Information about Europe Europa
Journal Issue

Belgium: Selected Issues

International Monetary Fund
Published Date:
March 2003
  • ShareShare
Information about Europe Europa
Show Summary Details

II. Fiscal Strategies for Population Aging11

A. Introduction

31. Like many industrialized nations, Belgium faces the problem of population aging in the coming decades. Its old-age dependency ratio is expected to rise from the current level of 40 percent to more than 60 percent in the next thirty years.12 According to a study by the OECD, pension expenditures in Belgium could rise by 3.3 percent of GDP during this period, slightly higher than the EU average (Table II.1). Cost of health care could add another 3 percent of GDP to the fiscal cost of aging in Belgium (Englert, 2002). While the risk of unsustainable pension financing may seem low at this stage—thanks to the large primary surpluses built up over the past ten years—Belgium faces an important fiscal challenge given its high level of public debt, which, at 104 percent of GDP in 2002, is one of the highest in the EU.

Table II.1.Projections of Pension Expenditure in the EU Countries(In percent of GDP)
Changes during
EU average10.410.411.513.013.613.32.62.9
Source: OECD (2001).
Source: OECD (2001).

32. Against this background, this chapter examines long-term fiscal strategies for meeting the fiscal burden of population aging in Belgium. The main conclusions are as follows:

  • The fiscal cost of population aging is projected to rise by 6.2 percent between now and 2050. This will be only partially offset by possible savings from lower spending on unemployment benefits. The fiscal pressures of population aging and the high debt-GDP ratio argue for a sustained surplus position.
  • Compared to an alternative of a balanced budget, a strategy of surpluses and swift debt reduction would guard against unforeseen shocks and the possibility of higher-than-forecast pension outlays. In addition, it would free up resources, stabilize the tax burden over time, and avoid the need to reduce pension generosity.
  • Delay in adjustment is costly, in that higher taxes are required to fully fund aging costs through a reduction in interest payments.
  • Further reforms of the pension system—specifically with regard to eligibility requirements for early retirement—are a complement to debt reduction, in dealing with fiscal cost of population aging. They would contain the budgetary cost of aging and free resources for alternative uses, notably tax cuts for the business sector to boost employment growth.

33. The Chapter is structured as follows: Section B presents estimates of the fiscal costs of population aging in Belgium, and discusses their sensitivity to underlying assumptions. Section C lays out strategies for meeting the fiscal costs of aging. Specifically, a strategy of swift debt reduction is discussed to see what budget surplus would be needed to reduce debt and interest payments rapidly enough to finance these costs. Alternative policy options are also discussed to show the consequences of delayed fiscal adjustment. Finally, reforms of the public pension and health care are discussed as a complement to fiscal consolidation. Section D concludes.

B. Fiscal Costs of Population Aging

Macroeconomic and demographic assumptions13

34. Main macroeconomic and demographic assumptions are presented in Table II.2. Macroeconomic assumptions include average labor productivity growth (dYt/dLt) of 1.75 percent (reflecting gains from assumed structural reforms), and real GDP growth of 1.9 percent on average during 2000–50. Real wages are assumed to grow in line with labor productivity and employment growth (dLt/dt) is equal to the difference between output growth (dYt/dt) and labor productivity growth (dYt/dLt). Furthermore, the real interest rate (rt) and the inflation rate (πt) are exogenously assumed to be fixed at 4 percent and 1.7 percent, respectively.

35. Demographic assumptions—namely population growth, evolution of age structure, life expectancy, and total fertility rate—are based on projections by the Belgian Federal Planning Bureau (FPB). Specifically, the total fertility rate, currently 1.6, would gradually rise to 1.75 in 2050; the life expectancy at birth would increase by 9 years for men and 7.5 years for women between now and 2050; and net immigration would be maintained at about the current level. These assumptions imply a slightly rising population over the 2000-2050 period, and a sharp rise of 29 percentage points in old-age dependency ratio (defined as those aged 60 years and over as a share of those aged 20–59 years). The rise in old-age dependency ratio is mainly due to increasing life expectancy for both Belgian men and women over the next 50 years, and is similar in magnitude to the average expected in EU countries.

Table II.2.Key Macroeconomic and Demographic Assumptions
Macroeconomic assumptions (in percent)Average
Real GDP growth3.
Labor productivity growth1.
Employment growth1.80.40.1-0.2-
Unemployment rate10.
Demographic assumptionsChanges
Population (millions)10.25310.52010.710.88810.9540.60.7
(a) 0-19 years2.
(b) 20-59 years5.
(c) 60+ years2.
Old-age dependency ratio (c)/(b) (percent)40.143.753.062.968.722.828.5
Key coefficients of the pension system:
Pensioners (millions)
Average pension/GDP per worker16.215.314.914.413.5-1.7-2.7
Pensioners (% of employed)54.153.663.676.987.922.833.8
Employment rate (age group 19-64) 1/59.061.563.064.865.65.86.6
Sources: Belgian Federal Planning Bureau; and Fund staff projections.

In percent of active population aged 19-64.

Sources: Belgian Federal Planning Bureau; and Fund staff projections.

In percent of active population aged 19-64.

36. One important assumption is that the employment rate will rise during the next 50 years, particularly for women and older people of working age (Table II.2). This increase would appear to imply significant structural reforms, notably pension reforms to reduce incentives for early retirement (more below).

Projections of pension expenditures

37. The projection of pension expenditure follows the approach of European Commission (EC, 1996), in which the ratio of pension expenditure (Ct) to GDP is expressed as the product of the following four ratios:

where POPp refers to the population at pensionable age and is measured by population of 60 years and up; POPw is the population at working age, measured by those with ages of 20-59 years; NP is the number of pensioners and its projections are taken from the official study on aging; and L is employment.

38. Therefore, the dynamics of pension expenditures can be decomposed into the dynamics of four ratios. The first (the old-age dependency ratio) represents the pure demographic component of expenditure dynamics, and is derived based on the demographic assumptions.14 The second (the eligibility ratio) is influenced by both demographic and legislative factors. The third (the transfer ratio) depends on legislative as well as economic factors, such as pension indexation and productivity growth. In this analysis, this ratio is used to capture the effect of pension reforms. The fourth ratio (the employment ratio) captures the influence of economic factors. In general, the eligibility and the transfer ratios move smoothly over time, since pension reforms usually are implemented gradually and the full effects of changes in pension coverage and benefit improvement are usually felt only after several decades (Chand and Jaeger, 1996).

39. Projections of pension expenditures are summarized in Table II.3 and Figure II.1.15 Pension expenditures are projected to rise by 3.2 percentage points of GDP between 2000 and 2050, slightly higher than the average increase in the EU countries during the same period. Projections of health-care spending and other social spending are based on the projections by the FPB. Specifically, health-care spending (including long-term care for the elderly) is projected to increase by 3.1 percent of GDP during this period, implying that age-adjusted spending will rise much faster than per capita GDP. Other social spending (mainly unemployment benefits) is projected to decline by 1.9 percent of GDP during 2000-50, as the unemployment rate is assumed to fall substantially by 2050, reflecting presumed labor-market reforms.16

Table II.3.Projections of Fiscal Costs of Aging(Base case, in percent of GDP)
Change during
Fiscal cost of aging:22.321.323.
of which: Public Pensions8.88.29.511.111.811.83.1
Public health care6.
Unemployment benefits & other7.
Implied ratios:
(a) Old-age dependency ratio0.410.440.530.630.660.690.28
(b) Eligibility ratio0.940.900.930.991.041.030.10
(c) Transfer ratio16.215.314.914.413.913.5-2.66
(d) Employment ratio1.
Sources: Belgian Federal Planning Bureau; and Fund staff projections.
Sources: Belgian Federal Planning Bureau; and Fund staff projections.

40. The relatively modest impact of aging on pension expenditure (particularly when compared to Germany, France, or the Netherlands) is mainly due to the declining transfer ratio (i.e., the ratio of average pension per beneficiary to average GDP per worker). If the transfer ratio were unchanged, ceteris paribus, pension expenditures could rise by 5.1 percent of GDP between 2000 and 2050, about 2 percentage points higher than in the base case.17 The transfer ratio has declined since 1990, mainly due to past pension reforms, most importantly because pensions are indexed to prices, not wages (Bogaert, 2000). As a result, real wage growth has outpaced the growth of real average pension payments (see Box II.1). Although each year the government can apply a “welfare adjustment” to allow pension payments to catch up with the national real income growth, this has been rare. The projections in Table II.3 assume a small welfare adjustment of about 0.5 percent of GDP during the period of 2000–50. Under this assumption, by 2050, the accumulated growth of real average pension would be about 40 percent lower that that of real wage growth (Figure II. 1). However, this assumption of a small welfare adjustment could well be on the low side, since, with the number of pensioners as a share of the employed expected to increase rapidly from 54 percent today to 88 percent in 2050, the political pressure for a much larger welfare adjustment would certainly rise.

Figure II. 1.Belgium: Projection of Public Pension Expenditures, 2000-2050

Sources: Belgian Federal Planning Bureau; and Fund staff estimates.

Box II.1.Pension System in Belgium

The pension system in Belgium has three pillars. The first is a mandatory pay-as-you-go public pension scheme, accounting for about 80 percent of total pension expenditure in Belgium. It provides general pensions for wage-earners and self-employed, and is financed by pension contributions (80 percent) as well as transfers from the central government (20 percent). The second pillar includes only private pension schemes, accounting for about 20 percent of total pension expenditure. These private schemes cover nearly 1/3 of employees in Belgian private sector and are not mandatory (Bogaert 2000). Third is individual retirement saving.

The benefit levels of Belgian public pension system are relatively low compared to the EU average, reflected in its low average replacement ratios (see Table below). In addition to the relatively low contribution rates in Belgium, the low average replacement ratio in Belgium is a result of the indexation rule as well as the prevailing early retirement. Although the legal pensionable age is 65, average early retirement age in Belgium, at 59 for men and 55 for women, is among the lowest in the EU countries.

Under the general scheme, a full pension is calculated based on the average wage during a 45-year career, which are indexed to current prices, multiplied by the replacement rate (60 and 70 percent for families with two and one income, respectively). There are wage ceilings applied to pension calculation, which are also indexed to prices. With pensions indexed to prices, the higher growth of wages have led to declining replacement ratios.

Relative benefit levelsIndexation rules Pensions/Ref. SalaryReplacement ratio Statutory/AverageNumber of years in reference salaryContribution rates (% of ave. wages)
BelgiumLowPrices/prices w. adj.60/40.74516.36
GermanyMediumWages/wages48.9/48.9Life long earning19.1
ItalyHighPrices/prices w. adj.80/69.15/10/life long32.7
NetherlandsLowWages/Flat rate/31Unrelated
Source: EC (2002)
Source: EC (2002)

Belgium’s pension system has gone through several reforms since the early 1980s, to ensure sustainable and adequate pensions for its population. In 1983, the ceiling on wages used in the computation of pension benefits was lowered, and was kept constant in real terms until 1996 (after which it was indexed to wages), contributing to lower average replacement ratios in Belgium. Although the government could take a “welfare adjustment” each year to allow the pensioners to benefit from the increase in real income, this was rare. Starting 1996, the retirement age for women is being gradually raised (from 60 to 65 years) and the eligibility for early retirement was restricted.

C. Strategies for Meeting the Cost of Population Aging

41. Fiscal costs of aging can be met by a swift reduction of the national debt to reduce the interest payments. The Ecofin Council considers the strengthening of public finances through rapid reduction of public debt a central element of its three-pronged strategy (which also include raising employment rates and reforming pension and health care) in dealing with budgetary cost of aging in its member states (EC, 2002). This strategy was also proposed by the Belgian High Finance Council for the period 2000–30 (HFC, 2002).

42. Specifically, this policy strategy (scenario A) envisages a significant reduction in debt-GDP ratio from 108 percent in 2001 to the Maastricht level of 60 percent of GDP by 2013, followed by further reductions to 23 percent by 2030 and 12 percent by 2050 (Figure II.2). Under this scenario, the estimated aging-related increase in pensions and health-care between 2000 and 2050 (6.2 percent of GDP) would be covered entirely by savings from interest payments resulting from lower public debt, without the need to raise taxes. Specifically, interest payments on the public debt would fall by 6.2 percent of GDP during 2000-50, with most of this reduction (5.5 percent) taking place during 2000–30.

43. Such policy strategy would require sustained budget surpluses and a gradual decline in real expenditure growth in the long run (Figure II.2). The budget surplus would need to rise to 0.7 percent of GDP in 2005 (as foreseen by the Stability Program for 2002–2005), and to 1.5 percent of GDP in 2010. This surplus would need to be maintained until 2015, after which surpluses could decline gradually, reaching a balanced budget in 2030. A balanced budget for 2030–50 would further lower the debt-GDP ratio to 12 percent by 2050.18 The growth of real non-age related primary expenditure would decline from an average of 2.5 percent in 2000–30 to about 1.5 percent in 2030–50.

44. The swift debt reduction would guard against unforeseen shocks and the possibility that pension outlays could prove higher than now forecast. As debt dynamics are highly sensitive to changes in interest rates (see figure), lower public debt can help to reduce the vulnerability of the Belgian public finances to interest rate shocks. Moreover, pension outlays may turn out higher than now forecast. For example, if “welfare adjustments” fully offset the endogenous fall in the transfer ratio, pension expenditure could rise by 5.1 percent of GDP during 2000-50, about 2 percentage points higher than in the base case. With no tax change and in the absence of reforms to pension and medical plans, a bigger budget surplus would then be needed.

Debt Dynamics

(In percent of GDP)

Source: Fund staff estimates.

Figure II.2.Belgium: Fiscal Projections, 2000-2050

Source: Fund staff estimates.

45. Delaying adjustment is costly, in that to fully fund aging costs through debt reduction requires higher taxes later on (Figure II.3).19 Based on the generational accounting framework, established by Gokhale and Kotlikoff (Kotlikoff, 2001), the government’s intertemporal budget constraint is:

where Tt,k stands for tax per capita in year t applied to population born in year k (k>t implies future generation) and Nt,k stands for population in year t born in year k. If the intertemporal budget constraint were not met, revenues would need to rise. The permanent adjustment required to close the gap is given by (r-g)(L-T)/GDP, where r and g refer to real interest rate and real growth rate. Hence, the later the adjustment takes place, the larger the gap (L-T) is, and the greater the needed permanent adjustment.

46. This is illustrated in Figure II.3. In contrast to scenario A, an alternative of running a balanced budget through 2030 (scenario B), which results a smaller fall in the debt-GDP ratio, allows a lower tax rate until about 2020, but requires a higher one thereafter in order to converge to the debt path under the scenario A. The cost of delaying in adjustment is further demonstrated with other two scenarios, where delaying adjustment till 2013 (scenario C) and 2023 (scenario D) would lead to permanently higher tax rates after 2013 and 2023, respectively.20

Figure II.3.Belgium: Adjustment Scenarios, 2003-2050

Source: Fund Staff estimates.

47. Further reform of the pension system, specifically reductions of the built-in incentives for early retirement, would be necessary to complement the fiscal consolidation effort. By European standards, the generosity of public pensions in Belgium is not high, nor is the country’s implicit pension debt (EC, 2002; and Holzmann et. al. 2001), thanks to past pension reforms (see Box II. 1). However, Belgium—among France, Italy, Luxembourg and Austria—faces the biggest challenge from lower activity rates, with employment rates for people in the 55–65 age group between 25 and 30 percent (Figure II.4). Although the legal pensionable age is 65, the average early retirement age in Belgium is among the lowest in the EU countries. Among other structural factors, low activity rates are found to be correlated in particular with the eligibility requirement for early retirement and/or disability program (EC, 2002). According to the same EC study, raising the effective retirement age by one year could lead to a reduction of pension spending up to 1.1 percent of GDP. Furthermore, pension reform that discourages early retirement would increase employment and enhance growth.

48. Fiscal costs of aging can also be contained by controlling health-care spending. Health-care costs increased in real terms by 4 percent a year in the past five years, and are set to rise by 5 percent in the 2003 budget (reflecting partly a broadening of coverage). New control mechanisms have recently been introduced, but even so the cost pressure from technological advances and population aging is likely to increase. The projection for healthcare cost, which is based on the study by the FPB, assumes that health-care spending will increase by 3.1 percent of GDP between 2000 and 2050, implying an average real growth of 2.7 percent. To reduce the growth of real health care spending from the current average of 4-5 percent to this level would certainly require policy measures.

49. Pension and health care reforms would free budgetary resources for alternative uses, notably tax cuts to boost employment. In an extreme case where all the savings from interest payments (6.2 percent of GDP) were used to reduce federal taxes, the revenue of the federal government (including social security contributions) could fall by about 5 percentage points of GDP between 2002 and 2050.21 This reduction of tax burden could, for example, reflect payroll tax reductions of 10 percent between 2002 and 2050, assuming that compensation of employees as a share of GDP is constant at its average level during 1990-2001 (about 52 percent). Based on an average payroll tax rate of 25 percent in 2001, this would mean a reduction of the payroll tax rate to 15 percent.

Figure II.4.Employment Rate For Age Group 55-65 in Selected EU Countries in 2001

(In percent of active population 55-65 years old)

Source: EC (2002).

D. Concluding Remarks

50. The fiscal cost of population aging is projected to rise by 6.2 percent by 2050, only partially offset by possible savings from lower spending on unemployment benefits. The fiscal pressures of population aging and the high debt-GDP ratio argue for a sustained surplus position. A policy strategy to prepare for population aging by rapidly reducing the public debt, as proposed by the High Finance Council, seems sensible. This strategy would guard against unforeseen shocks and the possibility of higher-than-forecast pension outlays. In addition, it would free up resources, stabilize the tax burden over time, and avoid a reduction in pension generosity. Delay in adjustment is costly, in that to fully fund aging costs through reduction in interest payment requires higher taxes later on. Reforms of the pension system—specifically with regard to eligibility requirements for early retirement—would also contain the budgetary cost of aging and free resources for alternative uses, notably tax cuts for the business sector to boost employment growth.


    BogaertHenry2000Long-term Fiscal Sustainability in Belgium,Banca d’ltalia Research Department Fiscal Sustainability Workshop.

    • Search Google Scholar
    • Export Citation

    ChandS. and A.Jaeger1996Aging Populations and Public Pension Schemes,IMF Occasional Paper 147.

    EnglertM.N.FasquelleM.FestjensM.LambrechtM.SaintrainC.Streel and S.Weemaes2002Perspectives Financières de la Sécurité Sociale 2000-2050: Le Vieillissement et la Viabilité du Système Légal des Pensions,Federal Planning Bureau Planning Paper No. 91Belgium.

    • Search Google Scholar
    • Export Citation

    European Commission2002Reform Challenges Facing Public Pension Systems: The Impact of Certain Parametric Reforms on Pension Expenditure,EC Economic Policy CommitteeJuly.

    • Search Google Scholar
    • Export Citation

    European Commission1996Public Pension Expenditure Prospects in the European Union: A Survey of National Projections.

    High Council of Finance2002Annual Report 2002Belgium.

    HolzmannR.R.Palacios and A.Zviniene2001On the Economics and Scope of Implicit Pension Debt: An International Perspective,Empirica Vol. 28.

    • Search Google Scholar
    • Export Citation

    KotlikoffL.2001Generational Policy,in the second edition ofThe Handbook of Public Economics.

    OECD2001Budgetary Challenges Posed by Aging Populations: The Impact on Public Spending on Pensions, Health and Long-term Care for the Elderly and Possible Indicators of the Long-term Sustainability of Public Finances,Economic Policy Committee Working Group on Aging.

    • Search Google Scholar
    • Export Citation

    van EwijkCasperB.KuipersH.ter ReleM.van de Ven and E.Westerhout2000Aging in the Netherlands,Netherlands Bureau for Economic Policy Analysis.

    • Search Google Scholar
    • Export Citation
11Prepared by Jianping Zhou.
12In this Chapter, the old-age dependency ratio is defined as those aged 60 years and over as a percentage of those aged 20-59 years, following the definition adopted by the Belgium’s High Council of Finance. Note that studies by the EU or the OECD usually define this ratio as those age 65 and over as a share of those aged 20-64.
13These assumptions are consistent with the “alternative” scenario in the paper by the Federal Planning Bureau: Perspectives financières de la Sécuité sociale 2000-2050, which was also presented in the 2002 annual report of the government Study Committe on Aging (Comité d’Etude sur le Vieillissement).
14However, this ratio can be affected by changes in retirement age. For example, should the government decide to increase retirement age, the old-age dependency ratio would fall.
15These projections are consistent with the “alternative” scenario in the study by the FPB, although the FPB has recently revised upward its projection of pension expenditure to 3.5 percent of GDP.
16The aging of population also implies an increasing number of pensioners relative to pension contributors. Therefore, future pension revenue would decline relative to pension spending if contribution rates are kept unchanged. In this paper, any changes in pension expenditure are net of changes in pension revenues.
17Bogaert (2000) shows a similar result.
18A deficit of 0.7 percent of GDP after 2030 would maintain the debt-GDP ratio at 23 percent during 2030–50.
19This framework was used in a study by the Netherlands Bureau for Economic Policy Analysis in analyzing budgetary consequences of aging in the Netherlands (see van Ewijk et al., 2002). It is important to emphasize that an analysis along these lines does not address the issue of the “optimal” fiscal policy, and the policy scenario discussed here would not necessarily yield an optimal debt level.
20Policy scenario A implies an immediate and permanent adjustment in 2002, and as a share of GDP tax rates are assumed to be constant in order to minimize deadweight loss. The rationale for “tax smoothing”, laid out in Barro (1979), is that the distortionary cost of a tax rises more than proportionally with the tax rate, implying that the total distortion is minimized when the tax rate is constant. Note that the efficiency gain from tax smoothing does not necessarily lead to intergenerational neutrality (i.e., equal net benefits across generations).
21Assuming that (i) non-aging-related primary expenditures grow in. line with GDP growth, hence the total primary expenditure as a share of GDP stay constant during 2002–50; and (ii) the revenues of local governments rise in line with GDP growth.

Other Resources Citing This Publication