Luxembourg: Selected Issues and Statistical Appendix

Luxembourg's impressive growth performance has been accompanied by regional specialization of production, high labor mobility, export-propelled growth, and the dominance of regional growth fluctuations. Luxembourg's public pension system faces the challenges of population aging and Luxembourg's small, open, and highly specialized economy. Luxembourg's labor market performance holds a seeming paradox, favorable labor market outcomes are coupled with rigid labor market institutions. The supervision of cross-border financial activities raises the difficult challenge of obtaining a complete, consolidated view of the operations of international banking institutions.

Abstract

Luxembourg's impressive growth performance has been accompanied by regional specialization of production, high labor mobility, export-propelled growth, and the dominance of regional growth fluctuations. Luxembourg's public pension system faces the challenges of population aging and Luxembourg's small, open, and highly specialized economy. Luxembourg's labor market performance holds a seeming paradox, favorable labor market outcomes are coupled with rigid labor market institutions. The supervision of cross-border financial activities raises the difficult challenge of obtaining a complete, consolidated view of the operations of international banking institutions.

II. Stress Testing Luxembourg’s Public Pension System9

A. The Issues

19. Luxembourg’s public pension system—and the social insurance system at large—face two policy challenges. First, as in other industrial countries, projected population aging will, all other things equal, put stress on public pension finances. And second, reflecting Luxembourg’s small, open, and highly specialized economy, the public pension system’s contribution base can be subject to possibly large and persistent shocks.

20. The premise of this chapter is that the design of Luxembourg’s overall pension system—i.e., the combination of public (first pillar) and private pension arrangements (second and third pillars)—should aim at providing a stable and sustainable source of retirement income. Given this premise, the two challenges raise different design issues. Population aging—by its nature a slow and largely predictable process—raises the issue of longer-run sustainability of the public pension system. What are the policy options that would help bridge the projected long-run financing gap in the public pension system? By contrast, the possibility of large and persistent shifts in the public pension system’s contribution base—by its nature a rapidly-evolving and largely unpredictable process—raises the issue of robustness of overall retirement income provision. What are the policy options that would help increase the stability of overall pension provision?

21. Restoring sustainability to a pay-as-you-go (PAYG) public pension system has been the subject of a large literature.10 Policy options include parametric adjustments in the basic PAYG policy levers (contribution rate, budget transfers, pension replacement rate, retirement age), administrative reforms, the building up of public pension reserves, and shifting part of pension provision to a funded second pillar. In the context of Luxembourg, the issue of sustainability of the public pension system was addressed extensively in previous staff publications.11 The present chapter focuses on the issue of robustness of Luxembourg’s pension system design in an economic environment of largely unpredictable and persistent shocks. The chapter’s main focus will be on illustrating a range of plausible longer-term scenarios for the public system’s finances.

22. Section B briefly reviews key characteristics of Luxembourg’s overall pension system. Section C lays out a simple analytic framework. Section D reports the results of stress testing exercises to illustrate the range of possible evolutions for the pension system’s dependency ratio. Section E briefly summarizes the main policy options.

B. A Sketch of Luxembourg’s Pension System

23. The provision of retirement income in Luxembourg is dominated by a mandatory pay-as-you-go (PAYG) public pension pillar.12 The two main reasons for the dominance of public pension provision are its wide coverage and high pension replacement rates. Almost all private sector workers (including cross-border workers) and the self-employed are enrolled in the public pension scheme. An average wage earner with a contribution record of 40 years is promised a pension benefit equivalent to some 70 percent of averaged lifetime earnings—one of the highest public pension replacement rates in the OECD.13 Civil servants are covered by separate pension plans.14 While the statutory retirement age in Luxembourg stands at 65, generous early retirement provisions are reflected in some of the lowest estimated transition ages to retirement in the OECD—58 years for men, and 55 years for women.15 Owing to the wide coverage of the public pension system, high replacement rates, and low effective retirement ages, total public pension spending (including civil service pensions) amounted to about 12 percent of GDP in 1999, one of the highest public pension spending ratios in the OECD.

24. The revenue side of Luxembourg’s first-pillar system relies on pension contributions (for private sector workers) and direct budget financing (for civil servants). The contribution rate for the private sector schemes—shared equally among employers, employees, and the state budget—has been stable at 24 percent of gross earnings since the mid-1980s. An earnings floor equivalent to the minimum wage is exempted from contributions, while the earnings ceiling for contributions is equivalent to five times the minimum wage. Luxembourg’s rapid employment growth since the early 1980s not only allowed to keep the level of the contribution rate constant, but it also underpinned the accumulation of a sizable pension reserve (some 21 percent of GDP at end-1999).16

25. The second or occupational pension pillar of Luxembourg’s pension system comprises voluntary supplementary company pension plans. Data on the significance of voluntary pension plans are scarce, but these plans are likely to have remained small in view of the dominant first-pillar scheme. A new law on supplementary pension schemes adopted in 1999 provides a flexible legal framework for the establishment of voluntary supplementary pension plans at the company level. The third pension pillar, voluntary individual income provisions for retirement, is likely to play a substantial role, particularly for persons with higher incomes.

C. A Framework

26. The main parameters affecting the financial condition of a stylized PAYG system can be brought out by considering the system’s basic budget constraint:

(1)N(t)[α(t)W(t)]= M(t)[β(t)W(t)],

where N(t) denotes the number of contributors to the system in period t, α(t) is the effective pension contribution rate, W(t) denotes average gross earnings subject to contributions, M(t) is the number of pensioners, and β(t) denotes the effective pension replacement rate. Equation (1) says that revenue and expenditure of a PAYG system have to match in each period.17 Accordingly, the equilibrium contribution rate is defined as:

(2)α(t)=β(t)[M(t)/N(t)],

where [M(t)/N(t)] is the pension system’s dependency ratio, and its inverse [N(t)/M(t)] is the pension system’s support ratio.

27. It is useful to extend the basic budget constraint (2) to allow for the analytical distinction between the two policy challenges discussed at the beginning of this chapter—population aging and variability in the public pension system’s contribution base due to regional shocks.18 Assume that the number of contributors amounts to a proportion γ(t) of the number of resident persons of working age N*. This proportion γ(t)—termed in the following the contributor coverage ratio—will depend on labor force participation, unemployment, and the number of cross-border workers. In large integrated economic areas with broad-based PAYG systems, the contributor coverage ratio will typically be close to the economy’s resident employment ratio (resident employment as a percent of resident working age population) and, apart from cyclical variations, it can be assumed to evolve rather smoothly. In Luxembourg’s particular case, the contributor coverage ratio exceeds the resident employment ratio by a large margin—about one-third of employment is accounted for by cross-border workers—and can also undergo potentially large variations in response to regional shocks to labor demand.

28. Similarly, assume that the number of pensioners corresponds to a proportion δ(t) of the number of resident elderly persons M*. In the case of Luxembourg, and again reflecting the possibility of large-scale inflows of cross-border workers, the proportion δ(t)—termed in the following the pensioner coverage ratio—can differ markedly from the proportion of resident pensioners as a percent of resident elderly persons.

29. Under these assumptions, the effective equilibrium contribution rate of a PAYG system can be re-written as:

(3)α(t)=β(t)[δ(t)/γ(t)][M*(t)/N*(t)],

where [M*(t)/N*(t)] is the system’s elderly dependency ratio for residents. Thus, equation (3) allows to separate analytically the issue of population aging, represented by the system’s elderly dependency ratio [M*(t)/N*(t)], from the issue of regional variability of employment flows—represented by the contributor and pensioner coverage ratios δ(t) and γ(t).

30. Figure II-1 illustrates the evolution of the public pension system’s dependency ratio and the underlying movements of the elderly dependency ratio and the two coverage ratios during 1980-99.

Figure II-1.
Figure II-1.

Luxembourg: Public Pension System Dependency Ratio, 1980-99

Citation: IMF Staff Country Reports 2000, 067; 10.5089/9781451824308.002.A002

Sources: United Nations, World Population Prospects (1998); Statec; and staff calculations.1/ Number of pensioners as a percent of the number of contributors.2/ Resident population aged 60 and over as a percent of the resident population aged 20-59.3/ Ratio of total number of contributors to resident population aged 20-59.4/ Ratio of total number of pensioners to resident population aged 60 and over.

D. Empirical Results

31. The model used for tracking a range of scenarios for the public pension finances over the period 2000-50 follows broadly the generic projection model structure outlined in Chand and Jaeger (1996, pp. 34-36). These exercises are illustrative and based on broad-brush assumptions that neglect many of the institutional details that characterize Luxembourg’s public pension system.19 The projection model’s key assumptions and underlying data sources include:

  • The demographic projection that underlies all stress testing exercises is based on an official population scenario for the period 1995-2050 (medium variant), which was developed by the Luxembourg Statistical Office (Statec). This demographic scenario assumes that persistent immigration inflows will mitigate the impact of population aging. As a consequence, the implied longer-term profile for Luxembourg’s elderly dependency ratio indicates less rapid population aging than in France and Germany, particularly toward the end of the projection horizon (Figure II-2).

  • Throughout the projection period 2000-50, labor force participation and unemployment rates are kept constant at their present levels.

  • Labor-augmenting technical progress is assumed to occur at a constant rate of 2 percent. Assuming balanced growth of the capital stock and effective labor input, real GDP grows at this constant rate plus the endogenously determined rate of growth of employment.

  • The projection model only simulates the evolution of the number of contributors and pensioners in the pension schemes for private sector workers (accounting for about 90 percent of all contributors). The share of elderly resident persons receiving a pension is kept constant at their present level. Cross-border workers are assumed to become eligible for a (reduced) pension after an average of 20 years of contributions. The number of civil service pensioners is assumed to grow by 1 percent a year.

Figure II-2.
Figure II-2.

Projections of Elderly Dependency Ratios in Luxembourg, France, and Germany, 2000-50 1/

(In percent)

Citation: IMF Staff Country Reports 2000, 067; 10.5089/9781451824308.002.A002

Sources: United Nations, World Population Prospects (1998); Statec; and staff calculations.1/ Resident population aged 60 and over as a percent of the resident population aged 20-59.

32. Three long-run scenarios for the period 2000-50 are used to study the range of possible pressures on Luxembourg’s public pension system:

  • The status quo scenario assumes that Luxembourg’s regional economic boom will persist in the longer run at the average growth rates observed over the last 20 years. Thus, real GDP growth is exogenously fixed at 5 percent a year during 2000-50, requiring employment (and therefore the number of contributors to the public pension system) to grow by 3 percent a year (Figure II-3). As the number of available resident workers is constrained by the demographic scenario and the assumptions of constant labor force and unemployment rates, this scenario implies massive inflows of cross-border workers—in this scenario, the ratio of cross-border to resident workers in 2050 would rise to about 6.

  • The average growth scenario assumes that the status quo persists only until 2005. In that year, the net inflow of cross-border workers is assumed to stop (with cross-border workers accounting for some 40 percent of total employment) and employment growth during the remainder of the projection period 2006-50 is driven by the demographic characteristics of the resident population and other model assumptions (Figure II-3).

  • Finally, the sharp economic reversal scenario also assumes that the status quo persists until 2005, but that the cross-border work force will afterwards linearly decline to zero during 2006-15, and remain at zero throughout the rest of the projection period (Figure II-3).

Figure II-3.
Figure II-3.

Luxembourg: Alternative Labor Market Scenarios, 2000-50

Citation: IMF Staff Country Reports 2000, 067; 10.5089/9781451824308.002.A002

Source: Staff calculations.1/ Constant annual real GDP growth of 5 percent during 2000-50.2/ Average annual real GDP growth of 2.2 percent during 2000-50.3/ Average annual real GDP growth of 1.2 percent during 2000-50.

33. Turning to the implications of these scenarios for the public pension system, in the status quo scenario the pension system dependency ratio remains roughly stable throughout the projection period (Figure II-4). Thus, the financing pressures on the system due to population aging are fully offset by the rapid rise in the contributor coverage ratio owing to massive inflows of cross-border workers. The implied pension cost as a percentage of GDP remains roughly constant over the projection horizon (Figure II-5).

Figure II-4.
Figure II-4.

Luxembourg: Stress Testing Scenarios for the Public Pension System, 2000-50

Citation: IMF Staff Country Reports 2000, 067; 10.5089/9781451824308.002.A002

Source: Staff estimates.1/ Number of pensioners as a percent of the number of contributors.2/ Ratio of total number of contributors to resident population aged 20-59.3/ Ratio of total number of pensioners to resident population aged 60 and over.
Figure II-5.
Figure II-5.

Luxembourg: Stress Testing Scenarios for Pension Expenditures, 2000-50

(In percent of GDP)

Citation: IMF Staff Country Reports 2000, 067; 10.5089/9781451824308.002.A002

Source: Staff estimates.

34. In the average growth scenario, the pension system’s dependency ratio and pension cost rise sharply as growth slows below the status quo path after 2005. With net inflows of cross-border workers assumed to come to an abrupt stop in 2005, the contributor coverage ratio is projected to stabilize over the remainder of the projection horizon. The projected runup in the dependency ratio and pension cost—relative to the favorable status quo scenario—reflects two factors. First, population aging in the form of the projected rise in the elderly dependency ratio of the resident population feeds fully through to the system dependency ratio and pension cost. And second, the pensioner coverage ratio continues to rise after 2005 as cross-border workers continue to reach the retirement age.

35. In the sharp economic reversal scenario, the pension system dependency ratio rises even more sharply than in the average growth scenario. This reflects mainly a sharply declining contributor coverage ratio as the stock of cross-border workers is assumed to decline to zero during 2006-15.

E. Policy Options

36. There are two main generic parametric PAYG policy levers to offset increases in the system dependency ratio—increases in the contribution rate or decreases in the pension replacement rate. However, both policy levers are likely to raise concerns if used to offset significant increases in the system dependency ratio. Sharp increases in contribution rates in response to an adverse economic shock or reversal could lead to a further deterioration in what would already be a difficult labor market situation. And the option of deep cuts in pension replacement rates would conflict with the objective of stable retirement income provision.

37. A second generic set of policy options would comprise raising the contributor coverage ratio or decreasing the pensioner coverage ratio, e.g., through increases in effective retirement ages or by providing new incentives to increase labor force participation rates. However, the effectiveness of these particular options in response to large adverse shocks or economic reversals may also be limited: (i) the measures may need to be phased in gradually, in particular in the case of increases of retirement ages; (ii) the economic environment may not be favorable for the required increases in labor force participation rates; and (iii) higher labor force participation rates, while helpful in providing additional resources in the short run, would further increase unfunded pension liabilities of the PAYG system.

38. More diversification of the financing of overall pension provision by moving to a diversified multipillar structure, perhaps benchmarked on Switzerland’s long-standing multipillar approach, would allow Luxembourg to make a significant portion of pension provision independent of local economic booms and busts. While the transition to a multipillar structure would likely involve significant transition costs, an exceptionally favorable fiscal outlook for the next few years should provide ample budgetary scope for far-reaching reforms.

39. Apart from spreading financing risks, moving to a more balanced multipillar structure would likely have further advantages: (i) the impact of shocks on public finances would be mitigated; (ii) future required social contribution rates could be lower than otherwise; (iii) the higher rates of return of funded schemes (relative to rates of return in PAYG schemes) would add to Luxembourg’s locational advantages in terms of labor mobility; and (iv) Luxembourg could draw on extensive local financial expertise to develop funded pension schemes. In this context, the recently adopted laws on international pension funds and supplementary pension schemes appear to have prepared the legal ground for a proactive pension policy.

References

  • Blöndal, Sveinbjörn, and Stefano Scarpetta, 1998, “The Retirement Decision in OECD Countries,” OECD Working Paper AWP 1.4.

  • Chand, Sheetal K., and Albert Jaeger, 1996, Aging Populations and Public Pension Schemes, IMF Occasional Paper No. 147.

  • Chand, Sheetal K., and Albert Jaeger, 2000, “Reforming the PAYG System,” in: World Bank Primer on Pension Reform, (at: http://www.worldbank.org/pensions).

    • Search Google Scholar
    • Export Citation
  • Disney, Richard, 1999, “OECD Public Pension Programs in Crisis: An Evaluation of the Reform Options,” in: World Bank Primer on Pension Reform (at: http://www.worldbank.org/pensions).

    • Search Google Scholar
    • Export Citation
  • Economist Club Luxembourg, 2000, “L’avenir de nos pensions,” (manuscript).

  • IMF Country Report No. 98/67, 1996, Luxembourg—Selected Issues.

  • IMF Country Report No. 96/47, 1998, Luxembourg—Selected Issues.

  • Ministère de la Sécurité Sociale, 1995, Etude actuarielle des regimes de pension.

  • United Nations, 1999, World Population Prospects. The 1998 Revision. Volume II: The Sex and Age Distribution of the World Population.

  • U.S. Social Security Administration, 1999, Social Security Programs throughout the World-1999.

9

Prepared by Albert Jaeger and Caroline Kollau.

10

See Disney (1999) for a review of reform options focusing on OECD countries.

11

See IMF Country Reports No. 96/47 and 98/67.

12

U.S. Social Security Administration (1999, pp. 225-27) provides a concise overview of Luxembourg’s social security system. A study published by the Economist Club Luxembourg (2000) summarizes the history of Luxembourg’s public pension system in an annex.

13

An OECD cross-country study of standardized gross pension replacement rates estimated a replacement rate for Luxembourg of 93.2 percent compared to an OECD average of 59.3 percent (1995 data). See Blöndal and Scarpetta (1998, Table II-3).

14

A civil service pension reform adopted in 1998 will-subject to a long transition period-reduce the more generous civil service pension benefits to the present levels of the private sector schemes.

16

The pension schemes for private sector workers are required to hold a contingency reserve equivalent to at least 1.5 to 2.5 years of benefits.

17

For simplicity, the contingency reserve is not incorporated in the budget constraint.

19

For an actuarial study covering the period 1995-2015 that includes a wealth of institutional detail, see Ministere de la Securite Sociale (1995).

Luxembourg: Selected Issues and Statistical Appendix
Author: International Monetary Fund
  • View in gallery

    Luxembourg: Public Pension System Dependency Ratio, 1980-99

  • View in gallery

    Projections of Elderly Dependency Ratios in Luxembourg, France, and Germany, 2000-50 1/

    (In percent)

  • View in gallery

    Luxembourg: Alternative Labor Market Scenarios, 2000-50

  • View in gallery

    Luxembourg: Stress Testing Scenarios for the Public Pension System, 2000-50

  • View in gallery

    Luxembourg: Stress Testing Scenarios for Pension Expenditures, 2000-50

    (In percent of GDP)