Abstract
Belgium has effected a remarkable fiscal adjustment, best illustrated by the decline in its public debt. While benefiting from an appreciable decline in interest rates, most of the underlying consolidation reflected a considerable increase in the tax burden, one of the highest in the Organisation for Economic Cooperation and Development. This paper analyzes the social transfer system in Belgium. Belgium has a very accessible and equitable health care system. The system is characterized by high input levels and service volumes.
I. Fiscal Consolidation and the Welfare State1
A. Introduction
1. Over the last decade, Belgium has effected a remarkable fiscal adjustment, best illustrated by the decline in its public debt from a high of 138 percent of GDP in 1993 to 100 percent at end 2003. While benefiting from an appreciable decline in interest rates, most of the underlying consolidation reflected a considerable increase in the tax burden, which has become one of the highest in the OECD. In addition, new challenges imply that the task is not yet over: potential growth has fallen behind that of other OECD countries, triggering a debate about the need to cut taxes, while population aging will soon begin to give rise to new budgetary pressures. Hence, identifying and implementing expenditure savings and measures to boost potential output become paramount.
2. Against this background, this chapter analyzes the social transfer system in Belgium.2 Three factors warrant its closer examination: first, it is a major component of all public spending, so that meaningful expenditure savings will require a reduction in social transfers; second, in international comparison the system exhibits important idiosyncrasies, which offer insights into fruitful directions for reform; and, third, in addition to the general efficiency gains that can be expected from trimming public expenditure—and the associated lowering of tax distortions—a streamlining of the social welfare system would remove significant disincentives to work.
3. This analysis suggests that there is sufficient scope to lower social transfers by some 1½ percent of GDP—above all by trimming unemployment-related spending—without putting social welfare achievements at risk. While the Belgian welfare state has produced excellent social welfare outcomes, it has done so at considerable cost. Budgetary outlays are large and in excess of most international comparators. Most of these expenditures are concentrated on work-income replacing transfers with adverse incentive effects on labor supply and, in turn, on output. Remedial measures that have relied on active labor market polices are unlikely to have helped much as indicated by their rather mixed track record in evaluations in a wide range of countries. Scaling back these policies offers scope for budgetary savings without adverse implications on labor market performance.
4. In what follows, section B places the Belgian welfare state in an international context, pointing out key differences with neighboring countries, the euro area, EU, and the OECD, which show that the social transfer system is both extensive and expensive. The quantitatively most important differences arise from unemployment-related expenditures. Section C examines the implications of these expenditures and their implied disincentive effects on employment. Section D critically discusses the traditional active employment policies, aimed at countering such disincentives. Section E concludes with an alternative reform proposal aimed at improving both work incentives and the fiscal position.
B. The Belgian Welfare State in International Perspective
General observations
5. Social transfer spending in Belgium remains substantial. Following the large buildup of the Belgian welfare state between the mid-1960s and the end of the 1970s, the system was significantly cut in the late 1980s (Deleeck (2003) and Cantillon et al., 2001). Still, with social expenditure amounting to 10½ percent of GDP in 2000 (excluding age- and health related spending), Belgium records the highest level against all comparators (Figure I.1).3 It is worthy to note that, while Belgium’s tax burden is high and social transfers are taxable, the above assessment of a generous welfare state remains valid, as social expenditures net of taxes remain at higher levels than in comparators (Adema, 2003).
Belgium: Social Expenditure, 1980-2000 1/
(In percent of GDP, excluding spending on old-age and health related spending)
Citation: IMF Staff Country Reports 2004, 048; 10.5089/9781451803198.002.A001
Source: OECD and Eurostat.1/ The “3 neighbors” are France, Germany and the Netherlands6. The incidence of transfers is high, as they accrue to a large part of the population. At close to 60 percent, Belgium has the highest benefit dependency ratio (including old-age transfers) among all OECD countries (Figure I-2), especially for women, of whom close to 100 percent receive some kind of benefit.4 When controlling for low labor force participation, and conducting the analysis in full-time-equivalents, benefit dependency exceeds 100 percent, the highest level in a recent cross-country study (NEI, 1999).5 Moreover, transfers are fairly evenly spread over the income distribution, with the top 20 percent highest income earners still receiving one fifth of their income from transfers (as compared to 40 percent for the bottom quintile).6
Belgium: Benefit Dependency Ratios, 1980-1999
Citation: IMF Staff Country Reports 2004, 048; 10.5089/9781451803198.002.A001
Sources: OECD and NEI (1999).7. Replacement rates for work-income replacing transfers are uneven, but high for the low paid (OECD 2003), with net replacement rates in excess of 90 percent for lone parents and joint earners, groups which make up ¾ of all unemployed (De Lathouwer, 2003). In addition, the real value of the benefits has risen over the 1990s (Ministry of Social Affairs 2002 and Cantillon et al., 2001). As benefits are based on wages up to a certain limit at which they are capped, the faster growth of real wages over the period has resulted in a convergence of the average benefit toward the maximum level, or, in other words, a significant compression of benefits paid (Cantillon et al., 2003).7 This compression explains the observed retrenchment of the standardized generosity of unemployment benefits, as defined in the OECD Summary Benefit Measure (Figure I.3); still generosity remains higher than elsewhere.
Belgium: OECD Summary Benefit Measure, 1979-99 1/
Citation: IMF Staff Country Reports 2004, 048; 10.5089/9781451803198.002.A001
Source: OECD (2002).1/ Defined as the average of the gross unemployment benefit replacement rates for two earnings levels, three family situations, and three durations of unemployment.8. Transfers are primarily geared toward income support—especially for the unemployed—with less emphasis on in-kind transfers and services (Table I.1). Unemployment-related transfers are significantly higher than elsewhere as is income support for survivors and families. In addition, family benefits through child allowances and the minimum guaranteed income have been drastically expanded between the mid-70s and mid-808, and since then held constant in real terms. On the other hand, and perhaps reflecting the above choices to replace labor market income, spending on other contingencies and particularly in-kind services in lieu of cash is lower, and at times much lower, than in other countries.
Social Transfer Spending (average 1990-98)
(In percent of GDP)
France, Germany, and Netherlands.
Comprising survivors, family-cash, and unemployment benefits.
Social Transfer Spending (average 1990-98)
(In percent of GDP)
| Belgium | OECD | EU-15 | euro area |
3 neighbors 1/ | |
|---|---|---|---|---|---|
| Disability cash benefits | 1.4 | 1.5 | 1.8 | 1.7 | 1.9 |
| Occupational injury payments | 0.4 | 0.3 | 0.3 | 0.3 | 0.3 |
| Sickness | 0.4 | 0.6 | 0.7 | 0.7 | 0.8 |
| Disability services | 0.1 | 0.8 | 0.9 | 0.5 | 0.7 |
| Survivors benefits | 2.6 | 1.1 | 1.4 | 1.6 | 1.1 |
| Family cash benefits | 2.1 | 1.4 | 1.5 | 1.4 | 1.5 |
| Family in kind benefits | 0.1 | 0.6 | 0.7 | 0.5 | 0.6 |
| Active labor market policies (ALMP) | 1.3 | 0.7 | 1.0 | 0.9 | 1.2 |
| Unemployment | 2.7 | 1.4 | 1.9 | 1.7 | 2.1 |
| Other contingencies | 0.3 | 0.5 | 0.4 | 0.4 | 0.6 |
| Memorandum items: | |||||
| Cash benefits | 9.7 | 6.6 | 7.8 | 7.6 | 8.0 |
| in-kind benefits | 0.2 | 1.3 | 1.6 | 1.1 | 1.3 |
| income-replacing benefits 2/ | 7.5 | 3.9 | 4.8 | 4.8 | 4.6 |
France, Germany, and Netherlands.
Comprising survivors, family-cash, and unemployment benefits.
Social Transfer Spending (average 1990-98)
(In percent of GDP)
| Belgium | OECD | EU-15 | euro area |
3 neighbors 1/ | |
|---|---|---|---|---|---|
| Disability cash benefits | 1.4 | 1.5 | 1.8 | 1.7 | 1.9 |
| Occupational injury payments | 0.4 | 0.3 | 0.3 | 0.3 | 0.3 |
| Sickness | 0.4 | 0.6 | 0.7 | 0.7 | 0.8 |
| Disability services | 0.1 | 0.8 | 0.9 | 0.5 | 0.7 |
| Survivors benefits | 2.6 | 1.1 | 1.4 | 1.6 | 1.1 |
| Family cash benefits | 2.1 | 1.4 | 1.5 | 1.4 | 1.5 |
| Family in kind benefits | 0.1 | 0.6 | 0.7 | 0.5 | 0.6 |
| Active labor market policies (ALMP) | 1.3 | 0.7 | 1.0 | 0.9 | 1.2 |
| Unemployment | 2.7 | 1.4 | 1.9 | 1.7 | 2.1 |
| Other contingencies | 0.3 | 0.5 | 0.4 | 0.4 | 0.6 |
| Memorandum items: | |||||
| Cash benefits | 9.7 | 6.6 | 7.8 | 7.6 | 8.0 |
| in-kind benefits | 0.2 | 1.3 | 1.6 | 1.1 | 1.3 |
| income-replacing benefits 2/ | 7.5 | 3.9 | 4.8 | 4.8 | 4.6 |
France, Germany, and Netherlands.
Comprising survivors, family-cash, and unemployment benefits.
9. In the aggregate, higher unemployment-related expenditure fully accounts for the difference in overall transfer spending between Belgium and its comparators (Figure I.4). Unemployment-related spending (including expenditure on active labor market policies) stands at some 4 percent of GDP, between 1 to 2 percentage points higher than in other countries. This elevated level is not the result of a particularly high unemployment rate, but rather due to high spending per unemployed (gearing ratio) (Figure I.5).8 This reflects the fact that the unemployment insurance system pays benefits to a large number of workers who are no longer searching for work and high active-labor-market-policy (ALMP) outlays, i.e., public expenditure on job creation.
Belgium: Composition of Social Transfers, 1985-2000
(In percent of GDP)
Citation: IMF Staff Country Reports 2004, 048; 10.5089/9781451803198.002.A001
Sources: OECD and Eurostat.Belgium: Gearing of Unemployment-Related Spending, 1985-2000 1/
Citation: IMF Staff Country Reports 2004, 048; 10.5089/9781451803198.002.A001
Sources: OECD and Eurostat, and staff calculations.1/ Gearing is defined as the share of unemployment-related spending to GDP divided by the OECD defined unemployment rate.10. Given the elevated level of unemployment-related outlays, most of the remainder of the analysis will focus on options for reform in this area. This does not imply that there are no possibilities for savings options in other transfer spending—exactly the opposite is suggested by the very high incidence for all social transfers—but instead that savings are probably easier identified in unemployment transfers. First, however, and to provide further motivation for reform, the economic and social implications of the welfare state are reviewed.
C. Implications of the Welfare State
11. Belgium has low poverty rates and high rates of income security, with thus favorable chances of societal participation for groups at risk from poverty. While any welfare system implies inherent and inevitable disincentives for work (Moffitt, 2000), some institutional aspects in Belgium accentuate these effects. Consequently, the employment performance of the economy has been lackluster, dampening potential output, and contributing to regional economic discrepancies.
Social Indicators
12. Welfare indicators are favorable. The general poverty level in Belgium is quite low by international comparisons. Any remaining deficiencies are likely limited to small groups, e.g., old people (65 and older) and the 18-24 year olds who appear to face a higher poverty risk (Forster, 2000). Notwithstanding some retrenchment in transfer spending over the 1990s (see above), the overall poverty rate was further reduced by almost 3 percentage points to some 7 percent by 2000, one of the lowest rates in the OECD, and, moreover, in contrast with mounting poverty in other countries. Finally, regional imbalances are smoothed, with the poverty rate in Wallonia—an area with considerably higher unemployment—only slightly in excess of the rest of the country (De Lathouwer, 2003).
13. Turning to the distribution of income, Belgium has a relatively low Gini coefficient, even though it has increased since the 1980s. Moreover, the share of low income in total has been drastically cut since the mid-1980s, with low-income incidence at the lowest level in the OECD. Not surprisingly, given the widely spread transfer system, income redistribution is mostly effected through the tax system rather than transfer payments (OECD 2003) with the tax level in turn being more important than the underlying progressivity of the tax rates.9
Employment Performance
14. Social transfer payments require revenue-raising operations which in general introduce distortions and losses in economic efficiency (Auerbach and Hines, 1999) and Krueger and Meyer, 2002). The most important of these distortions affects the labor supply decision, as it is impossible to tax leisure (see also Lucas, 2003). The empirical growth literature has accordingly identified the tax wedge as the most problematic aspect. In line with its high revenue ratio, Belgium has the highest tax wedge in the OECD (Figure I.6),10 notably higher than France, where the tax wedge has been singled out as the key factor in explaining lower per capita income as compared to the US (Prescott, 2003). For Belgium, similar results are implied by Drèze and Modigliani (1981), Daveri and Tabellini (2000) and Konings and Roodhooft (1997).11
Belgium: Tax Wedge, 1980-2001
(In percent)
Citation: IMF Staff Country Reports 2004, 048; 10.5089/9781451803198.002.A001
Source: OECD.15. The transfer system—in addition to its indirect adverse effects through the need to raise revenue—also contains significant direct disincentives to work. Generous welfare transfers, particularly to specific groups result in higher reservation wages and lower Participation.12 Thus, economic efficiency gains may be generated if such transfers are cut back and incentives for work strengthened.
16. The Belgian employment rate, at some 60 percent, is among the lowest internationally and, based on the OECD definition, Belgium records the highest non-employment rate of working-age households (Figure I.7). The overall employment rate is pulled down by low participation of the young and old, and overall low employment rates in the Walloon region (Table I-2).
Belgium: Non-Employment Rates for Working-Age Households
(In percent)
Citation: IMF Staff Country Reports 2004, 048; 10.5089/9781451803198.002.A001
Source: OECD.Employment Rates by Age Group, 2002
(In percent)
Employment Rates by Age Group, 2002
(In percent)
| Belgium | Flanders | Wallonia | EU-15 | ||
|---|---|---|---|---|---|
| 15-64 years old | 59.6 | 63.1 | 55.6 | 64.2 | |
| male | 68.1 | 71.4 | 64.9 | 72.9 | |
| female | 51.1 | 54.6 | 46.3 | 55.5 | |
| 15-24 years old | 28.6 | 33.3 | 22.4 | 40.5 | |
| 55-64 years old | 25.7 | 24.7 | 25.7 | 39.8 | |
Employment Rates by Age Group, 2002
(In percent)
| Belgium | Flanders | Wallonia | EU-15 | ||
|---|---|---|---|---|---|
| 15-64 years old | 59.6 | 63.1 | 55.6 | 64.2 | |
| male | 68.1 | 71.4 | 64.9 | 72.9 | |
| female | 51.1 | 54.6 | 46.3 | 55.5 | |
| 15-24 years old | 28.6 | 33.3 | 22.4 | 40.5 | |
| 55-64 years old | 25.7 | 24.7 | 25.7 | 39.8 | |
17. These patterns are consistent with the particularly large reservation wages that transfers provide to groups with the lowest participation rates.
After an initial 9-month waiting period, workers can qualify for unemployment benefits even without a formal work history.
Unemployment benefits are in principle open-ended, conditional on job-search requirements. 13 With more than half of all unemployment being in excess of one year, participation. Thus, economic efficiency gains maybe generated if such transfers are cut back and incentives for work strengthened. Belgium records one of the highest incidences of long term unemployment among OECD countries.14 Attempts to mitigate the disincentives from open-endedness of benefits called for strict sanctions, e.g., with an unemployed’s failure to show up for an interview at the public job search agency automatically triggering a temporary loss of benefits, and repeated failure resulting in the permanent loss of all benefits. However, in the absence of an alternative social safety net for the unemployed, these sanctions have not been enforced—helped by the fact that the regional job-search agencies need to report any offenses to the federal unemployment insurance before the latter can stop benefit payments. A modification introduced in 2001 which limits the sanctions and was thus hoped to trigger more actual enforcement, has also failed to meet expectations.
There are generous early-retirement provisions. While the traditional “prepension” system of early retirement payments is being phased out, a new generous scheme for “older unemployed” has been instituted and the overall number of those on early retirement schemes has slightly increased with the equivalent of 6 percent of the labor force in early retirement in 2000. Belgian public spending on early retirement, ½ percent of GDP in 2000, remained by far the highest in the OECD—more than three times the average of the euro zone or its 3 neighboring countries (Figure I.8). Meanwhile, the average retirement age—53 into the old-age unemployment scheme—continued to fall.
Differential replacement rates imply disincentives for labor supply for the low skilled. Unemployment benefits are capped at a fraction of the average wage, thus providing lower replacement rates for higher income earners, where the impact of replacement incomes on the reservation wage rapidly diminishes. Indeed, survey evidence shows that low wage unemployed are not interested in finding a job, and large employment traps have opened up (D’Addio et al., 2002). Accordingly, a considerable share of the unemployed has never worked (De Lathouwer, 2003).15
Benefit levels do not adequately reflect regional productivity variations, such that the transfer system promulgates the adverse effects of limited wage flexibility. There is evidence that the substantial differences in regional productivity (c.f. Bell, 2003) are only insufficiently incorporated into actual wages.16 Since wages form the basis for the determination of unemployment benefits, a larger difference between reservation wages and labor-market clearing equilibrium wages is implied in geographic areas with lower productivity and, thus, higher unemployment than elsewhere in the country. Moreover, open-ended benefits prevent this discrepancy from adjusting with unemployment duration. These observations are consistent with the persistent unemployment in Belgium’s southern areas—reflected, for example, by a Walloon unemployment rate of 10 percent in 2001, compared to 4 percent in Flanders. In Flanders, consistent with the less binding reservation wages, participation is higher, even among the traditional problem groups, except for the old, where generous early retirement provisions are at play.17
Other traditional problem groups do not show adverse employment records. Notably the employment rate of the disabled is relatively higher in Belgium, as compared internationally, which again is consistent with the relatively less generous and broad benefit entitlements (especially as regards eligibility) in Belgium (OECD, 2002).
Belgium: Public Expenditure on Early Retirement, 1985-2000
(In percent of GDP)
Citation: IMF Staff Country Reports 2004, 048; 10.5089/9781451803198.002.A001
Source: OECD.D. Employment Policy
18. Resolving Belgium’s poor employment performance has been on the policy agenda for some time. However, the link between the functioning of the labor market and the welfare system has so far not been systematically focused on, with, instead, greater reliance being placed on active remedial measures in the form of additional expenditure on ALMP, see Annex I.1. There are several noteworthy aspects.
Belgium has consistently outspent most international comparators on ALMPs. In 2000, the last year for which comparable data are available, Belgium spent some 1 ⅓ percent of GDP on ALMPs, compared to spending of less than 1 percent of GDP in the EU, the euro area, and the OECD, though Belgian spending is in line with its closest neighbors (Figure I.9). Moreover, this spending level has remained fairly constant—and in excess of most comparators—since the beginning of the sample in 1985.
ALMPs in Belgium also reach a larger segment of the labor force, but are likely spread thin. With a large volume of different schemes—220 were identified in the most recent audit of such policies (CSE, 2003)—ALMP participation exceeds unemployment18 (FPB, 2003). At the same time, spending per participant is significantly lower.
The composition of Belgian ALMPs has undergone significant changes. While expenditure on public employment services has remained largely flat over the 1985-2000 period, there have been some increases in expenditure on training—concentrated on training for already employed—and a slight decline in ALMP spending for the handicapped. The most pronounced change, however, was the 50 percent cut of public employment programs, to ½ percent of GDP, while expenditure on private sector wage subsidies grew by an equivalent amount.
The mix of Belgian ALMPs is different from international comparators. On an international scale, the last 15 years have witnessed a decline in the relative importance of public employment search assistance and training while spending on problems groups (such as youth and the disabled) gained prominence. Belgium, on the other hand, has continued to place considerable importance on public employment programs, while emphasizing employment subsidies to a significantly larger degree than elsewhere (Figure I.10).
Belgium: ALMP Spending, 1985-2000
(In percent of GDP)
Citation: IMF Staff Country Reports 2004, 048; 10.5089/9781451803198.002.A001
Source: OECD.Belgium: Composition of ALMPS, 1985-2000
(In percent of GDP)
Citation: IMF Staff Country Reports 2004, 048; 10.5089/9781451803198.002.A001
Source: OECD and Eurostat.Evaluating ALMP
19. With few evaluations of Belgian ALMPs (Cockx, 2003, Cockx and Ridder, 2001, Cockx et al., 1998 and Lopez-Novella, 2003) it is useful to survey the relevant European empirical evaluation literature to distill some findings that can be expected to generalize into the Belgian context. As economic social policy evaluation was pioneered for ALMPs policies—notably the study of different training schemes in the US (see Heckman et al., 1999)—a vast body of literature is available and methodological issues to be addressed in any evaluation are well known (Box I.1).
20. In general, and even with the methodological caveats, the numerous empirical evaluations permit a fairly robust assessment that the case for ALMP is not a particularly strong one. Table I.3 summarizes the key findings of evaluations of European ALMP.19 No matter which evaluation method is utilized, the majority of studies fails to find any statistically significant beneficial effects of ALMP. Some exceptions to this rule relate to carefully targeted and limited interventions, schemes that combine on-the-job training with extensive monitoring—particularly youth measures—as well as subsidy schemes. On the other hand, the record for other training programs and even employment programs is poor.
Key Results of Evaluations of European Active Labor Market Policies
Key Results of Evaluations of European Active Labor Market Policies
| Training Programs | Youth measures | Enployment programs | |||||||
|---|---|---|---|---|---|---|---|---|---|
| Overall effect | Overall effect | Overall effect | |||||||
| Negative | No effect | Positive | Negative | No effect | Positive | Negative | No effect | Positive | |
| 1. Microeconometric studies | Regner (2001), earnings | Larsson (2000) annual earnings | Kluve et al. (1999). employment | Dolton et al. (1994a); unemployment hazard, net of time in training 8/ | Larsson (20000): earning | Jenssen (1999), unemployment duration | Kluve et al. (1999), employment | Payne et al. (1996); hourly wages 10/ | |
Lechner (2001); employment probability, earning Edin (1988); weekly earnings Andersson (1993); annual earnings |
Raaurm and Torp (2001), earnings Bell et al. (1999), productivity Gerfin and Lechner (2002), employment lensen et at. (1993); unemployment Westergard-Nielsen (1993); hourly wages 2/ Thierry and Sollogub (1993); employment hazard Breen (1991), employment rate Ridder (1986): employment hazard, unemployment hazard Coocks (2003): unemployment hazard Ackurn (l99l): hourly wages Harkman rial. (1996); employment rate, hourly wage Torp et al. (1993): employmen rate Payne et al.(1996): hourly wages |
van Ours (2001); job Folding rate Brodety et al. (2001); employment Zweimaller and Winter- Ebner (l996); unemployment Bonnat et al. (1997) imemployment hazard, employmenl hazard Kraus et al. (1997); employment hazard O’connell and MeGinnity (1997); empolyment rate, wage de Koning (1991) unemployment rete, hourly wage Axelsson (1989); annual earnings payne et al. (1996); employment rate |
Brodaty et al. (2001): employment O’Connelt and McGinnity (1997); employment rate, wage 4/ de Koning (1991), unemployment rate 6/ Main and Raffe (1983); employmenl rate 7/ Whitfiled and Bourlakis (1991); hourly wage 8/ Main (1991); employment rate 9/ Green et al. (1996); hourly wage 8/ Dolton et al.(l992); employment rate, hourly wage 8/ Dolton et at, (1994b); hourly wage 8/ |
Gerfin and Lechner (2002), employment Breen (1991), employment rate 4/ Main (1985); employment rate 8/ Whitfiled and Bourlakii (l99l); employment rate 8/ Main and Shelly (1990); employment rata 8/ o’Higgins (l994); employment rate 8/ |
Gerfin and Lechner (2002), employment Ridder (1986); employment hazard 5/ Cockx and Ridder (2003); employment hazard |
Ridder (1986); unemployment hazard |
|||
| II. Experimental studies | |||||||||
| III. Macro studies 1/ | OECD (1993); unemployment 11/ | ||||||||
Key Results of Evaluations of European Active Labor Market Policies
| Training Programs | Youth measures | Enployment programs | |||||||
|---|---|---|---|---|---|---|---|---|---|
| Overall effect | Overall effect | Overall effect | |||||||
| Negative | No effect | Positive | Negative | No effect | Positive | Negative | No effect | Positive | |
| 1. Microeconometric studies | Regner (2001), earnings | Larsson (2000) annual earnings | Kluve et al. (1999). employment | Dolton et al. (1994a); unemployment hazard, net of time in training 8/ | Larsson (20000): earning | Jenssen (1999), unemployment duration | Kluve et al. (1999), employment | Payne et al. (1996); hourly wages 10/ | |
Lechner (2001); employment probability, earning Edin (1988); weekly earnings Andersson (1993); annual earnings |
Raaurm and Torp (2001), earnings Bell et al. (1999), productivity Gerfin and Lechner (2002), employment lensen et at. (1993); unemployment Westergard-Nielsen (1993); hourly wages 2/ Thierry and Sollogub (1993); employment hazard Breen (1991), employment rate Ridder (1986): employment hazard, unemployment hazard Coocks (2003): unemployment hazard Ackurn (l99l): hourly wages Harkman rial. (1996); employment rate, hourly wage Torp et al. (1993): employmen rate Payne et al.(1996): hourly wages |
van Ours (2001); job Folding rate Brodety et al. (2001); employment Zweimaller and Winter- Ebner (l996); unemployment Bonnat et al. (1997) imemployment hazard, employmenl hazard Kraus et al. (1997); employment hazard O’connell and MeGinnity (1997); empolyment rate, wage de Koning (1991) unemployment rete, hourly wage Axelsson (1989); annual earnings payne et al. (1996); employment rate |
Brodaty et al. (2001): employment O’Connelt and McGinnity (1997); employment rate, wage 4/ de Koning (1991), unemployment rate 6/ Main and Raffe (1983); employmenl rate 7/ Whitfiled and Bourlakis (1991); hourly wage 8/ Main (1991); employment rate 9/ Green et al. (1996); hourly wage 8/ Dolton et al.(l992); employment rate, hourly wage 8/ Dolton et at, (1994b); hourly wage 8/ |
Gerfin and Lechner (2002), employment Breen (1991), employment rate 4/ Main (1985); employment rate 8/ Whitfiled and Bourlakii (l99l); employment rate 8/ Main and Shelly (1990); employment rata 8/ o’Higgins (l994); employment rate 8/ |
Gerfin and Lechner (2002), employment Ridder (1986); employment hazard 5/ Cockx and Ridder (2003); employment hazard |
Ridder (1986); unemployment hazard |
|||
| II. Experimental studies | |||||||||
| III. Macro studies 1/ | OECD (1993); unemployment 11/ | ||||||||
Key Results of Evaluations of European Active Labor Market Policies (concluded)
Where possible, the results report The employmenl effect corrected for participation in ALMP programs, at calculated by Calmfors et al. (2002).
Positive effect for unskilled male workers.
Positive employment hazard for workers with diploma.
Measure evaluated was temporary employment program in Ireland.
For older workers, no effect found for younger workers.
Measure evaluated was on-the-job training.
Measure evaluated was training, positive effect found for females.
Measure evaluated was training.
Measure evaluated was training, positive effect found for "disadvantaged" youth.
Measure evaluated was subsidized employment in public and non-profit sectors.
Interaction effect between GDP and public employment expenditure.
Positive effect for long-term unemployment.
Effect found in 1990s; no effect found in 1980s
Key Results of Evaluations of European Active Labor Market Policies (concluded)
| Employment subsidies | Job search assistance | Summary ALMP measure | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Overall effect | Overall effect | Overall effect | ||||||||
| Negative | No effect | Positive | Negative | No effect | Positive | Negative | No effect | Positive | ||
| I. Macroeconometric studies | Kluve et al.(1999), employment | Bell et al. (1999), productivity van Ours (2001); job finding rate Bonnal et al. (1997); unemployment hazard, employment hazard 3/ |
Gerfin and Lechner (2002), employment Rosholm (1999); employment hazard |
Dolton ind O’Neill (1996); unemiployment hizard | Bjorklund (1993); unemployment hazard | Sianesi (2001); employment probability | ||||
| II. Experimental studies | van den Berg and van der Klaauw (2000); unemployment hazard | Bjorkland and Regner (1998); employment rate | ||||||||
White and Lakey (1992); employment rate |
||||||||||
| III. Macro studies 1/ | Beltmann and Jackman (1996); employment | OECD (l993); unemployment Forslund and Krueger (1994); unemployment Bellmann and Jackman (1996); unemployment Jackman et al. (1996); unemployment 12/ Elmeskov et al. (1996); unempolyment Elmeskov et at. (1996); unemployment |
Layard et al.(1991); unemployment Heylen (l993); unemployment Zetterberg (1993); unemployment Nickel and Layard (1999); unemployment Blanchard and Wolfers (2000); unemployment Estevao (2003); employment 14/ |
|||||||
Where possible, the results report The employmenl effect corrected for participation in ALMP programs, at calculated by Calmfors et al. (2002).
Positive effect for unskilled male workers.
Positive employment hazard for workers with diploma.
Measure evaluated was temporary employment program in Ireland.
For older workers, no effect found for younger workers.
Measure evaluated was on-the-job training.
Measure evaluated was training, positive effect found for females.
Measure evaluated was training.
Measure evaluated was training, positive effect found for "disadvantaged" youth.
Measure evaluated was subsidized employment in public and non-profit sectors.
Interaction effect between GDP and public employment expenditure.
Positive effect for long-term unemployment.
Effect found in 1990s; no effect found in 1980s
Key Results of Evaluations of European Active Labor Market Policies (concluded)
| Employment subsidies | Job search assistance | Summary ALMP measure | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Overall effect | Overall effect | Overall effect | ||||||||
| Negative | No effect | Positive | Negative | No effect | Positive | Negative | No effect | Positive | ||
| I. Macroeconometric studies | Kluve et al.(1999), employment | Bell et al. (1999), productivity van Ours (2001); job finding rate Bonnal et al. (1997); unemployment hazard, employment hazard 3/ |
Gerfin and Lechner (2002), employment Rosholm (1999); employment hazard |
Dolton ind O’Neill (1996); unemiployment hizard | Bjorklund (1993); unemployment hazard | Sianesi (2001); employment probability | ||||
| II. Experimental studies | van den Berg and van der Klaauw (2000); unemployment hazard | Bjorkland and Regner (1998); employment rate | ||||||||
White and Lakey (1992); employment rate |
||||||||||
| III. Macro studies 1/ | Beltmann and Jackman (1996); employment | OECD (l993); unemployment Forslund and Krueger (1994); unemployment Bellmann and Jackman (1996); unemployment Jackman et al. (1996); unemployment 12/ Elmeskov et al. (1996); unempolyment Elmeskov et at. (1996); unemployment |
Layard et al.(1991); unemployment Heylen (l993); unemployment Zetterberg (1993); unemployment Nickel and Layard (1999); unemployment Blanchard and Wolfers (2000); unemployment Estevao (2003); employment 14/ |
|||||||
Where possible, the results report The employmenl effect corrected for participation in ALMP programs, at calculated by Calmfors et al. (2002).
Positive effect for unskilled male workers.
Positive employment hazard for workers with diploma.
Measure evaluated was temporary employment program in Ireland.
For older workers, no effect found for younger workers.
Measure evaluated was on-the-job training.
Measure evaluated was training, positive effect found for females.
Measure evaluated was training.
Measure evaluated was training, positive effect found for "disadvantaged" youth.
Measure evaluated was subsidized employment in public and non-profit sectors.
Interaction effect between GDP and public employment expenditure.
Positive effect for long-term unemployment.
Effect found in 1990s; no effect found in 1980s
21. In Belgium, ALMPs appear to have focused on the less useful measures. The high incidence of programs over a large group of participants in Belgium contrasts with the highly targeted and involving programs that appear to offer some benefit, especially on-the-job training (Gerfin and Lechner, 2002). In addition, the record of public employment programs, on which still ⅓ of one percent of GDP is spent in Belgium, is particularly poor from the perspective of moving participants into private-sector jobs (Cockx and Ridder, 2001).20 Experimental studies point to the benefits of job search assistance, which are also relatively underdeveloped, but, as indicated in Box L1, experimental results may not easily generalize.
The Evaluation Problem
Assume that an individual i’s participation in an ALMP program is captured by the indicator variable D, such that if D=1 an individual participates in a program, or D=0, if not. In the language of the evaluation literature, D=1 indicates “treatment” (i.e., participation in ALMP, e.g., a training program). Assume further that the policy maker is interested in the outcome variable Y (e.g., earnings or employment status), with the treatment status formally being captured in a suffix. In order to evaluate the impact of an ALMP for an individual i, one would want to know:
Accounting for the fact that the treatment will not yield the same result for all workers, these individual treatment effects would need to be aggregated into a population average.
Two fundamental problems arise at this stage, and are indeed inherent in evaluating any social program. The first stems from the fact that one cannot observe the same individual contemporaneously in two states, having benefited from a program and not having benefited from a program, i.e. Yli and YOi do not exist at the same time. The literature has developed several approaches to overcome this problem by constructing a counterfactual, each with its own strengths and weaknesses (for a fuller discussion of this critical issue see Heckman et al., 1999 and Kluve and Schmidt, 2003). The second set of problems arises from the possibility that in an economy, any individual effect can easily be offset, or overturned, on a more aggregate level. For example, workers who find employment after participating in ALMP may have been hired anyway (deadweight loss), or they may be replacing other workers (substitution effect), hi addition, the cost of financing ALMPs may result in adverse general equilibrium effects, such as loss of employment from higher taxes, that are not captured in the evaluation sample.
The following methodological approaches maybe distinguished:
Microeconometric studies seek to explicitly model the individual choice to participate in an ALMP. Hence, they condition the participation in an ALMP on a set of observable variables X. With this information, the bias
The experimental approach seeks to randomly assign individuals into treatment and control groups, such that the unobservable effect of nontreatment of program participants can be estimated by the nonparticipant population outcomes:
Macroeconometric studies focus directly on the population averages, interpreting different time periods and/or countries as different outcomes of one basic experiment. They do not concern themselves with the details of the individual ALMP treatment effect on individuals, thus offer little hope of revealing any insight at all into key structural questions of the individual labor supply decision. In addition, the links between ALMP (typically individual specific programs) and the aggregate variables (e.g. unemployment) are highly tenuous, subject to simultaneity bias or reverse causality (e.g., spending on training might increase when unemployment rises, with no causality running between the former and the latter), and thus require strong and untestable assumptions to identify any effect. Macro studies would, however, be able to estimate aggregate deadweight and substitution effects.
22. It also appears that the regional differences in ALMPs are not well tailored to the specific labor market problems in different geographic areas. The Belgian regions and communities are pursuing a different mix of ALMP (Table I.4). The high share of training programs in Wallonia and the German community appears at odds with the observation that reservation-wage and minimum-wage unemployment are important, which would preferably be addressed by wage subsidies. In contrast in Flanders, the less problem-group specific ALMPs and the larger share of wage subsidies are not generally in tune with the Flemish labor market. Finally, coordinating a large volume of policies is reflected in high overhead costs in the smaller areas of Brussels and the German Community.
Share of Specific ALMP Programs, 2002 1/
Shares express the number of specific programs in relation to all programs.
Share of Specific ALMP Programs, 2002 1/
| Wallonia | Flanders | Federal Government |
German Community |
Brussels | ||
|---|---|---|---|---|---|---|
| (Shares in percent, unless otherwise indicated) | ||||||
| Training measures | 54.8 | 29.2 | 15.8 | 48.4 | 27.7 | |
| Classroom training | 28.6 | 16.7 | 6.3 | 22.6 | 2.8 | |
| On-the-job training | 14.3 | 8.3 | 6.3 | 12.9 | 13.9 | |
| Search training | 11.9 | 4.2 | 3.2 | 12.9 | 11.0 | |
| Employment subsidies | 11.9 | 22.9 | 30.2 | 6.5 | 19.4 | |
| Public employment programs | 21.4 | 10.4 | 12.7 | 12.9 | 16.7 | |
| Public employment services | 2.4 | 0.0 | 1.6 | 3.2 | 11.1 | |
| Coordinating measures | 11.9 | 18.8 | 4.8 | 22.6 | 25.0 | |
| Total number of ALMP measures | 42 | 48 | 63 | 31 | 36 | |
Shares express the number of specific programs in relation to all programs.
Share of Specific ALMP Programs, 2002 1/
| Wallonia | Flanders | Federal Government |
German Community |
Brussels | ||
|---|---|---|---|---|---|---|
| (Shares in percent, unless otherwise indicated) | ||||||
| Training measures | 54.8 | 29.2 | 15.8 | 48.4 | 27.7 | |
| Classroom training | 28.6 | 16.7 | 6.3 | 22.6 | 2.8 | |
| On-the-job training | 14.3 | 8.3 | 6.3 | 12.9 | 13.9 | |
| Search training | 11.9 | 4.2 | 3.2 | 12.9 | 11.0 | |
| Employment subsidies | 11.9 | 22.9 | 30.2 | 6.5 | 19.4 | |
| Public employment programs | 21.4 | 10.4 | 12.7 | 12.9 | 16.7 | |
| Public employment services | 2.4 | 0.0 | 1.6 | 3.2 | 11.1 | |
| Coordinating measures | 11.9 | 18.8 | 4.8 | 22.6 | 25.0 | |
| Total number of ALMP measures | 42 | 48 | 63 | 31 | 36 | |
Shares express the number of specific programs in relation to all programs.
23. While there is some evidence that targeted subsidies raise employment (Estevao, 2003) they are not problem-free. Belgium has effected a shift from other ALMPs to such subsidies over the last 10 years, particularly by promoting low-wage employment through cuts in social security contributions. This may have improved the performance of ALMPs, but the subsidies to the low-wage earners also reflected concerns to limit the adverse incentive effects from the concurrent boost of social benefits (Cantillon et al., 2002 and De Lathouwer, 2003). In addition these targeted cuts have also created “wage” traps at slightly higher levels of the income distribution—where the reductions run out—and the highest marginal tax rates emerge (Figure I.11). Finally, these subsidies also entail new (tax) expenditures for the budget. Against this background, a general tax cut, financed through expenditure cuts—centered on the least effective transfer programs—holds promise.
Belgium: Marginal Effective Tax Rates
Citation: IMF Staff Country Reports 2004, 048; 10.5089/9781451803198.002.A001
Source: OECD (2003)E. Concluding Remarks: An Alternative Approach—Scaling Back the Welfare State
24. Unemployment-related transfers in Belgium are appreciably higher than in a wide sample of comparator countries. Considerable budgetary savings on the order to 1½-2 percent of GDP could be realized by:
a. sharply curtailing early-retirement provisions: this should yield steady state savings of ½ of one percentage point of GDP;
b. phasing out open-ended unemployment benefits; e.g., ending unemployment benefits after one year, as is currently discussed in Germany and the Netherlands, would save up to 1 percent of GDP in unemployment insurance transfers—with these saving being partially offset by the need to increase minimum income payments, in particular if other structural distortions like elevated sectoral minimum wages are not redressed;
c. substantially scaling back ALMPs through phasing out public employment programs and halving training expenditures: savings could amount to ⅔ A of one percentage point of GDP;
d. additional substantial savings could result from a better targeting of transfer payments to a reduced number of beneficiaries, their quantification is, however, beyond the scope of this paper.
25. Such savings would be key in underpinning the twin government targets of fiscal consolidation and tax cuts, while vitalizing the labor market. Savings of 1½ percentage point of GDP would lower primary spending by an equivalent amount and, if implemented over a four-year horizon, would reduce real primary spending by almost 1 percent per year. Moreover, flanked by additional reforms of labor market institutions, benefit reform would induce considerable additional employment based on the empirical work surveyed above. As this employment would primarily benefit currently marginalized labor market groups, the increased supply would contribute to economy-wide wage moderation and the preservation of competitiveness.
26. What about social welfare? It is difficult to unambiguously predict the actual effect of the proposed reforms. A standard neoclassical model would show that benefit recipients who after a cut in benefits would have to accept work—and thus forego leisure—would be worse off. Society as a whole, however, would of course benefit from a reduction of the deadweight loss associated with low participation, and this societal gain may be redistributed to the previous benefit recipients in ways that eschew disincentives to work (e.g., through general tax cuts). Moreover, the adverse effect on the welfare of previous benefit recipients may be less pronounced in practice, particularly if there are nonmonetary social benefits from work socialization or if burdensome administrative control requirements are reduced.
27. There is also a fairness dimension, given that the current targeted wage subsidies cause the highest marginal effective tax rates for some groups of low wage earners. As regards income distribution, this outcome is undesirable. Moreover, with the current tax cuts likely resulting in less redistribution—and given the need to generally lower marginal tax rates—removing the highest marginal taxes on low earners would undoubtedly be helpful in limiting a rise in inequality.
ANNEX I.1 Active Labor Market Policies
ALMP generally seek to boost the labor market chances (especially earnings and employability) of so called problem groups. Depending on the diagnosis of the unemployment problem, different strands of ALMP may be pursued (c.f. Calmfors, 1993 and OECD, 1993).
Search (frictional) unemployment could be lowered by measures to improve workers’ and employers’ information about each other, such as public job search agencies, and targeted efforts to help workers search better.
Structural unemployment may be addressed through government programs to redress the mismatch between labor supply and demand. For example, training measures could help workers develop the skills required by employers, or mobility assistance could provide incentives for workers to move to geographic areas with excess labor demand.
Minimum-wage unemployment—unemployment that arises when, for legal or collective bargaining reasons, the wage cannot fall to a level that would clear the labor market—could be addressed through measures to lower the effective minimum wage. The same channel would work for replacement rates—and thus reservation wages—that are too high. Measures in this regard are wage subsidies, e.g. through targeted cuts in nonwage labor costs to the employer, e.g. lower social security contributions.
Aggregate demand induced unemployment, would call for macroeconomic stabilization policies. However, direct deficit-financed employment programs in the public sector would also be helpful in this context.
These different motivations call for different policies, and it is important to realize that ALMP measures can have different effects, depending on the nature of the unemployment problem. For example, increased training will do nothing to lower frictional unemployment, and its effects on minimum-wage unemployment are also far from clear,21 Similarly, job search agencies will not be able to redress skill mismatches or binding minimum wages, while public employment may e.g. worsen frictional unemployment, as job seekers could hold out for a safe job in the public sector.
References
Ackum, S., 1991: Youth Unemployment, Labor Market Programs, and Subsequent Earnings, Scandinavian Journal of Economics, Vol. 93, No. 4, pp. 531- 43.
Adema, W., 2003: Net Social Expenditure, OECD, mimeo.
Andersson, H., 1993: Choosing among Alternative Nonexperimental Methods for Estimating the Impact of Training: New Swedish Evidence, Swedish Institute for Social Research, mimeo.
Auerbach, A. and J. Hines, 1999: Taxation and Economic Efficiency, in: Auerbach, A. and M. Feldstein (eds.) Handbook of Public Economics, Volume 3, Chapter 21.
Axelsson, R., 1989: Svensk Arbetsmarktnadsutbildning, Umea University, mimeo.
Bell, G., 2003: Regional Economic Development in Belgium, Belgium: Selected Issues, IMF Country Report No. 03/50.
Bell, B., Blundell, R, and J. van Reenen, 1999: Getting the Unemployed Back to Work: The Role of Targeted Wage Subsidies, International Tax and Policy Finance, Vol. 6, pp. 339- 360.
Bellmann, L. and R. Jackman, 1996: The Impact of Labor Market Policy on Wages, Employment and Labor Market Mismatch, in Schmid, G., O’Reilly, J. and K. Schomann (eds): International Handbook of Labour Market Policy and Evaluation, pp. 89- 114, Edward Elgar, Aldershott.
Björklund, A., 1993: Evaluations of Training Programs: Experiences and Suggestions for Future Research, Wissenschaftszentrum Berlin Discussion Paper 89
Björklund, A., 1994: Evaluations of Swedish Lab or Market Policy, International Journal of Manpower, Vol. 15, No. 5, pp. 16- 31.
and H. Regner, 1998: Experiment al Evaluation of European Labour Market Policy, in Schmid, G., O’Reilly, J. and K. Schomann (eds): International Handbook of Labour Market Policy and Evaluation, pp. 89- 114, Edward Elgar, Aldershott.
Blanchard, O. and J. Wolfers, 2000: The Role of Shocks and Institutions in the Rise of European Unemployment: The Aggregate Evidence, Economic Journal, Vol. 110 (March), pp. 1- 33.
Bonnal, L, Fougère, D. and A Sérandon, 1997: Evaluating the Impact of French Employment Policies on Individual Labor Market Histories, Review of Economic Studies, Vol. 64, pp. 683- 713.
Brodaty, T, Crépon, B and D. Fougère, 2001: Using Matching Estimators to Evaluate Alternative Youth Employment Programs: Evidence for France, 1986-88, in: Lecher, M. and F. Pfeiffer (eds): Econometric Evaluation of Labor Market Programs.
Breen, R 1991: Assessing the Effectiveness of Training and Temporary Employment Schemes: Some Results for the Youth Labor Market, The Economic and Social Review, Vol. 22, No. 3, pp. 177- 98.
Calmfors, L., 1994: Active Labor Market Policy and Unemployment — A Framework for the Analysis of Crucial Design Features, OECD Economic Studies 22.
Cantillon, B. De Maesschalk, V. and R. Van Dam, 2001: Welvaartsvastheid en Adequaatheid van de sociale Minima, Centrum voor Sociaal Beleid Berichten, November.
Cantillon, B. De Lathouwer, L. Marx, I., Van Dam, R. and K. Van den Bosch, 2002: Sociale Indicatooren, Centrum voor Sociaal Beleid.
Cantillon, B. Marx, I. and De Maesschalk, V., 2003a: DeBodem van de Welvaartsstaat van 1970 tot nu, en daarma, Centrum voor Sociaal Beleid Berichten, March.
_________, (2003b): Le Trilleme de la Securite Sociale: Passe, Futur et Present, in Revue de la Securite Sociale Beige, pp. 397- 433.
Cockx, B., 2003: Vocational Training of Unemployed Workers in Belgium, in Applied Economics Quarterly / Konjunkturpolitik, Vol. 49, No.1, pp 23- 48.
_________, and G. Ridder, 2001: Social Employment of Welfare Recipients in Belgium: An Evaluation, in: Economic Journal, Vol. III (April), pp. 322- 352.
Cockx, B., Van der Linden, B., and A. Karaa, 1998: Active Labor Market Polices and Job Tenure, Oxford Economic Papers, Vol. 50, pp. 685- 708.
Conseil Superieur de l’Emploi (CSE), 2003: Inventaire des Mesures en Faveur de I’Emploi, Juin.
D’Addio, A., De Greef, I. and M. Rosholm, 2002: Assessing Unemployment Traps in Belgium Using Panel Data Sample Selection Models, IZA Discussion Paper No. 669.
Daveri, F. and G. Tabellini, 2000: Unemployment, Growth and Taxation in Industrial Countries, Economic Policy, No. 30, pp. 49- 104.
De Koning, J., 1993: Measuring the Placement Effects of Two Placement Schemes for the Long-Term Unemployed, Empirical Economics, No. 18, pp. 447- 68.
De Lathouwer, L., 2003: Reforming the Passive Welfare State: Belgium’s New Income Arrangements to Make Work Pay in International Perspective, Paper for the Conference “Social Security and Participation in Economic and Social Life”, 10th International Research Seminar on Issues in Social Policy, FISS, Stiguna Sweden.
Deleeck, H., 2003: Securite Sociale et Distribution des Revenus en Belgie 1966-1999, mimeo.
Dolton, P., Makepeace, G. and J. Treble, 1992: Public and Private Sector Training of Young People in Britain, in Lynch, L. (ed.): Training and the Private Sector, pp. 261- 81, University of Chicago Press, Chicago, IL.
Dolton, P., Makepeace, G. and J. Treble, 1994a: The Young Training Scheme and the School-to-Work Transition, Oxford Economic Papers, Vol. 46, No. 4, pp. 629- 57.
Dolton, P., Makepeace, G. and J. Treble, 1994b: The Wage Effect of YTS: Evidence from YCS, Scottish Journal of Political Economy, Vol. 41, No. 4, pp. 444- 53.
Dolton, P. and D. O’Neill, 1996: The Long-Run Effect of Unemployment Monitoring and Work Search Programs: Some Evidence from the UK, University of Newcastle-upon-Tyne, mimeo.
Drèze, J. and F. Modigliani, 1981: The Tradeoff between Real Wages and Employment in an Open Economy (Belgium), European Economic Review, Vol. 15, pp. 1- 40.
Edin, P.-A., 1988: Individual Consequences of Plant Closure, PhD. Dissertation, University of Uppsala.
Elmeskov, J., Martin, J. and S. Scarpetta, 1996: Key Lessons for Labour Market Reforms: Evidence for OECD Countries’ Experiences, Swedish Economic Policy Review, Autumn, Vol. 5, No. 2, pp. 205- 58.
Estevao, M., 2002: Regional Labor Markets in Belgium, Belgium Selected Issues, IMF Country Report 02/43.
Estevao, M., 2003: Employment and Wage Effects of Active Labor Market Polices, France Selected Issues, SM/03/331.
Forslund, A. and A. Krueger, 1994: An Evaluation of the Swedish Active Labor Market Policy: New and Received Wisdom, in Freeman, R., Topel, R. and B. Swedenburg (eds): The Welfare State in Transition, The University of Chicago Press, Chicago, IL.
Gerfin, M. and M. Lechner, 2002: A Microeconometric Evaluation of the Active Labor Market Policy in Switzerland, Economic Journal, Vol. 112, October, pp. 854- 93.
Green, F., Hoskins, M. and S. Montgomery, 1996: The Effects of Company Training, Further Education, and the Youth Training Scheme on the Earnings of Young Employees, Oxford Bulletin of Economics and Statistics, Vol. 58, No. 3, pp. 469- 88.
Harkman, A., Jansson, F. and A. Tamas, 1996: Effects, Defects, and Prospects — An Evaluation of Labor Market Training in Sweden, Swedish National Labor Market Board, mimeo.
Heckman, J. J., Lalonde, R., and J. Smith, 1999: The Economics and Econometrics of Active Labor Market Programs, in: Ashenfelter, O. and D. Card (eds): Handbook of Labor Economics, Volume 3, pp. 1865- 2097, Elsevier Science, Amsterdam.
Heylen, F., 1993: Labour Market Structures, Labour Market Policy and Wage Formation in the OECD, Labour, Vol. 7, No. 2, pp. 25- 51.
Jackman, R., Layard, R. and S. Nickell, 1996: Combating Unemployment, Is Flexibility Enough? Discussion Paper No. 293, Centre for Economic Performance, LSE, London.
Jensen, P., Pedersen, P., Smith, N. and N. Westergard-Nielson, 1993: The Effects of Labor Market Training on Wages and Unemployment: Some Danish Results, in: Bunzel, H., Jensen, P. and N. Westergard-Nielson (eds): Panel Data and Labor Market Dynamics, Contributions to Economic Analysis, No. 222, pp. 227- 54, North Holland, Amsterdam.
Jenssen, P., 1999: The Danish Youth Unemployment Programme, Center for Labor Market and Social Research, Aarhus, mimeo.
Kluve, J. and C. Schmidt, 2002: Active Policy Evaluation, Economic Policy, October, pp. 410- 48.
Konings, J, and F. Roodhooft, 1997: How Elastic is the Demand for Labor in Belgian Enterprises? Results from Firm-Level Accounts Data, 1987-94, in De Economist, No. 145, pp. 229- 241.
Kraus, F., Puhani, P. and V. Steiner, 1997: Employment Effects of Publicly Finances Training Programs — The East German Experience, Zentrum fur Europaische Wirtschaftsforschung Discussion Paper No. 97- 33.
Krueger, A. and B. Meyer, 2002: Labor Supply Effects of Social Insurance, in: Auerbach, A. and M. Feldstein (eds): Handbook of Public Economics, Vol. 4, Chapter 33, pp. 2327- 92, Elsevier Science, Amsterdam.
Larsson, L., 2000: Evaluation of Swedish Youth Labour Market Programmes, Uppsala University, Dept of Economics Working Paper 2000- 6.
Laurent, S., 2002: Panel Data Analysis of Individual Wages, mimeo.
Layard, R., Nickell, S. and R. Jackman 1991: Unemployment: Macroeconomic Performance and the Labor Market, Oxford University Press, Oxford.
Lechner, M., 2001: Program Heterogeneity and Propensity Score Matching: Application to the Evaluation of Active Labor Market Policies, Review of Economics and Statistics, Vol. 84, No.2, pp. 205- 20.
Lopez-Novella, M., 2003: Effets de certain susides temporaires a I’embauche: une analyse microeconometrique des plans plus et du plan avantage a I’embauche, Federal Planning Bureau Working Paper, No. 16- 03, Brussels.
Lucas, R. E., 2003: Macroeconomic Priorities, American Economic Review, Vol. 93, No. 1 (March) pp. 1- 14.
Main, B., 1991: The Effects of the Youth Training Scheme on Employment Probability, Applied Economics, Vol. 23, No. 2, pp. 367- 72.
Main, B. and D. Raffe, 1983: Determinants of Employment and Unemployment among School Leavers: Evidence from the 1979 Survey of Scottish School Leavers, Scottish Journal of Political Economy, Vol. 30, No. 1, pp. 1- 17.
Main, B. and D. Raffe, and M. Shelly, 1990: The Effectiveness of the Youth Training Scheme as a Manpower Policy, Economics, Vol. 57, pp. 495- 514.
Moffitt, R., 2002: Welfare Programs and Labor Supply, in Auerbach, J. and M. Feldstein (eds.) Handbook of Public Economics, Vo. 4, Chapter 34, pp. 2393- 431, Elsevier Science, Amsterdam.
NEI, 1999: Benefit dependency ratios, An analysis of nine European, countries, Japan and the US, Rotterdam, Netherlands
Nickell, S. and R. Layard, 1999: Labor Market Institutions and Economic Performance, in: Ashenfelter, O. and D. Card (eds): Handbook of Labor Economics, Vol. 3. Elsevier Science, Amsterdam.
O’Connell, P. and F. McGinnity, 1997: What works, Who Works? The Employment and Earnings Effects of Active Labour Market Programmes Among Young People in Ireland, Work, Employment and Society, Vol. 11, No. 4, pp. 639- 61.
O’Higgins, N., 1994: YTS, Employment, and Sample Selection Bias, Oxford Economic Papers, Vol. 46, No. 4, pp. 605- 28.
OECD, 1993: Employment Outlook, Paris.
OECD, 2002: Benefits and Wages, OECD Indicators, Paris.
OECD, 2003: Economic Surveys: Belgium, Volume 2003/1, Paris.
Payne, J., Lissenburg, S. and M. White, 1996: Employment Training and Employment Action: An Evaluation by the Matched Comparison Method, Policy Studies Institute, London.
Prescott, E., 2002: Prosperity and Depression, American Economic Review, Papers and Proceedings, Vol. 92, No. 2, pp. 1- 15.
Raaum, O. and H. Torp, 2001: Labor Market Training in Norway - Effect on Earnings, Labor Economics.
Regnér, H. 2001: A Nonexperimental Evaluation of Training Programs for the Unemployed in Sweden, Labor Economics.
Ridder, G., 1986: An Event History Approach to the Evaluation of Training, Recruitment and Employment Programs, Journal of Applied Econometrics, Vol. 11, pp. 109- 126.
Rosholm, M., 1999: Evaluating Subsidized Employment Programmes in the Private and Public Sector, Center for Labour Market and Social Research, Aarhus, mimeo.
Scarpetta, S., 1996: Assessing the Role of Labour Market Policies and Institutional Settings on Unemployment: A Cross-Country Study, OECD Economic Studies, Vol. 26, No. 2, pp. 43- 98.
Sianesi, B., 2001: An Evaluation of the Active Labour Market Programmes in Sweden, University College London, mimeo.
Smets, J., 2003: Mobilisation generalepour I’emploi, Conference pour l’emploi 2003, Brussels, Conseil Superieur de TEmploi.
Thierry, P. and M. Sollogub, 1993: Les politiques francaises de l’emploi en faveur des jeunes - une evaluation econometrique, Revue Economique Vol. 46, No. 3, pp. 549- 59.
Torp, H., Raaum, O., Hereas, E. and H. Goldstein, 1993: The First Norwegian Experiment, in Jensen, K. and P.K. Madsen (eds): Measuring Labor Market Measures, Danish Ministry of Labor, Copenhagen, pp. 97- 140, Copenhagen, Denmark.
van den Berg, G.J. and B. van der Klaauw, 2000: Counseling and Monitoring of Unemployed Workers: Theory and Evidence from a Social Experiment, Free University of Amsterdam, mimeo.
van Ours, J.C., 2000: Do Active Labor Market Policies Help Unemployed Workers to Find and Keep Regular Jobs?, Tilburg University, CentEr Working Paper 00- 10.
Westergard-Nielsen, N., 1993: The Effects of Training: A Fixed-Effect Model, in Jensen, K. and P.K. Madsen (eds): Measuring Labor Market Measures, Danish Ministry of Labor, Copenhagen, pp. 167- 200, Copenhagen, Denmark.
White, M. and J. Lakey, 1992: The Restart Effect: Does Active Labor Market Policy Reduce Unemployment? Policy Studies Institute, London, mimeo.
Whitfield, K. and C. Bourlakis, 1991: An Empirical Analysis of YTS, Employment and Training, Journal of Economic Studies, Vol. 18, No. l, pp. 42- 56.
Zetterberg, J., 1995: Unemployment, Labour Market Policy and the Wage Bargaining System, in Johannesson, J (ed.): Unemployment, Labour Market Policy and the Wage Bargaining System, EFA, Stockholm, pp. 57- 115.
Zhou, J. 2003: Fiscal Strategies for Population Ageing, Belgium: Selected Issues, IMF Country Report N. 03/50.
Zweimueller, J. and R. Winter-Ebner, 1996: Manpower Training Programmes and Employment Stability, Economica, Vol. 63, pp 113- 30.
Prepared by Gerwin Bell.
In line with OECD convention (OECD 2002), social transfers are defined to exclude pensions (which were discussed in Zhou, 2003). Health related spending is taken up in Chapter 2.
Unless otherwise indicated, the data are based on OECD statistics through 1998, and staff estimates for 1999 and 2000 on the basis of preliminary data obtained from the OECD and Eurostat. The comparator aggregates are calculated as unweighted averages.
In practical terms, this is the result of individuals accumulating several benefits, rather than the entire population receiving a benefit.
The elevated levels of the benefit dependency ratio have not been affected by the significant fiscal adjustment during the 1990s. Thus if there was any reduction in the breadth of social transfers, it must have been offset by the decline in employment in full-time-equivalent terms.
Hence, income redistribution is not substantially affected by the transfer system, but rather through taxes, with the top quintile paying ⅔ of taxes, compared to only 4 percent for the bottom one (OECD 2003). See also section C.
This pattern also implies that in due course, as more and more benefits reach the maximum statutory level and the real value of the maximum is not adjusted, the replacement rate would start to decline.
For consistency across countries, OECD data on unemployment rates were used.
The tax increases in the early part of the 1990s are the major reason for the further equalization of the income distribution over the period. For the currently phased-in tax cuts, this finding suggests that if there is a desire to keep income redistribution at its current scope, a greater measure of progressivity would need to be introduced into the tax-benefit system (see also the discussion below).
This finding holds for various definitions of the tax wedge used by the OECD. Figure I.8 depicts the average tax wedge, including employer social security contributions.
A long-run labor demand elasticity of -1 ¾, estimated by Konings and Roodhooft (1997) corresponds to the estimate of-2 by Dreze and Modigliani (1981). Daveri and Tabellini (2000) whose cross-country sample includes Belgium also report substantial elasticities.
It should of course be noted that there are also other factors at play, for example institutional aspects, like minimum wages, the collective labor bargaining regime, or employment-protection legislation.
Some exceptions to the generally open-ended system were introduced during the 1990s. Most importantly, benefit duration for unemployed who are not considered to be the heads of households can be limited to 1.5 times the length of the average unemployment spell, taking account of the sex, age and residence of the unemployed. For example, the limit for women stretches from 2½ years (under 36 years, residing in Ostend) to 8 years (over 45 years, and residing in Mons) (De Lathouwer, 2003).
In 2000, long-term unemployment accounted for 56 percent of the total, compared with 32 percent in the OECD and 44 percent in the euro area.
For example, one third of all unemployed women have never worked.
Laurent (2001) estimates that wages in Wallonia are some 7 percent lower than in the rest of the country.
Other institutional factors aggravate the geographically differentiated employment effects. Most importantly, collective wage agreements have traditionally set sectoral minimum wages considerably in excess of the national minimum wage, thus serving to impede a low-wage job market even if reservation wages from unemployment benefits were not binding.
These observations are based on annual OECD data on participant’s inflows into ALMPs. To the extent that ALMPs in other countries last longer than in Belgium, these results would need to be nuanced.
A large part of such employment programs consist of workfare arrangements for welfare recipients. Cockx and Ridder (2001) show that the local welfare agencies that were running these schemes selected the most employable candidates from the welfare pool into workfare. Such workfare took care of certain local government services and was paid for by the employment agency. For the welfare participants the scheme was attractive because it resulted in higher unemployment benefits when the workfare program was left after one year. In total, the scheme was found to significantly lower the exit rate from unemployment.
In the first place, one would not expect any effect. However, in a collective bargaining model, where the union sets the minimum wage taking into account the cost of job loss (Layard and Nickel, 1991), an effect might arise: the newly trained may increase labor supply, and thus the expected duration of unemployment after a job loss, which in turn increases the expected cost of a job loss, and would in the end need to be resolved through lowered wages. Nevertheless, to the extent that the newly trained are not perceived as potential replacements, as is likely, given their lack of experience, this channel is probably not very important, in which case measured employment may actually increase (i.e., by the newly trained who would not find a job).
References
Arrow, Kenneth J., 1963, Uncertainty and the Welfare Economics of Medical Care, American Economic Review, Vol. 53, No. 5, pp. 941- 973.
Culyer, Anthony J., 1988, Health Expenditures in Canada: Myth and Reality; Past and Future, Canadian Tax Foundation: Paper No. 82
Cuttler, David M., 2002, Health Care and the Public Sector, in A. Auerbach and M. Feldstein (eds) Handbook of Public Economics, Amsterdam: Elsevier.
Darby, M.R. and E. Kami, 1973, Free Competition and the Optimal Amount of Fraud, Journal of Law and Economics, Vol. 16, pp. 67- 88
Docteur, Elizabeth and Howard Oxley, 2003, Health-Care Systems: Lessonsfrom the Reform Experience, OECD Health Working Paper No. 9.
Federaal Planbureau, 2002, Verkenning van de Financiele Evolutie van de Sociale Zekerheid 2000-2050, Occasional Paper, Brussels
Gerdtham, Ulf-G. and Bengt Jonsson, 2000, International Comparisons of Health Expenditure, in A.J. Culyer and J.P. Newhouse (eds) Handbook of Health Economics, Amsterdam: Elsevier.
Grossman, Michael, 1972, The Demand for Health: A Theoretical and Empirical Investigation, New York: NBER.
Grossman, Michael, 1999, The Human Capital Model of the Demand for Health, NBER Working Paper No. 7078.
Koen, Vincent, 2000, Public Expenditure Reform: the Health Care Sector in the United Kingdom, OECD Economics Department Working Paper No. 256.
Newhouse, J.P., 1977, Medical Care Expenditure: A Cross National Survey, Journal of Human Resources, Vol. 6, No. 3 (summer).
OECD, 1992, Economic Survey of the United States, Paris
OECD, 1999, Economic Survey of Belgium, Paris
OECD, 2003 a, Assessing the Performance of Health Care Systems: A Framework for OECD Surveys, Mimeo, Paris.
OECD, 2003b, Spending on Health and Long Term Care: Projections to 2050 Revisited, Mimeo, Paris.
Tirole, Jean, 1988, The Theory of Industrial Organization, Cambridge (MA): MIT Press.
Weisbrod, Burton A., 1991, The Health Care Quadrilemma: An Essay on Technological Change, Insurance, Quality of Care, and Cost Containment, Journal of Economic Literature, Vol. 29, No. 2, pp. 523- 552.
Prepared by David Hofman.
Forty percent of hospitals are public, with the remainder being mostly non-profit private institutions.
Under the public insurance scheme, coverage for the self-employed is less extensive than for workers. The former group is only insured against “major risks”
Lower governments’ competence in health care issues is limited to health education, preventive care and implementation of certain federal decisions.
Reimbursement levels as well as maximum co-payments are set in terms of official reference prices, which still leaves patients exposed to the possible charging of premiums by physicians. The vast majority of physicians, though, adhere to the reference prices (see ¶ 35).
The Netherlands is currently planning a reform that will end the dichotomy between public and private schemes, and will introduce a single uniform and mandatory insurance package for the entire population. To some extent, the reform entails a switch towards the universal public insurance systems prevalent in other European countries, although it is expected that the new system will be managed by private, commercial insurance companies.
The United States government’s tax policies and regulations still have a considerable impact on health care market outcomes.
Health care is a so called “credence good”, for which the quality—and indeed often the very necessity of purchase—is unobservable to the consumer, even after the product has been consumed (Tirole, 1988, and Darby and Kami, 1973).
The age group subdivision here—which differs somewhat from usual definitions of working age and elderly populations—-has been chosen such as to be consistent with the 2002 FPB forecasting exercise.
Although important, moral hazard is not the only phenomenon hampering the proper functioning of the health care market. Other distortions arise from various information problems, externalities, and adverse selection. See Cuttler (2002) for an overview of sources of market failure and rationales for government intervention.
