This Selected Issues Paper on Belgium provides an overview of the extent of trade and financial openness of Belgium and the links to particular countries. With an export-to-GDP ratio of 79 percent, Belgium belongs to the most open economies in Europe and also globally. Its exports are highly concentrated with a share of three-fourths of total merchandise exports accounted for by the European Union, of which close to two-thirds go to Germany, France, and the Netherlands.

Abstract

This Selected Issues Paper on Belgium provides an overview of the extent of trade and financial openness of Belgium and the links to particular countries. With an export-to-GDP ratio of 79 percent, Belgium belongs to the most open economies in Europe and also globally. Its exports are highly concentrated with a share of three-fourths of total merchandise exports accounted for by the European Union, of which close to two-thirds go to Germany, France, and the Netherlands.

III. TOWARDS JOB-CREATING LABOR MARKET REFORM1

1. The new government program aims to raise the share of employed persons in the working age population by 5 percentage points from 67.6 percent in 2010 to 73.2 percent by 2020. The share of the employed in the working age population is below that in most EU peers. Belgium’s low employment rate is a constraint to growth and weakens public finances. It is a result of both relatively high unemployment among the labor force and a relatively low labor market participation rate. The unemployment benefit system grants wide access to modest benefits while the pension system allows early exit from the labor market. The reforms agreed in December 2011 are a first step in reforming the benefit system that can be built upon to bring it more in line with EU standards, and provide incentives to boost activity.

2. This note proceeds in the following manner. Section A illustrates the key characteristics of the Belgian labor market, especially highlighting the dispersion in regional unemployment rates and low participation rates. Sections B and C describe the key elements of the unemployment and pension benefit systems, focusing on some of the labor market incentives inherent in the system. In December 2011, the government announced a series of labor market reform initiatives, which will also be sketched out in Sections B and C. Section D includes a brief overview of the active labor market policies pursued by the various levels of government while Section E illustrates some impediments to labor mobility. Section F takes a first cut at estimating the impact of the unemployment benefit system on regional employment in a regional panel data set. Section G concludes and makes some policy recommendations.

A. Characteristics of the Labor Market

3. Owing to weak labor market participation, employment accounts for a relatively small share of the working age population by eurozone standards. For example, the unemployment rate in 2010 was somewhat below the euro area average. This, however, did not translate into a higher employment rate than the euro area average because labor market participation was weak: At 67¾ percent, the overall activity rate in 2010 was well below the eurozone average (71½ percent). Despite an increase over the past two decades, low labor force participation by EU standards has been a persistent feature of the Belgian labor market. Throughout the decade from 2001-10, Belgium’s activity rate has been 3¼-4½ percentage points point below the euro area average. Within the eurozone, only Italy, Malta, Greece, Portugal, and Luxembourg have at times shown lower activity rates.

uA03fig01

Labor Market Activity, 2010

(Percent of working age population)

Citation: IMF Staff Country Reports 2012, 056; 10.5089/9781475502398.002.A003

Source: Eurostat.

4. During the most recent downturn, employment growth has held up well, reflecting government support. Belgium has a large public sector and makes extensive use of employment subsidies in various forms. As a result, public employment (in public administration, education, or the government-supported health sector) and subsidized or supported employment (including workers on subsidized part-time employment and service voucher workers) have accounted for more than ninety percent of all employment growth from 2000 to 2010.2 During the most recent downturn of 2007-10, public and publicly supported employment growth has partially compensated for employment losses in the private sector.

uA03fig02

Employment Growth, 2007–2010

(Percent)

Citation: IMF Staff Country Reports 2012, 056; 10.5089/9781475502398.002.A003

Source: Eurostat.
uA03fig03

Contribution to Employment Growth

(Percent)

Citation: IMF Staff Country Reports 2012, 056; 10.5089/9781475502398.002.A003

Sources: ONEm; NBB; and Eurostat.1/ Includes workers supported by the government, service voucher workers, and workers reducing their worktime on government support.

5. The dispersion of unemployment rates is, by EU standards, exceptionally large in Belgium. In 2010, unemployment rates ranged from 3.8 percent in West-Flanders to 17.3 percent in Brussels. Only Spain and France, both about twenty times Belgium’s size by area, have a greater regional dispersion in unemployment rates.3 Within regions, the dispersion of unemployment rates in Wallonia is as high as that of Portugal and that in Flanders as low as that in Sweden. The dispersion in regional unemployment rates is most pronounced among young workers below the age of 30. The highest youth unemployment rates are in communities in Brussels and Wallonia. In contrast to other EU member countries, including the neighboring countries, there has been no significant convergence in unemployment rates over the past decade.

uA03fig04

Regional Unemployment Rates, 2010

(NUTS 2 level, percent)

Citation: IMF Staff Country Reports 2012, 056; 10.5089/9781475502398.002.A003

Source: Eurostat.
uA03fig05

Standard Deviation in Regional Unemployment Rates, 2010

(NUTS2, percent)

Citation: IMF Staff Country Reports 2012, 056; 10.5089/9781475502398.002.A003

Source: Eurostat.

Convergence in unemployment rates 1/ 2000-2010

Coefficient estimates of the change in unemployment rates on lagged unemployment rates. A negative coefficient indicates convergence in unemployment rates. A zero coefficient indicates an estimate insignificantly different from zero at the 5 percent significance level.

6. In contrast, the dispersion of labor force participation rates is broadly in line with that in EU peers. Low activity rates are a country-wide phenomenon. The dispersion of labor force participation rates across provinces is relatively modest and in line with other countries of similar population size or Germany.

uA03fig06

Regional Labor Force, 2010

(NUTS2, percent of working age population)

Citation: IMF Staff Country Reports 2012, 056; 10.5089/9781475502398.002.A003

Source: Eurostat.
uA03fig06a

Dispersion in Regional Unemployment Rates by Age Group, 2010

(Percent of labor force)

Citation: IMF Staff Country Reports 2012, 056; 10.5089/9781475502398.002.A003

Sources: Eurostat; and ONEm.

7. Disparities in regional labor market outcomes may reflect structural differences that are amplified by labor market policies. Unemployment benefits, which are moderate by EU standards, are determined at the national level, but activation policies—and in this context, some of the monitoring efforts—are implemented at the regional level. Regional agencies refer job seekers who violate job search requirements to the national agency that imposes sanctions. For older workers, the level of pension benefits is similarly moderate by EU standards, but various early-retirement schemes encourage labor market dropout well before the official retirement age of 65 years. These are not necessarily reflected in unemployment rates but reduce the share of employed in the working age population. Active labor market policies, including job search assistance, training, direct job creation, and employment support, are widely used and more costly than in other countries with similar unemployment rates.

uA03fig07

Expenditures on Active Labor Market Measures, 2009

(Percent of GDP)

Citation: IMF Staff Country Reports 2012, 056; 10.5089/9781475502398.002.A003

Sources: Eurostat; and OECD.
uA03fig08

Net Replacement Rate of Unemployment Benefits, at 67% of Average Wage, 2009

(Percent)

Citation: IMF Staff Country Reports 2012, 056; 10.5089/9781475502398.002.A003

Source: OECD.Note: For single-earner family with two-children at 67 percent of average wage.
uA03fig09

Net Replacement Rate of Pension Benefits, at 50% of Average Wage

(Percent)

Citation: IMF Staff Country Reports 2012, 056; 10.5089/9781475502398.002.A003

Sources: OECD Pensions at a Glance (2011).Note: Most recent data available, 2008 for Belgium.

B. Unemployment Benefits

8. Unemployment benefits are overall moderate by euro area standards but openended. With a net replacement rate of 60 percent for the average worker, they are initially lower than in other OECD countries, but in contrast to other OECD economies decline only marginally over time. In principle, unemployment benefits are 60 percent of the last salary up to a cap for one year. After the first year, the benefit drops to 55 percent of the last salary thereafter for single workers. For workers with a family, the benefit level is maintained even after the first year in the absence of other family income. If additional family income is available, unemployment benefits are reduced to 40 percent of the last salary but no lower than a lump sum of Ĩ474.50 per month. Workers above the age of 50 years have been eligible for higher benefits and benefits for specific family situations can be significantly higher.4 Unemployment benefits were paid indefinitely and not means-tested.

9. An important element of the reform agenda agreed in December 2011 was to make unemployment benefits more degressive over the duration of unemployment. The net replacement rate of benefits for the initial three months of unemployment is raised to 65 percent of the last wage, but will then drop back to the current level of 60 percent for the following nine months and is further reduced for all household types in several steps to a lump sum which is only little above the social assistance level. This lump sum is reached at the latest after four years of unemployment but much earlier for those unemployed with short working careers. Hence, the 5 percent increase in the net replacement rate of unemployment benefits during the first three months of employment will be matched by a reduction in benefits by the time the unemployment spells extends into 1¼-2 years (depending on working career length) and further thereafter. Unemployed above 55 years (previously 50 years) will remain eligible for higher unemployment benefits.

10. Eligibility requirements for unemployment benefits have been lenient for some groups. In principle, eligibility for unemployment benefits is conditional on active search efforts and availability for a suitable job. A suitable job is defined—during the first six months of job search—as a similar job to the pre-unemployment one within a radius of 25 km or a total travel time not exceeding four hours a day. Compliance with these eligibility requirements is verified by the regional public employment services from the start of the unemployment spell and by the benefit-providing federal authorities only after 18 months of unemployment and then at 12 months intervals (although regions monitor compliance more frequently). The reform agenda of December 2011 tightens eligibility criteria: A suitable job is redefined to include a perimeter of 60 kilometers and the frequency of federal monitoring of compliance with eligibility criteria will be doubled.

11. The strength of regional monitoring depends on resources allocated towards the task. For example, in Brussels one coach is responsible for 170 unemployed, in Wallonia for 140 unemployed, and in Flanders for 70 unemployed. For workers over the age of 50, eligibility requirements for unemployment benefits are not enforced by the federal authorities. This likely contributes to high unemployment among older workers. For Australia (Borland and Tseng, 2007) and the U.S. (Klepinger, Johnson, and Joesch, 2002), stricter job search requirements have been found to reduce the duration of unemployment spells.

12. Unemployment benefits are also extended under several special schemes:

  • Various career-break benefits exist. Those benefits are below unemployment benefits and are available for workers who take full-time career breaks for five years and/or reduce their working time for five years. No time limitation exists for workers above 50. Career breaks count fully towards pension entitlement. In the third quarter of 2011, about 270,000 workers received career-break benefits. The reforms agreed in December 2011 raise the minimum qualifying age for unlimited career breaks and part-time employment with benefits to 55 years, with some exceptions (e.g. long careers in physically demanding jobs) where the minimum age remains 50 years. In total, career breaks before the age of 55 are limited in the private sector to a maximum of four years. In addition, career breaks will only partially count towards pension benefits.

  • Under the temporary unemployment program, the employer reduces temporarily working hours to accommodate reduced operations. Workers are compensated with partial unemployment benefits. These arrangements typically involve employer top-up payments, so that the net income loss of workers is limited. The program was originally only available to blue-collar worker but was extended as one of the 2009 crisis measures to white-collar workers. The scheme has been widely used. In the first quarter of 2010, more than 256,000 workers received temporary unemployment benefits. This number had fallen below 116,000 workers by the third quarter of 2011.

  • Following a waiting period of nine months, young school leavers are entitled to unemployment benefits (called “waiting allowances”), although at less than full levels, without ever having contributed to the system. In the third quarter of 2011, about 110,000 unemployed received such waiting allowances. As part of the reform agreed in December 2011, the waiting period is extended to 12 months, waiting allowances are limited to 36 months, and young unemployed will be subject to more stringent monitoring of job search efforts.

  • Involuntary part-time workers registered as job seekers for a full-time job can register for unemployment benefits. The total income from benefits and part-time employment can reach the income from a full-time job. Nearly 50,000 part-time workers were registered under the program in the third quarter of 2011.

The Administration of Unemployment Benefits

Revenue collection and budget management responsibilities are centralized in the National Social Security Office (RSZ-ONSS). The responsibility for unemployment benefit delivery lies with the national unemployment fund (RVA-ONEM). For most people, local trade union offices act as paying agents for RVA-ONEM. RVA-ONEM assesses eligibility for benefits and imposition of sanctions, also upon referral from regional agencies. Participation in activation and job placement programs, is determined by regional public employment services. There are four regional employment services: Actiris in Brussels, the Flemish Employment and Vocational Training Agency in Flanders (VDAB), Forem in Wallonia, and Arbeitsamt for the German-speaking Community. Private employment agencies are also active, especially in Flanders. Actiris and the VDAB have a cooperation agreement to place unemployed from Brussels in Flanders, if they cannot find a job in Brussels. Although the national rules are uniform, implementation has differed in practice. For example, in 2007, Wallonia and Brussels imposed half as many sanctions (2.51 and 2.02 sanctions per hundred unoccupied job seekers, respectively) as Flanders (6.92 sanctions per hundred unoccupied job seekers). However, this has gradually changed and the relative ranking was reversed in 2010.

C. Pension System

13. Aging costs are expected to rise by 5½ percent of GDP over the course of the next five decades to about 31 percent of GDP. The High Council of Finance (2011) estimates that three quarters of this increase will be due to rising pension cost. The cost in percent of GDP of the public pension scheme for private and public sector employees will increase by almost half each. IMF (2011) suggests that this increase is about 1¼ percentage point of GDP above the eurozone average.

uA03fig10a

Aging Cost

(Percent of GDP)

Citation: IMF Staff Country Reports 2012, 056; 10.5089/9781475502398.002.A003

Source: High Council of Finance (2011).

14. Pillar I pensions are the predominant part of the Belgian pension system. The public pension system for private employees differs from that for self-employed and public workers (that for public employees typically being more generous than that for private employees).5 In 2003, the Vandenbroucke Law promoted sectoral pension plans determined by collective labor agreements, which differ between economic sectors and are administered by the social partners. In 2006, about 51 percent of the active population was covered by such sectoral agreements. Agreements typically are defined contributions (Trampusch at al., 2010). Finally, voluntary private pension insurance is available. For an average wage earner, such voluntary pension may add net replacement rates of 20 percent (OECD 2011).

15. Five types of benefits are available to private sector employees above the age of 50 (55 from 2012): unemployment benefits at lenient conditions, the pre-pension system (unemployment benefits with an employer top-up), career-break benefits, the early retirement system, and the standard pension system from the age of 65. For workers above the age of 55, expenditures on unemployment, pre-pension, and career-break benefits accounted for 1¼ percent of GDP while pension payments amounted to 5½ percent of GDP in 2010.6

  • From the age of 50, workers on standard unemployment benefits do not have to search actively for work. The reform agenda agreed in December 2011 stipulates that this threshold age be raised to 55 years from 2013 and 58 years from 2016. Pension entitlements are accumulated at the last salary except when unemployment benefits are at the minimum lump sum. Workers often receive employer top-up payments and after one year of unemployment a supplement to their normal unemployment benefits.

  • From the age of 50, workers who lose their job due to collective dismissals, restructuring or company losses, are entitled to “pre-pensions”. From the age of 58 in most industries, “pre-pensions” are open to workers dismissed for any reason. Pre-pensions are unemployment benefits topped up by employers and with no obligation for job search. Almost one quarter of workers on unemployment benefits above the age of 50 are in the pre-pension scheme. Pension rights are accumulated throughout the period of pre-pension at the last full salary. The reforms agreed in December 2011 raise the minimum age for pre-pensions due to collective dismissal and restructuring to 55 years from 2013. In other cases, the current minimum age of 58 years is raised to 60 (from 2012 for new collective agreements, from 2015 for existing collective agreements). From 2013, recipients of pre-pension benefits below the age of 60 years would be required to search new jobs to qualify. The years below the age of 60 on pre-pensions will no longer count fully towards the career length applied toward pension benefits. To avoid companies selectively dismissing older workers, collective dismissals must reflect the age structure of the firm to qualify for pre-pension benefits.7

  • Career breaks benefits have been available for workers above the age of 50 for an unlimited period, while they were limited to five year for younger workers. The reforms agreed in December 2011 increased the minimum age for an unlimited duration of career break benefits to 55 years. In some cases, the minimum age remains 50 years (e.g. long careers in physically demanding jobs).

  • Early retirement pension benefits make the standard pension scheme available to workers aged 60 or above with a full career of 35 years. Benefit levels are adjusted for shorter career lengths below 45 years by 2¼ percent per annum—well below the actuarially neutral rate of 5-6 percent per annum. A small pension supplement of Ĩ2 per day worked is provided for those working above the age of 61. No supplementary pension rights are accumulated during the period of early retirement. Under the reforms agreed in December 2011, the minimum qualifying age for early retirement would be raised in six-month steps to 62 years and the minimum career would be extended in annual steps to 40 years by 2016.

  • From the age of 65, workers are eligible for pension benefits. Full benefits accrue for those above 65 years with a full career of 45 years. Benefits are reduced by 2¼ percent per year for shorter careers—again, well below the actuarially neutral rate of 5-6 percent. For the private sector, pension benefits are 60 percent of the average wage over the whole career. In the public sector, pension benefits are based on the average salary of the last five years. Under the reforms agreed in December 2011, pensions for civil servants under the age of 50 will be based on the average wage of the last ten years from 2012. Pension (like unemployment benefits) are automatically adjusted by a subindex of the consumer price index (the “health index”). This contrasts with Germany (where inflation indexation is automatically corrected for the dependency ratio and growth in contributions) and France (where the indexation is regularly renegotiated).

16. The standard pension benefit level is broadly in line with EU standards, but the various early retirement and pre-pension schemes have higher replacement rates. The net replacement rate of the standard pension scheme is about 82 percent for low-wage workers (compared with an EU27 average of 82 percent) and 52 percent for high-wage workers. Since pre-pension schemes include top-up payments by employers, the net replacement rate of labor market exits for older workers before the official retirement age can be significantly higher and creates increased pension rights under the standard pension scheme.

17. As a result of incentives to retire early, the effective retirement age was 59 years in 2009. With Austria, Belgium has the widest gap between the official and the effective retirement age among OECD countries. Common use of early-retirement and pre-pension schemes partly reflects relatively strict employment protection. This encourages employers to substitute outright worker dismissal with placement into benefit programs such as pre-pensions, career breaks and part-time work, or early retirement. In 2008, according to the OECD’s employment protection index, Belgium’s employment protection in cases of mass redundancies was the second most rigid after Italy’s in the whole sample of OECD countries included in the ranking. For normal redundancies, Belgium ranked third among its Western European peers, after France, Greece, and Luxembourg.

uA03fig11

Retirement Age

(Years)

Citation: IMF Staff Country Reports 2012, 056; 10.5089/9781475502398.002.A003

Source: OECD Pensions at a Glance, 2011.Note: Latest available data. Does not include announced pension reforms.

18. Access to early pension and quasi-pension benefits is fairly uniform across the country. Compared with unemployment benefits, there is little dispersion in the share of the over-55 year-olds who receive some form of non- pension old-age support. While regions differ in the particular instrument of granting quasi-pensions, in aggregate they provide them to similar shares of the older working age population.

uA03fig12

Pension and Quasi-Pension Benefit Recipients Between 50-64 years, 2010

(Percent of population aged 50-64 years)

Citation: IMF Staff Country Reports 2012, 056; 10.5089/9781475502398.002.A003

Sources: ONEm; Eurostat; and IMF staff estimates.

D. Active Labor Market Policies

19. Belgium’s expenditures on active labor market policies are the second highest in the OECD (1.4 percent of GDP in 2009). Active labor market measures include job search assistance but also training, direct job creation, fiscal incentives for recruitment and employment maintenance, and fiscally supported employment schemes. Startup incentives for business creation are less than in other OECD countries.

uA03fig13

Expenditures on Active Labor Market Policies, 2009

(Share of total)

Citation: IMF Staff Country Reports 2012, 056; 10.5089/9781475502398.002.A003

Source: OECD.
uA03fig13a

Expenditures on Active Labor Market Policies, 2009

(Percent of GDP)

Citation: IMF Staff Country Reports 2012, 056; 10.5089/9781475502398.002.A003

Source: OECD.

20. In Belgium, subsidized employment and employment incentives account for the bulk of expenditures on active labor market measures. More than eighty percent of expenditures on active labor market measures are due to employment subsidies in various forms.8 Especially Brussels employs these programs. Job search assistance accounted for only 6 percent of total expenditures on active labor market measures. This is less than half of the OECD average (14 percent). OECD (2007) describes how other countries have increasingly focused their expenditures on job search assistance, with participation in subsidized employment programs only after job search fails for an extended period. This is in line with a review of evaluation studies that finds particularly strong effects from job assistance and sanctions (Kluve, 2006).

E. Impediments and Incentives for Mobility

21. Labor market mobility in general is limited in Belgium.

  • At a regional level, limited labor market mobility is often attributed to insufficient language skills.

  • At the sectoral level, Vandenbrande (2000) documents the low mobility rates in and out of most sectors in Belgium compared with Sweden, Finland, and Norway. It is possible that the predominance of industry-financed training (rather than public vocational training) discourages worker training that is portable across industries.

22. The centralized wage setting mechanism compresses wage differentials and thus dulls incentives for labor market mobility. Biannual wage negotiations for the private sector start with a national agreement that sets an upper limit to wage increases composed of two elements, expected inflation for the next two years and a supplementary margin. This margin serves as a reference point for all sectors on top of the automatic price indexation.9 The result is a relatively homogeneous wage structure across sectors and regions which does not adjust flexibly to local and sectoral labor market conditions. Davies and Hallet (2001) highlight the lack of regional wage differentials as one of the key reasons for persistent regional unemployment differentials in EU countries. In addition, the seniority premium in Belgium is among the highest in Western European countries, after Luxembourg, Italy, and Greece (Mastribuoni and Taddei, 2011), thus making it difficult for unemployed older workers to find new employment.

uA03fig14

Hourly Wages, Regional Averages, 2004

(Euros)

Citation: IMF Staff Country Reports 2012, 056; 10.5089/9781475502398.002.A003

Source: Eurostat.

F. The Importance of Labor Market Policies

23. A key question in any analysis of social benefits is to what extent the business cycle causes unemployment and low labor market participation and to what extent social benefits do. The business cycle is a national or EU-wide phenomenon. In a perfectly flexible labor market, workers from regions and industries in a slump would relocate to rising regions and industries. Social benefits and regulatory rigidities may hinder this relocation.

24. In this section, the effect of the interaction of social benefits with regional population structures is analyzed econometrically using a dataset of subnational labor markets outcomes from Eurostat’s regional database. By focusing on subnational labor market outcomes, the interaction of social benefits with regional population characteristics, notably its age structure, can be distilled. The regressions are used as a tool to identify statistical patterns rather than a comprehensive analysis of the determinants of regional unemployment rates. In this spirit, they control for an EU-wide business cycle by including time dummies and for country characteristics beyond those of immediate interest here or those unavailable for data reasons by including country dummies.10 (While this precludes the separate inclusion of country-level variables, it is a comprehensive control for country-level omitted variables.) Regional labor market data at the NUTS3 level—the municipal level, in the case of Belgium—to 2009 is used. Labor market outcomes such as unemployment or employment are stocks and, hence, highly persistent. The autocorrelation caused by this persistence would bias the estimates. To mitigate the persistence somewhat, annual averages are created over the periods 2000-2003, 2004-2007, and 2008-2009.11

25. The variable of interest here is the municipal employment rate—the share of employed in the working age population. This variable will inter alia be driven by disincentives inherent in the unemployment system to both drop out of the labor market entirely and to remain unemployed. In the literature on unemployment differentials in the U.S. and Canada, the focus is on regional differences in social security systems (for an overview see Elhorst, 2000). In European countries, in contrast, social security benefits are typically set and applied nation-wide. Of interest here is therefore the interaction between the population structure in a region and the generosity of the country-wide unemployment system—both for young and for old workers. Young workers tend to have below-average wages. Hence, for young workers, the focus here is on the interaction between the net replacement rate at wages one-third below the average wage and the share of the provincial population below the age of 25. Older workers tend to have average or above-average wages. Hence, for workers from the age of 55 onwards (i.e. those still well below the official retirement age), the regressions focus on the interaction of their share in the provincial population with the net replacement rate of unemployment benefits for workers at average wages.12 In sum, the following regression is run:

erjit=sharejityoung+sharejityoungXbenefititlowwage+sharejitold+sharejitoldXbenefititaveragewage+countrydummyi+timedummyt+ϵjit

26. The results suggest that, in this sample, more generous unemployment benefit systems were associated with a stronger correlation between the population age structure and the employment rate. Column I in Table III.1 shows the correlation between the age structure and municipal employment rates: a greater share of people below 25 years and above 55 years was associated with a lower employment rate. Column II introduces the two interaction terms: the share of below-25 year olds interacted with the net replacement rate for worker with below-average incomes and the share of above-55 year olds interacted with the net replacement rate for workers with average wages. The results in Column II suggest that the negative correlations shown in Column I were driven by the benefit system. By itself, a greater share of below-25 year olds was now associated with greater employment rates while that of workers above 55 years no longer had any impact. This is in line with earlier evidence (for a summary, see Box 3 of IMF, 2011). The interaction terms with benefit levels are significantly negative: the higher net replacement rates of unemployment benefits, the greater the drag of a large share of workers at either end of the working age population on employment rates. Column III raises the bar higher: it also includes country-time dummies to control for country-specific business cycles. Again, controlling for the interaction terms, a greater share of below-25 year olds was associated with a greater employment rate and the share of above-55 year olds had no significant impact. Where the unemployment benefit system was generous, however, it dampened the positive link between employment and the share of people below 25 years (albeit with marginal significance) while it turned a larger population of above-55 year olds into a drag on employment rates. Finally, the regressions in Columns I-III fall short from other studies in the literature that control for other provincial- level characteristics, such as industry structure and unit labor cost (Lopez-Bazo, Barrio and Artis, 2002). Comparable municipal-level data are not available for the whole period in this sample. However, they can be controlled for by using a provincial fixed-effects regression. All provincial characteristics, including the age structure, would be captured in the provincial fixed effects. Results are shown in Column IV of Table III.1. Again and as expected, the interaction between unemployment benefits and the share of the youngest or oldest members of the working age population was significantly negative.

Table III.1.

Panel OLS Regression. Dependent Variable: Employment Rate

(percent)

Note: Robust pvalues in brackets.*** p<0.01, ** p<0.05, * p<0.1, § p<0.15.Country and time dummies are not reported here.

27. These statistical correlations suggest that, until the recent reforms, the Belgian unemployment benefit system may have been sufficiently generous to turn a high share of below-25 year olds and above-55 year olds into a drag on employment rates. Consider, for example, the overall effect of a 1 percentage point increase in the share of below-25 year olds. The overall effect is composed of both the level effect (estimated by the coefficient 0.60 in Column II of Table III.1) and the interaction with the net replacement rate of unemployment benefits for below-average wage earners that prevailed in Belgium on average during 2008-10 (estimated with a coefficient of -0.564 in Column II of Table III.1). Hence, the overall effect of a 1 percentage point increase in the share of below-25 year olds—shown in the coefficient line at its intersection with the vertical line in the left figure below—was to reduce the employment rate by ½ percentage point. Similarly, given the Belgian unemployment benefit system, the overall effect of a 1 percentage point increase in the share of above-55 year olds was to reduce the employment rate by ¼ percentage point.

uA03fig15

Overall Effect on the Employment Rate of the Share of Below 25-Year Olds Depending on the Generosity of Unemployment Benefits

(95 percent confidence interval)

Citation: IMF Staff Country Reports 2012, 056; 10.5089/9781475502398.002.A003

Source: IMF staff estimates.Note: Defined as sum of regression coefficients in Column II on share of below-25 year olds and interaction term with net replacement rate .
uA03fig16

Overall Effect on the Employment Rate of the Share of Above 55-Year Olds Depending on the Generosity of Unemployment Benefits

(95 percent confidence interval)

Citation: IMF Staff Country Reports 2012, 056; 10.5089/9781475502398.002.A003

Source: IMF staff estimates.Note: Defined as sum of regression coefficients in Column II on share of above-55 year olds and interaction term with net replacement rate.

28. What might be expected if the net replacement rates of Belgian unemployment benefits were reduced by, say, 5 percentage points? Compared to the pre-2012 system, a 5 percentage point cut is approximately what an unemployed can expect after 1½-2 years of unemployment. Based on the coefficient estimates from Column II of Table III.1 and the age structure of Belgium’s regions, the potential impact on employment rates can be derived. A 5 percentage point cut in the net replacement rate for all unemployment benefits might raise the employment rate by 2-3 percentage point. This is not insignificant in view of the government’s target to raise the employment rate by 5 percentage points.

G. Further Reform Needs

29. Raising employment is key to strengthening potential growth and ensuring fiscal sustainability. In particular, the combined unemployment and pension benefit systems provide incentives for labor market exit well before the official retirement age of 65 years. Pre-pension and early retirement benefits and lenient provisions for benefits during career break and unemployment spells encourage workers above 50 years to reduce their labor market participation. An illustrative regression suggests that the current unemployment benefits distort the impact of the population age structure and reduces regional employment rates in Belgium.

30. The labor market and pension reforms agreed in December 2011 are steps towards raising employment (Appendix Table A.1). They span the unemployment benefit system, including various paths to early labor market exit, as well as pension reform. In the unemployment benefit system, search requirements will be enforced more strictly, the benefit duration will be reduced for the young, benefit levels will be made more degressive over the duration of unemployment. In the unemployment benefit system for older workers, higher benefits will only begin at the age of 55 years, the pre-pension system will become less accessible because of tightened age and career conditions and will become less attractive for employers with higher taxes on employer top-up payments. Pre-pension benefits will become altogether inaccessible before the age of 55 years in cases of dismissal in restructuring and loss-making companies. In the pension system, minimum ages for qualifying for early retirement pension benefits will be raised and the benefit levels reduced for participation in career break and pre-pension schemes. Additional measures will be phased in more slowly. In 2013, these include working time calculations based on annual wages and enforcement of job search requirements for workers from the age of 55. In the pension system, the minimum age for early retirement benefits will be gradually increased to 62 years by 2016. Key to the success of these reforms will be their consistent implementation, including monitoring and removing if needed loopholes, e.g., in the system of sickness and disability benefits.

31. Nevertheless, further reforms may be needed to increase the employment rate by 5 percentage points as articulated in the government program. These could include some of the following specific measures:

  • Unemployment benefits: Unemployment benefits distort labor market incentives to reduce employment rates. Further reforms of benefits and eligibility requirements for workers below 25 years (including in waiting allowances) and above 55 years—supplemented by enhanced search support—could help increase the employment rate. The regression estimates in this note suggest that the government’s emphasis on activating the youngest and oldest labor market participants is the right one.

  • Pension and quasi-pension benefits: The combined unemployment and pension benefit systems provide incentives for labor market exit well before the official retirement age of 65 years. The recent reforms strengthen incentives to remain in the labor market beyond the age of 50 years but may be too modest to raise the effective retirement age well above 60 years. In addition, pension benefits could be made actuarially neutral at a discount rate of 5-6 percent.

  • Search assistance: Employment agencies allocate a smaller share of their resources to job search support than their OECD peers. A reallocation towards job search assistance and training could help to better match job seekers and vacancies across regions and provinces.

  • Wage setting mechanism: The wage setting mechanism produces exceptionally small regional and sectoral variation. Allowing greater wage flexibility across sectors and regions in the wage bargaining mechanism could help encourage labor market mobility.

References

  • Klepinger, D.H., T.R. Johnson and J.M. Joesch (2002), “Effects of Unemployment Insurance Work-Search Requirements: The Maryland Experiment”, Industrial and Labor Relations Review, No. 56, pp. 322.

    • Search Google Scholar
    • Export Citation
  • Kluve, J. (2006), “The Effectiveness of European Active Labour Market Policy”, IZA Discussion Paper, No. 2018, Bonn.

  • Borland, J. and Y.P. Tseng (2007), “Does a Minimum Job Search Requirement Reduce Time on Unemployment Payments? Evidence from the Jobseeker Diary in Australia”, Industrial and Labor Relations Review, Vol. 60, No. 3, April.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Davies Sara and Martin Hallet (2001), “Policy responses to regional unemployment: Lessons from Germany, Spain, and Italy,” European Economic-Economic Papers No. 161.

    • Search Google Scholar
    • Export Citation
  • Elhorst, Paul, (2000), “The Mystery of Regional Unemployment Differentials: A Survey of Theoretical and Empirical Explanations,” Journal of Economic Surveys, Vol. 17, Issue 5, pp. 709748.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • OECD (2007), “Activating the Unemployed: What Countries Do,” in OECD Employment Outlook 2007, Paris: OECD, pp. 207242.

  • OECD (2011), “Pensions at a Glance 2011: Retirement-Income Systems in OECD and G20 Countries,” Paris: OECD.

  • Economic Survey Belgium (2011), OECD.

  • IMF, (2011), “The Challenge of Pension Reform in Advanced and Emerging Countries,” IMF Board Paper (forthcoming).

  • Lopez-Bazo, Enrique, Tomas del Barrio, and Manuel Artis, (2002), “The regional distribution of Spanish unemployment: a spatial analysis,” Papers in Regional Science, Vol. 81, No. 3, pp. 365389.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Mastrobuoni, Giovanni and Taddei, Filippo, (2011), “Age Before Beauty? Productivity and Work vs. Seniority and Early RetirementNetspar Discussion Paper No. 05/2011-062. Available at SSRN: http://ssrn.com/abstract=1889054.

    • Search Google Scholar
    • Export Citation
  • Trampusch, Christine; Pierre Eichenberger; Micha de Roo; Robin, Bartlett Rissi; Isabelle Bieri; Isabelle; Laura Schmid; and Simon Steinlin (eds.) (2010), “Pension in Belgium,” REBECA (Research on Social Benefits in Collective Agreements). Database, Part 2 ‘Social Benefits in Collective Agreements’. SNF-Project No. 100012-119898. Institute of Political Science, University of Berne.

    • Search Google Scholar
    • Export Citation
  • Vandenbrande, Tom (2000), “First exploration of the Belgian HRST data”, OECD Focus Group Paper.

  • High Council of Finance (2011), “Comité d’Etude sur le Vieillissement,” Annual Report.

  • Adema, Willem, (2006), “Social Assistance Policy Development and the Provision of a Decent Level of Income in Selected OECD Countries,” OECD Social, Employment, and Migration Working Paper No. (2006)38.

    • Search Google Scholar
    • Export Citation

Appendix Table A.1.

Labor Market and Pension Reform Measures for 2012

Source: SPF Emploi, Travail, et Concertation Sociale, NBB.
1

Prepared by Franziska Ohnsorge.

2

Government-supported employment is assumed to include workers on service vouchers, those employed but supported by ONEm, and those employed part-time with ONEm support. Workers on service vouchers are typically workers who provide services to private households, such as cleaning. In 2004, these services are encouraged by consumer subsidies and preferential tax treatment, thus drawing many of them from the informal economy into the formal sector.

3

The standard deviation of unemployment rates scales the comparison by the number of regions, which are significantly fewer in Belgium than in France and Spain.

4

OECD (2011) estimated that for a worker with a family and an income one-third below the average wage the replacement rate is 86 percent.

5

Public sector pension schemes differ for civil servants, contractual employees, and public enterprises, with some differences also between levels of government. For civil servants at the federal, regional, and municipal levels, pension benefits are calculated on the basis of the past five years of service and are indexed to the wage scale which has historically increased by more than in the private sector (OECD, 2011). The retirement age of 65 is mandatory and several schemes of early retirement, invalidity benefits and waivers exist. Neither employers nor employees pay social security contributions for pensions for federal, regional, and municipal civil servants.

6

Workers on career-break or time-credit benefits, with top up payments by employers are similar to pre- pensioners.

7

A part-time pre-pension system that was available to employees working part-time will no longer be available from 2012.

8

Employment incentives include the Plan Win-Win introduced during the 2008/09 crisis. Under the Plan Win- Win, an employer receives €500–1100 per month for hiring a young, old, or long-term unemployed worker for a limited period of time.

9

Automatic price indexation is established by law for the public sector only but de facto often applied in the private sector. The base index for the automatic indexation has occasionally been redefined to remove temporary shocks. Currently, the indexation is based on the “health index” which excludes fuel prices and some items with heavy excise taxes from the CPI.

10

Lopez-Bazo, Barrio and Artis (2002) also control for region-specific unit labor cost, shares of agriculture, and human capital. The full list of covariates is not included here. The most recent labor force survey available from Eurostat (2008) only provides country-wide wage cost data, hence would be captured by the country dummy. Regional data on the share of agriculture in employment or value added are also not available. Regional educational attainment, while available, was insignificant when included in the regression below.

11

The regression results in Table III.1 are based on NUTS3 data. At the time of writing, the NUTS2 data—at the provincial level for Belgium—was available until 2010 while NUTS3 data was available only until 2009. If NUTS2 data were used, the results for are similar except that the coefficient on the share of below-25 year olds would lose significance in Column II of Table III.1. For the regression in Column III of Table 1 there would be insufficient degrees of freedom to estimate the coefficients precisely.

12

Ideally, the net replacement rate for pension benefits would be included in the regression. However, to date the dataset only includes pension data from the 2011 vintage of the OECD’s Pensions At A Glance. Hence, the panel dimension is lost; in the cross-section alone, coefficient estimates on pension benefits are insignificant.

Belgium: Selected Issues Paper
Author: International Monetary Fund