its effects are qualitatively similar to changes in the marginal income tax rate. To keep matters simple, households’ labor income is assumed to be taxed according to a single marginal tax rate, tM, of which T0 is tax exempt, so that tax revenue from labor income is given by: -T0+tMwL.78 With a single labor tax rate and no tax exempt threshold (or zero-rated income band), average tax rates and marginal tax rates would theoretically coincide. However, Hungary does not have a zero-rated income band. Obviously, in a tax system of various tax bands—as is the case for Hungary—average and marginal rates always differ.

4. From the household’s optimization problem the following expression characterizing the optimum can be derived: UV/UC = wC=w(1 -tM)/p(1+tC), where wC denotes the real consumer wage. Workers equate the marginal rate of substitution between leisure and consumption to the real consumer wage—which features the marginal tax rate rather than the average tax rate since the former is the relevant tax rate for the consumer’s decision to supply an additional hour of labor. From this and the household budget constraint, the labor supply equation, Ls = ls (wC, G), results which features the consumer wage and transfer income as arguments. In relative changes this yields:79


where the elasticity of labor supply with respect to transfer income is ∈SG ≡ αL(1-ω)>0, ω denotes the share of net wage income in total household income, αLV/(ωV+L)>0 measures the intensity of leisure consumption, and the wage elasticity of labor supply is given by:


where ∈SC = Vσ>0, ∈SI = Vω>0, ∈SPV(σ–1) stand for the compensated wage elasticity, the income elasticity, and price elasticity of labor supply, respectively.80 A higher wage or a lower consumer price exerts two (converse) effects on labor supply. First, it makes leisure more expensive and thus increases the supply of labor (i.e., the substitution effect, represented by ∈SC). Second, households feel more wealthy, inducing them to consume more leisure and thus supply less labor (i.e., the income effect, denoted by ∈SI). An upward sloping labor supply curve—with ∈SU>0—is obtained if the substitution effect dominates the income effect. This requires that the elasticity of substitution between consumption and leisure, σ, is sufficiently large (i.e., σ>ω). Accordingly, the larger σ, the more elastic the labor supply curve.

5. An increase in the marginal labor tax rate shifts the labor supply curve to the left thereby decreasing the number of hours supplied at a given real wage. However, an increase in the average tax rate shifts the labor supply curve to the right and thus raises the number of hours worked at a given real wage rate. This somewhat counterintuitive result can be explained by the income effect in labor supply: household will work harder to compensate for the income loss. Conversely, through the income effect, a rise in government transfers depresses labor supply, particularly if the share of transfers in household income is large and a lot of leisure is consumed by the household.

B. The Effects of Labor Taxation on Employment and Wages

6. There are several ways to close the model. By assuming full wage flexibility it is assured that labor market equilibrium, Ls = Ld, is obtained implying that any observed unemployment is of voluntary nature. Alternatively, the disequilibrium interpretation assumes a real consumer wage which is exogenously fixed above a level that ensures full employment.81 Accordingly, employment is determined by labor demand and unemployment is the defined by (Ls-Ld)/Ld. In the following a fully clearing labor market is assumed.

7. To study the employment and wage effects of taxation, labor market equilibrium is perturbed by small changes in the government’s transfer and taxation instruments. From this, the relative change in employment can be derived:82


and the relative change in the wage rate is given by:


A measure of the progressivity of the labor tax system is given by the elasticity of after-tax labor income with respect to pre-tax income:


An increase in the marginal tax rate, t~M>0, combined with an increase in the tax exempt threshold to keep the average tax rate constant so that taxes become more progressive (i.e., z>0) decreases employment and raises the wage rate through the negative substitution effect in labor supply. However, a higher average tax rate, t~A>0, through a decline in the tax exempt threshold, raises the progressivity of the tax system, raises employment and depresses the wage rate. Employment declines and wages are pushed up if the government increases transfers or raises consumption taxes. Lowering the marginal labor tax reduces the progressivity of the tax system and increases employment.

8. The model can be easily used to study the issue of tax incidence which draws on the notion that the agent upon whom the statutory tax is levied is not necessarily the person who effectively pays the tax. Assume now, for simplicity, that the personal income tax is characterized by a flat rate without a tax exempt threshold so that the marginal and average tax on labor are equal. To simply matters further it is assumed that households do not receive any transfers from the government and productivity does not change. From the model, the relative change in the real producer and consumer wage follow:


which depends on the tax-shifting coefficient, ρ≡∈D/(∈SU + ∈D)>0, and the relative change in the “tax wedge” which is defined as the sum of the respective tax rates. The tax-shifting coefficient measures the degree to which employers are able to shift the burden of labor taxes to employees. Tax shifting crucially depends on the values of the wage elasticities of labor demand and supply. Under “normal” assumptions about labor supply and demand elasticities, the tax-shifting coefficient takes on values in the range between zero and one. Tax shifting by employers is large—with p close to unity—if labor supply is relatively inelastic and labor demand is relatively elastic. In that case, the consumer wage decline is large whereas the change in employment is small. A number of special cases can be distinguished. In case of a vertical labor supply curve, ρ=1, an increase in the labor tax rate is fully borne by employees implying that the consumption wage falls one-for-one with the increase in the labor tax rate. However, employees can fully escape the burden of labor taxes if labor supply is perfectly elastic but employment declines a lot. This is equivalent to the case of a rigid real consumer wage at the level of full employment. In the rare case of a backward bending labor supply curve—su<0 and ρ>1—employers can overshift the burden of labor taxes.

9. In a standard neoclassical labor market model the “Invariance of Incidence Proposition” (IIP) holds which says that it does not matter for the effect on employment, consumer wages, and producer wages whether the statutory labor tax is levied on employers or on employees. Ergo, swapping higher employees’ social security contributions for an equal decrease in employers’ social security contributions does not affect economic efficiency; consumption wages and the level of employment remain unaffected. However, the IIP breaks down when some of the underlying assumptions are relaxed. Allowing for a low degree of competition on the labor market resulting from rigidities such as regulations (i.e., minimum wages) and trade unions, invalidates the proposition. In such a case, employees can shift a larger part of the tax burden than under perfect competition. Similarly, capital income, which is taxed under the personal income tax but not under employers’ social security taxes modifies the proposition. Swapping an increase in the personal income tax for an equal decrease in employers’ taxes would alleviate the tax burden on labor encouraging a substitution of labor for capital in production. Because personal income taxes are progressive whereas social security taxes are generally levied at a flat rate, shifting the revenue raising task from social security taxes to personal income taxation may increase the tax burden on labor.


Hungary: Summary of Major Taxes as of November 1, 1998

article image
article image
article image
article image
article image
article image
article image
article image
article image
article image
article image
article image

Prepared by Jenny Ligthart.


Vamosi–Nagy et al. (1998) provide further details on the developments in the Hungarian tax system over the last 10 years. Newbery and Révész (1997) report that during 1988–96 the top marginal PIT rate was changed seven times whereas the number of tax brackets changed on six occasions.


Note that, tax revenue to GDP is not all an encompassing indicator of tax distortions at the micro level as it does not include the excess burden (i.e., the welfare costs of taxation in excess of public revenues collected) arising from behavioral responses to taxes and public goods financed with these taxes.


Key features of the Hungarian tax system can be found in Appendix II.


Compared to other European OECD countries, Hungary has done well during the last five years. In these countries, the (unweighted) tax revenue to GDP ratio continued to increase from 38 percent in 1985 to 41 percent in 1996 reflecting the increased pressure on social spending resulting from the aging of the population and high levels of unemployment


See Ruggiero (1998) for further details on the 1995 public finance reform program.


This measure weights the marginal rates by the percentage of taxable base that falls in the various tax brackets.


A tax credit system is more equitable from a distributional point of view. Under a deduction system, individuals at higher marginal rates derive greater absolute benefits than those at lower marginal rates.


Calculated using CIT rates for 1995 as listed in Owens (1997).


The nominal tax wedge includes contributions for old-age pensions, health care, and unemployment insurance paid by both employers and employees. The effective tax wedge includes, in addition, VAT and excise taxes.


The model extends Ligthart and van der Ploeg (1995) to allow for various tax instruments, social security contributions, and public spending.


The model abstracts from important aspects of reality such as the effect of minimum wages and trade unions on equilibrium wages and employment.


See Section D for values of the elasticities and some illustrative calculations of the employment effects of cuts in labor taxes.


According to an OECD (1994b) study there is no clear evidence that the level of taxation does generally affect the level of household saving. Leibfritz et al. (1997) provide more details on how taxation may affect total saving and domestic investment.


From the theory of public finance it is well known that the distortionary effect of taxes increases with the square of the tax rate.


Barro (1990) has shown in a simple endogenous growth model in which public spending enters a constant returns to scale production function that the relation between the size of the government and the per capita growth rate is an inverse U-shaped curve. On the upward-sloping section of the curve, the efficiency enhancing role of public spending exceeds the distortionary effect of taxation.


Easterly and Rebelo (1993) point out that the “evidence that tax rates matter for growth is disturbingly fragile” as the negative correlation disappears when the initial level of income is controlled for.


Granted to families with children up to the age of 16 (20 in the case of students) and those who are chronically ill regardless of age. The family allowance is a tax free amount which varies according to the number of children in the family.


This analysis focuses only on tax rate effects and does not allow for changes in the tax base caused by changes in wages or employment. See the subsection on the revenue effects of tax reform for a further elaboration on this.


The low elasticity of labor demand scenario takes ∈D=0.02 from Alesina and Perotti (1994) and the high elasticity of labor demand scenario takes ∈D=0.30 from Hamermesh (1993). The following labor supply estimates for the U.S. are used: ∈SU=0.03, ∈SC=0.95, and ∈SI=0.98 (Hausman (1985)), Other empirical studies of married men’s labor supply find zero and in some cases even negative uncompensated wage elasticities of labor supply for the U.S. and some European countries. In practice, these aggregate elasticities vary in value between different groups depending on hours worked and family composition. The effect of public spending on employment is abstracted from.


Intuitively, a rise in the labor tax rate has two effects on public revenues. First, a tax rate effect which increases public revenues. Second, a negative tax base effect as higher tax rates erode the tax base through lower employment. On the downward-sloping section of the Laffer curve the second effect dominates the first.


The revenue loss due to tax credits in 1999 amounts to 27 percent of progressive personal income taxes before credits.


Newbery and Révész (1997) have analyzed the distributional impact of revenue-neutral indirect tax reforms that yield the same set of relative consumer prices in each year. They found that these reforms had an adverse effect on the income distribution during 1989–96.


The main environmental instruments are penalties and (fixed) charges on environmentally damaging products, the revenues of which are earmarked for environmental purposes.


Public revenue from environmental taxes is still limited. They typically account for less than 1 percent of GDP in EU countries (Owens (1997)).


See OECD (1994a) for further details on this issue.


Retained earnings are favored anyway as Hungary employs a classical system of corporate income taxation with double taxation: enterprises pay taxes on income taken out as dividends and households pay a tax on dividend income.


The ratio of CIT revenues to the CIT base.


The number of hours available to the household has been normalized to unity to keep matters simple.


Because the fixed capital stock is rented from households, profit income accrues to working households. For simplicity, it is assumed that profits are not included in the household’s budget constraint and thus do not affect its labor supply decision.


Government expenditures consist of transfer income which are financed by taxes on labor income and consumption: G + T0 = (tE + tA)wL + tCC.


The following definitions are employed: t~MdtM/(1-tM),t~AdtA/(1-tA),andt~CdtC/(1+tC).


Without transfer income, ∈SU and ∈SP are equal to V(σ-1).


The inflexibility of real consumer wages can, for example, be explained by the presence of trade unions who set a wage above the market clearing level to maximize the utility of its members.


In the following, changes in the producer price are omitted.