Abstract
This paper examines a few selected aspects of adjustment and restructuring in the Czech enterprise sector. It examines financial performance of the Czech enterprise sector, drawing on data reported by nonfinancial enterprises to the Czech Statistical Office (CSO) in 1993. The findings reveal that significant progress has been made in the nonfinancial aspects of adjustment, but a large part of the enterprise sector has been slow to adopt measures to achieve sustainable financial viability.
III. Pension System Trends and Reforms: A Comparison of the Czech Republic, Hungary and Poland 1/
1. Introduction
As the countries of central and eastern Europe move toward market-based economic systems, the role of Government in all aspects of production and distribution is being re-evaluated. Social security--broadly defined as the protection of individuals from fluctuations in real income due to unemployment, disability, old age, family circumstances, etc.--previously existed implicitly due to the Government’s extensive control over the allocation of resources. As the Government’s allocative role is reduced, explicit social insurance schemes are required to replace the implicit mechanisms of the old system.
This note addresses the current and projected states of pension systems in three countries in central and eastern Europe; namely, the Czech Republic, Hungary, and Poland. The analysis concentrates on the financial positions of the pension systems, and projections of future expenditure trends, and does not address such issues as the distortionary impact of the financing mechanisms and benefit rules, or the implied distributional consequences. The broad objective is to evaluate the urgency of reform in the Czech Republic in comparison with the situation in Hungary and Poland, and to provide some information on the possible reform options under consideration in the Czech Republic.
2. Current financial positions of pension systems
This section examines the current levels of expenditures and revenues of the pension systems in the three countries. Total pension expenditure depends on the number of pensioners and the average pension received, which in turn are determined by the eligibility requirements (e.g., standard retirement age, allowance of early and/or postponed retirement, rules for determining partial or full disability, bereavement rules, rules about working while receiving a pension, etc.) and the benefit rules (e.g., relation to past profession, earnings, and years of contributions, and indexation rules). 2/
Financing of the three pension systems is nominally provided by proportional taxes on labor income. The revenues from these taxes may not be sufficient to cover expenditures on an annual basis, and it is unlikely that they will provide a sufficient revenue source over the long term. Without increases in tax rates and broader effective tax bases, the government budget must cover any pension shortfalls.
These spending and financing rules have direct effects on the incentives of individuals to work, retire, and save, and can also present individuals and employers with substantial arbitrage opportunities (with no necessary change in behavior). Such incentives increase the net financial cost of the system, and both the direct and associated distortions lower economic welfare.
a. Spending patterns
Although the underlying determinants of aggregate pension spending relate to the micro design of the system discussed above, expenditures as a proportion of GDP, σ, can be usefully decomposed according to the following identity:
where S is total pension spending, b is the average benefit received by a pensioner, w is the average wage, p is the number of pensioners and n is the number of workers, ρ = b/w defines the replacement rate, δ = p/n is the system dependency ratio, and α = nw/GDP is labor’s share in output.
A number of points about these definitions need to be made. First, pension benefits, b, may be measured on a pre- or post-tax basis. 1/ Second, wages may be defined as gross producer wages, net take-home wages, or taxable wages received by the employee (that is, labor costs not including the employer share of social security contributions, etc.). These alternatives allow the replacement rate to be defined in a number of different ways: two definitions employed in this note are (i) the gross replacement rate, ρg, which is the ratio of after-tax pension benefits to the gross producer wage, and (ii) the taxable wage replacement rate, ρt, which is the ratio of after-tax pension benefits to the taxable wage received by the employee. It should be noted that the definition of wages used in the replacement rate must be consistent with that used in measuring the share of labor income in GDP.
A second point to note is that b is the average benefit received by a pensioner, and not the average pension, because some individuals receive more than one pension. Thirdly, the system dependency ratio differs from the standard demographic dependency ratio (defined as the ratio of the number of individuals above the retirement age to the number of working age individuals) due to the existence of less than 100 percent labor force participation, unemployment, early retirement, and non-old-age pensioners.
Table 3.1 presents data on pension spending as a proportion of GDP for the three countries in 1993, broken down according to equation (1). The replacement rate and share of labor in output are presented on both gross and taxable wage bases (with Polish pensions measured after-tax), and both demographic (δd) and system (δs) dependency ratios are shown.
Pension Data, 1993
(In percent)
Includes pension, health, and other social insurance contributions.
Defined as the ratio of total pension expenditures to gross payroll tax base.
Pension Data, 1993
(In percent)
| Czech Republic | Hungary | Poland | ||||
|---|---|---|---|---|---|---|
| Dependency ratios | ||||||
| Demographic (δd) | 35.5 | 36.0 | 35.0 | |||
| System (δS) | 47.4 | 66.0 | 69.0 | |||
| Gross Outlay Measures | ||||||
| Replacement rate (ρg) | 37.5 | 21.3 | 59.0 | |||
| Labor’s share of GDP(αg) | 49.0 | 70.4 | 30.5 | |||
| Taxable wage outlay measures | ||||||
| Replacement rate (ρt) | 51.0 | 39.0 | 85.6 | |||
| Labor’s share of GDP (αt) | 36.0 | 38.5 | 25.7 | |||
| Expenditure share of GDP (σ) | 8.7 | 9.9 | 12.4 | |||
| Financing | ||||||
| Social security contribution rate 1/ | 48.0 | 54.0 | 37.0 | |||
| of which: | ||||||
| Employee share | 12.0 | 10.0 | --- | |||
| Employer share | 36.0 | 44.0 | 37.0 | |||
| Pension contribution rate | 27.2 | 30.5 | 37.0 | |||
| of which: | ||||||
| Employee share | 6.8 | 6.0 | --- | |||
| Employer share | 20.4 | 24.5 | 37.0 | |||
| Gross effective pension tax rate | 18.4 | 21.3 | 25.5 | |||
| Gross balanced budget tax rate 2/ | 16.9 | 23.8 | 36.3 | |||
Includes pension, health, and other social insurance contributions.
Defined as the ratio of total pension expenditures to gross payroll tax base.
Pension Data, 1993
(In percent)
| Czech Republic | Hungary | Poland | ||||
|---|---|---|---|---|---|---|
| Dependency ratios | ||||||
| Demographic (δd) | 35.5 | 36.0 | 35.0 | |||
| System (δS) | 47.4 | 66.0 | 69.0 | |||
| Gross Outlay Measures | ||||||
| Replacement rate (ρg) | 37.5 | 21.3 | 59.0 | |||
| Labor’s share of GDP(αg) | 49.0 | 70.4 | 30.5 | |||
| Taxable wage outlay measures | ||||||
| Replacement rate (ρt) | 51.0 | 39.0 | 85.6 | |||
| Labor’s share of GDP (αt) | 36.0 | 38.5 | 25.7 | |||
| Expenditure share of GDP (σ) | 8.7 | 9.9 | 12.4 | |||
| Financing | ||||||
| Social security contribution rate 1/ | 48.0 | 54.0 | 37.0 | |||
| of which: | ||||||
| Employee share | 12.0 | 10.0 | --- | |||
| Employer share | 36.0 | 44.0 | 37.0 | |||
| Pension contribution rate | 27.2 | 30.5 | 37.0 | |||
| of which: | ||||||
| Employee share | 6.8 | 6.0 | --- | |||
| Employer share | 20.4 | 24.5 | 37.0 | |||
| Gross effective pension tax rate | 18.4 | 21.3 | 25.5 | |||
| Gross balanced budget tax rate 2/ | 16.9 | 23.8 | 36.3 | |||
Includes pension, health, and other social insurance contributions.
Defined as the ratio of total pension expenditures to gross payroll tax base.
There is a wide degree of variation among the countries in many respects. In 1993, the share of GDP devoted to pensions differed significantly: it was highest in Poland and least in the Czech Republic. The relatively low share in the Czech Republic is explained mainly by that country’s low system dependency ratio, which in 1993 was two thirds of that for Poland 1/; the Czech Republic’s gross replacement rate was significantly less than Poland’s, but this effect was offset by the significantly higher gross share of labor in output. Hungary spent a smaller proportion of GDP on pensions than Poland despite having a similar system dependency ratio. This difference was due to the fact that, although the gross share of labor in output in Hungary was more than double of that in Poland, the gross replacement rate was only about one third of Poland’s.
b. Financing
Pension benefits are financed through payroll taxes and from the general government budget. Pension payroll taxes are part of general social security taxes, and in the Czech Republic and Hungary legal liability for payment is split between employers and employees. Because the allocation of legal tax liability between employee and employer differs among countries, it is useful for comparative purposes to calculate also the effective pension tax rates on gross labor income (including the employers’ contributions). Since the payroll taxes are proportional, the effective tax rates represent both average and marginal rates, and provide information on both the relative burden and distortionary effects of the payroll tax. 2/ Finally, because the annual cash flow positions of the systems differ, it is useful to calculate a “balanced budget” pension tax rate for each country. This is the ratio of average pension expenditures to the average gross wage, and represents the effective gross payroll tax that would be required to finance the system on a cash flow basis.
The fact that the Czech pension system ran a surplus in 1993 means that the balanced budget tax rate was lower than the effective statutory tax rate. The Hungarian system had a deficit of about 12 percent of pension revenues, and the deficit of the Polish system was over 40 percent of revenues. While the effective tax rate of the Polish system is higher than that of the other two countries, the low share of labor in GDP means that it is not sufficiently high to cover current pension costs. In fact, the tax rate required to balance the Polish pension system may be higher than the 36 percent reported in Table 3.1, since this figure is calculated assuming that the tax base is held fixed. However, with such a high marginal tax rate (in addition to personal income taxes), reduced labor supply and increased informal sector activity are likely to reduce the base and necessitate even higher rates.
3. Spending projections
For a useful analysis of the financial position of the pension systems, the description of current expenditures and revenues needs to be supplemented with information on expected longer-term trends. Such projections are fraught with uncertainty, but rough calculations at least are required, based on demographic projections, assumed growth rates, and policy stances. A full analysis for all countries is beyond the scope of this note, and this section only presents some basic data on expected future behavior. The next section presents more detailed projections for the Czech Republic under various reform scenarios.
Table 3.2 presents data on projected demographic trends and pension spending as a proportion of GDP. The first point to note is that while all countries currently have similar demographic dependency ratios (about 35 percent), those of the Czech Republic and Poland are projected to reach significantly higher levels than that of Hungary by 2020. However, despite its more favorable demographic projection, Hungary’s system dependency ratio increases proportionately slightly more than Poland’s. 1/ Indeed, by 2020, the countries’ system dependency ratios are projected to converge somewhat relative to their current distribution. This is due mainly to the fact that the Czech unemployment rate is assumed to increase above its current level of 3.3 percent, while unemployment is currently significantly higher in Hungary and Poland, and is projected to fall over the long term.
Pension Projections, 2020
(In percent)
Pension Projections, 2020
(In percent)
| Czech Republic | Hungary | Poland | ||
|---|---|---|---|---|
| Dependency ratios | ||||
| Demographic (δd) | 45.8 | 40.0 | 48.0 | |
| System (δs) | 70.2 | 75.0 | 77.0 | |
| Gross replacement rate (pg) | 33.1 | 21.3 | 59.0 | |
| Expenditure share of GDP (σ) | ||||
| Labor’s share of GDP constant | 12.5 | 11.2 | 13.9 | |
| Labor’s share of GDP two thirds | 15.5 | 10.6 | 35.8 | |
Pension Projections, 2020
(In percent)
| Czech Republic | Hungary | Poland | ||
|---|---|---|---|---|
| Dependency ratios | ||||
| Demographic (δd) | 45.8 | 40.0 | 48.0 | |
| System (δs) | 70.2 | 75.0 | 77.0 | |
| Gross replacement rate (pg) | 33.1 | 21.3 | 59.0 | |
| Expenditure share of GDP (σ) | ||||
| Labor’s share of GDP constant | 12.5 | 11.2 | 13.9 | |
| Labor’s share of GDP two thirds | 15.5 | 10.6 | 35.8 | |
If replacement rates and the share of labor in GDP remain fixed, then changes in the proportion of GDP devoted to pension expenditures are driven solely by changes in the system dependency ratio, according to equation (1). In this case, the Czech Republic experiences a larger increase in the share of GDP devoted to pension spending than Hungary or Poland, due to the proportionately larger deterioration in the system dependency ratio.
However, as the economies restructure, the share of labor in output is likely to approach western European levels over time. For illustrative purposes, the table reports projections of pension spending assuming the gross share of labor in output in 2020 is two thirds (while retaining a constant replacement rate). It is clear that, since the Czech Republic and Poland currently have relatively low labor shares, spending as a proportion of GDP in these countries will increase more if the share of labor in GDP increases to two thirds. The effect is particularly noticeable for Poland, where the labor share is currently very low.
This extreme development of Polish spending is unlikely to occur in practice, since if the share of labor in GDP more than doubles, the replacement rate is unlikely to remain fixed. The political process by which benefit levels are set is then relevant, and the replacement rate is unlikely to be a useful parameter to keep constant. 1/
The comparison, however, suggests that, from having the lowest level of spending relative to GDP, the Czech Republic, without effective reform of its pension system, could quickly converge to the levels of Hungary and Poland. The low level of spending and the lack of financial difficulties at present, thus should not be seen as an excuse not to proceed with reform.
4. Long-term pension projections in the Czech Republic 2/
The pension system in the Czech Republic is seen by advocates of reform to be inefficient, inequitable, and in danger of insolvency in the not too distant future. The reforms aim to both improve the medium- to long-term financial position of the system, and to reduce the distortions it imposes on individuals’ decisions about work, retirement, and saving. This section looks at some financial projections and indicators of the system as it now stands and under various reform scenarios, abstracting for the moment from individual responses to the changes, which may be of second order importance.
There are three main components of the reforms: (i) unification of the treatment of individuals in different professions; (ii) increasing the retirement age; and (iii) changing the calculation of benefits. The first component can be expected to have long-run efficiency enhancing effects through its impact on the allocation of labor in the economy, but is largely ignored in the following discussion. The second component has clear implications for the development of dependency ratios and, thus, expenditures. The third component--changing the benefit formula--has two effects. First, by possibly linking benefits more closely to contributions, it may reduce labor supply distortions. Thus, while the average benefit could be left unchanged, changing the distribution of benefits across individuals with different wages (generally in favor of the better off) is argued to increase labor supply, earnings, and total saving. The second effect is that, to the extent that the change in the benefit formula lowers the average nominal pension benefit, expenditures are reduced directly. This can be thought of, in general, as the effect on the replacement rate. Note that under conditions of inflation and real wage growth, making pension benefits more directly related to wages affects the ratio of new pensions to wages, and hence the average replacement rate (see below). The main focus of the analysis that follows is thus on component (ii) (the retirement age effect) and the second part of component (iii) (the replacement rate effect).
a. Essence of the reforms 1/
The increase in the retirement age will reduce the dependency ratio to a level below its current trend path. 2/ This has the impact of increasing the number of working individuals, and hence the social security contribution base, as well as reducing the number of pensioners receiving benefits.
Currently, pension benefits bear little relationship to lifetime earnings and contributions. In practice, nominal wages from the last 5 working years are averaged (without indexation), and then “reduced” according to a piecewise linear increasing but sharply concave transformation function. The brackets of this function are set in nominal terms (as opposed to being, for example, some multiple of the average wage), and most individuals end up on the flattest part. Pension benefits are proportional to the reduced average wage, and are thus virtually independent of average real lifetime wages. As well as the obvious distortions that such a scheme may impose on economic decision-making (which are mostly ignored here), this arrangement means that, without explicit indexation rules, benefits are essentially constant in nominal terms and in times of inflation fall in real terms. To alleviate this problem, newly awarded pensions are adjusted upwards on an ad hoc basis so that they are equal to benefits received by existing pensioners. Existing pensions are indexed, more or less, to the Consumer Price Index (CPI), so the real value of both average and new pensions remains close to constant over time. In an environment of real wage growth, this clearly means that the replacement rate, measured as the ratio of average pension benefits to average wages, is on a downward trend.
If, on the other hand, pensions were linearly related to past wages, then even if existing pensions were indexed only to the CPI, the replacement rate would not fall, assuming a constant dependency ratio. To see this, note that while the real benefits of existing pensioners would fall relative to average wages, each year a new cohort of just-retired individuals enters retirement (with a pension effectively indexed to the real wage in the first year) and the last cohort of retired individuals dies off (removing the group contributing least to the average benefit).
b. Long-term projections
Section 3 presented long-term projections of the Czech, Hungarian, and Polish pension systems assuming a constant replacement rate. This section presents projections for the Czech system under two sets of alternative assumptions. The first is that pension benefits increase only with CPI inflation (and not with wages), so that the replacement rate falls over time. Clearly, expenditures will increase more slowly under such a scenario, and the share of GDP devoted to pensions will actually fall. The second scenario is that of the current reform proposal. Under this proposal, the current (1995) replacement rate will be retained for 10 years, after which it will be allowed to fall (as the role of private savings vehicles is envisaged to increase).
The results of these exercises are presented in Table 3.3. Also included are the latest estimates of the projected outturn in 1995. If current pension rules are followed, and in particular, if pension benefits increase at the rate of CPI inflation, the share of GDP devoted to pensions by 2020 will be just 6.7 percent (second column of Table 3.3). 1/ This is despite the fact that the system dependency ratio is projected to be over 70 percent, and occurs because the gross replacement rate is only about 18 percent. The contribution rate (as a proportion of taxable wages) necessary to finance the system is just 19 percent.
Alternative Projections of the Czech Pension System 1/
(In percent)
Projections to 2020 assume the share of wages in GDP is constant at its 1995 level.
Alternative Projections of the Czech Pension System 1/
(In percent)
| 1995 Projection | 2020 CPI Indexation | 2020 Wage Indexation | 2020 Reform | ||
|---|---|---|---|---|---|
| Replacement rates | |||||
| percent of taxable wage (ρt) | 45.0 | 24.1 | 45.0 | 36.6 | |
| percent of gross wage (ρg) | 33.1 | 17.7 | 33.1 | 26.9 | |
| System dependency ratio (δs) | 53.1 | 70.2 | 70.2 | 62.1 | |
| Expenditure share of GDP (σ) | 9.5 | 6.7 | 12.5 | 8.9 | |
| Required contribution rate | |||||
| (on taxable wages) | 26.8 | 19.0 | 35.5 | 25.8 | |
Projections to 2020 assume the share of wages in GDP is constant at its 1995 level.
Alternative Projections of the Czech Pension System 1/
(In percent)
| 1995 Projection | 2020 CPI Indexation | 2020 Wage Indexation | 2020 Reform | ||
|---|---|---|---|---|---|
| Replacement rates | |||||
| percent of taxable wage (ρt) | 45.0 | 24.1 | 45.0 | 36.6 | |
| percent of gross wage (ρg) | 33.1 | 17.7 | 33.1 | 26.9 | |
| System dependency ratio (δs) | 53.1 | 70.2 | 70.2 | 62.1 | |
| Expenditure share of GDP (σ) | 9.5 | 6.7 | 12.5 | 8.9 | |
| Required contribution rate | |||||
| (on taxable wages) | 26.8 | 19.0 | 35.5 | 25.8 | |
Projections to 2020 assume the share of wages in GDP is constant at its 1995 level.
Such a low replacement rate is not likely to be considered satisfactory, but maintaining it at its 1995 level of 45 percent with no other changes will lead to increasing expenditures as a proportion of GDP, reaching 12.5 percent in 2020 (third column of Table 3.3). The required contribution rate then would be over 35 percent of taxable wages. By increasing the retirement age and indexing pensions to wage developments for the next 10 years, the reforms are projected to reduce the growth of spending, so that by 2020, the proportion of GDP devoted to pensions will be 8.9 percent (fourth column of Table 3.3). The system dependency ratio is expected to fall to 62 percent from its no-reform projected level of 70 percent. The contribution rate (as a proportion of taxable wages) required to finance the system will then stabilize at about 25 percent.
An important assumption underlying the projections is that the share of wages in GDP will remain constant. However, this share has been rising even over the last few years (increasing from 36 percent in 1993 to 38 percent in 1994), so a further increase towards levels of western Europe may be expected. Maintaining the projected replacement rates and demographic developments of the reform proposal, an increase of the labor share of GDP to two thirds would raise the share of GDP to 11 percent.
References
Central Statistical Office, Polish National Accounts by Institutional Sectors 1991-1992 (Warsaw, Poland, April 1994).
Fehér, C. R. Palacios, and D. Vittas, “Reforming Public Pensions in Hungary,” (unpublished paper, World Bank, 1994).
International Monetary Fund, Czech Republic--Recent Economic Developments, SM/94/194 (Washington, 7/20/94).
Kopits, G., R. Holzmann, G. Schieber, and E. Sidgwick, Social Security Reform in Hungary, Fiscal Affairs Department (Washington: International Monetary Fund, October 1990).
Preker, A., “Republic of Poland: The Pensions System: Old Age, Survivors, and Disability (OASD),” (unpublished paper, World Bank, 1994).
Perraudin, W.R. and T. Pujol, “A Framework for the Analysis of Pension and Unemployment Benefit Reform in Poland,” Staff Papers, International Monetary Fund (Washington), Vol. 41, No. 4, (1995), pp. 643-74.
World Bank, Hungary: Structural Reforms for Sustainable Growth, (unpublished Country Economic Memorandum, October 1994).
Prepared by William Jack.
A detailed description of the pension systems in the three countries can be found in International Monetary Fund (1994) for the Czech Republic, World Bank (1994) for Hungary, and Perraudin and Pujol (1995) for Poland.
This distinction is irrelevant in the Czech Republic and Hungary where pensions are currently untaxed. In Poland, pensions were made taxable in 1992.
A major cause of this difference appears to have been the number of individuals receiving disability pensions. Although 1993 data are not available for the Czech Republic, in 1992, 17 percent of Czech pensions were for disability, while the corresponding number for Poland was 37 percent. The lower unemployment rate in the Czech Republic also contributed to the divergence.
Of course, the net tax burden and net distortions of the entire tax system are of ultimate concern, but the effective tax rate reported above provides information on the contribution of the payroll tax to these variables.
The reasons for this are not apparent from the available sources, although differences in projected unemployment, early retirement, and disability pension take-up may have contributed.
That is, if a greater share of personal income is received in the form of wages, a lower replacement rate is sufficient to maintain a given relationship between total pre- and post-retirement income. It should be noted, however, that assuming a constant share of wages in GDP, Perraudin and Pujol’s projections imply that under current pension rules, the gross replacement rate will increase to about 70 percent by 2020, and that spending on pensions (not including unemployment benefits) will be about 16 percent of GDP. The figures reported in Table 3.2 use Perraudin and Pujol’s projections, but assume a constant replacement rate for purposes of comparison with the Czech Republic and Hungary.
Data reported in this section were provided by the Ministry of Labor and Social Affairs, Czech Republic, or estimated on the basis thereof.
For a more thorough discussion of the details of the Czech reform proposals, see International Monetary Fund (1994).
The standard retirement age is increased, on a phased-in basis, by 2 years to 62 for men, and by 4 years to 57-61 for women (depending on the number of children reared).
This projection assumes that the share of wages in GDP is constant.