How effective are hiring subsidies in reducing long-term unemployment among prime-aged jobseekers? Evidence from Belgium

Sam Desiere; Bart Cockx

doi:10.2478/izajolp-2022-0003

Figures & Tables

The evolution of long-term unemployment in Flanders.Source: Monthly average from Arvastat (https://arvastat.vdab.be/, accessed April 2021).

Employment stability: share of individuals aged 45–48 years hired in Month 5 (not eligible for the subsidy) vs. in Month 6 (eligible for the subsidy) employed for 1–65 consecutive months after being hired.Notes: For clarity, only the 95% confidence interval for individuals hired in Month 6 are shown. The main message is that the employment rate of individuals hired in Month 5 (not eligible for the subsidy) is slightly higher than the employment rate of individuals hired in Month 6 (eligible for the subsidy), but the difference between both groups is not statistically significant. The CIs are constructed by regressing the employment rate on the number of months after registration (dummies), dummies indicating the group (jobseekers hired in Month 5 vs. hired in Month 6), and the interaction terms between these dummies.

Exploring the effect of the abolition of the subsidy in a DiD framework: the job-finding rate of jobseekers aged 45–48 years as a function of elapsed unemployment duration, before and after the reform.Notes: The vertical dashed line indicates the month in which jobseekers become eligible for the hiring subsidy prior to the reform. The pre-reform period consists of jobseekers who registered at the PES in 2015. The post-reform period consists of jobseekers who registered in 2017. DiD, difference-in-differences; PES, public employment service.

The effect of abolition of the subsidy on the probability of finding employment lasting 1–12 (consecutive) months.Notes: Each graph shows the triple-interaction term with its 95% confidence interval obtained by estimating 12 triple-difference regressions following the benchmark specification as in Table 2. The outcome is an indicator taking the value of one if the jobseeker found a job and remained employed for 1–12 (consecutive) months, and zero otherwise. Full results of the triple-difference regressions are reported in Table A4 in Appendix A.

The evolution of long-term unemployment in Flanders among jobseekers aged 40–48 years.Source: Monthly average from Arvastat (https://arvastat.vdab.be/, accessed April 2021).

The job-finding rate of jobseekers aged 40–43 years as a function of the elapsed unemployment duration, before and after the reform.Notes: The vertical dashed line indicates the month in which jobseekers aged 45–48 years become eligible for the hiring subsidy prior to the reform.

The effect of the abolition of the subsidy on the probability of finding employment lasting 1–12 (consecutive) months using the more stringent definition of employment.Notes: Each graph shows the triple-interaction term with its 95% CI obtained by estimating 12 triple-difference regressions following the benchmark specification as in Table 2. The outcome is an indicator taking the value of one if the jobseeker found a job and remained employed for 1–12 (consecutive) months, and zero otherwise. The definition of employment follows a more stringent definition of work, thereby excluding temporary work lasting between 10 days and 20 days per month.

The impact of abolition of the hiring subsidy on the job-finding rate: full results of the benchmark specification

	Benchmark
Unemployment duration (base level: Month 1)
Month 2	−0.0407^*** (0.00183)
Month 3	−0.0841^*** (0.00181)
Month 4	−0.0998^*** (0.00183)
Month 5	−0.109^*** (0.00185)
Month 6	−0.111^*** (0.00213)
Month 7	−0.122^*** (0.00213)
Month 8	−0.129^*** (0.00213)
Month 9	−0.132^*** (0.00214)
Month 10	−0.141^*** (0.00212)
Month 11	−0.141^*** (0.00215)
Month 12	−0.143^*** (0.00217)
Post-reform	0.0250^*** (0.00275)
Age 45–48 years	−0.00487^*** (0.00144)
Age 45–48 years ## post-reform	0.0101^*** (0.00261)
Post-reform ## long-term unemployed	−0.0231^*** (0.00247)
Post-reform ## age 45–48 years ## long-term unemployed	−0.0118^*** (0.00358)
Year of registration at the PES (base level: 2012)
2013	−0.00498^*** (0.00147)
2014	0.00309^** (0.00154)
2015	0.0131^*** (0.00159)
2016	0.0115^*** (0.00205)
2017	0.0162^*** (0.00283)
2018	0.0205^*** (0.00298)
Education (base level: high level)
Low	−0.00236^** (0.00116)
Medium	0.00423^*** (0.00115)
Man	0.00759^*** (0.000845)
Proficiency in Dutch (base level: no proficiency)
Limited proficiency	−0.00967^*** (0.00332)
Good proficiency	0.00455 (0.00325)
Excellent proficiency	0.00491 (0.00329)
Origin (base level: EU-15)
EU-13	−0.0196^*** (0.00261)
Other	−0.0250^*** (0.00110)
Constant	0.197^*** (0.00387)
No. of unemployment spells	104,437
No. of observations	602,916
R²	0.032

Placebo tests for the benchmark analysis

	Parallel trend	Placebo test for the short-term unemployed
Age 45–48 years ## long-term unemployed ##
2012	0.00494 (0.00613)
2013	0.00441 (0.00534)
2014	−0.000794 (0.00559)
2016 Q1–Q2	0.00574 (0.00762)
2016 Q3–Q4	−0.00516 (0.00715)
2017	−*0.0101^ (0.00616)**
2018	−*0.0137^ (0.00721)**
Age 45–48 years ## after reform ## unemployed for (base level: 5 months)
Less than 5 months		0.00134 (0.00668)
6–12 months		−0.0112^* (0.00650)

No. of unemployment spells	104,437	104,437
No. of observations	602,916	602,916
R²	0.032	602,916

Pre-reform parallel trend (F-test, p-value)	0.74
Post-reform parallel trend (F-test, p-value)	0.59
No effect on the short-term unemployed (p-value)		0.84

The effect of the subsidy for different populations

	Job-finding rate	Job stability: remaining employed for 6 consecutive months
Benchmark specification	−0.0118^*** (0.00358)	−0.00408 (0.00286)
Population: all jobseekers	−0.0101^*** (0.00269)	−0.00521^** (0.00209)
Population: all jobseekers who worked 4 months prior to registration + jobseekers whose unemployment history is not known	−0.0110^*** (0.00330)	−0.00417 (0.00265)

The effect of the abolition of the hiring subsidy on finding a job and remaining employed for 1–12 consecutive months (Figure 4, left panel in the main text plots the results)

	Employed for at least:
	1 month	2 months	3 months	4 months	5 months	6 months	7 months	8 months	9 months	10 months	11 months	12 months
Effect	−0.0118^*** (0.00358)	−0.00934^*** (0.00337)	−0.00786^** (0.00322)	−0.00445 (0.00308)	−0.00459 (0.00297)	−0.00408 (0.00286)	−0.00361 (0.00273)	−0.00383 (0.00264)	−0.00347 (0.00256)	−0.00191 (0.00248)	−0.000313 (0.00238)	−0.000546 (0.00230)
No. of unemployment spells	104,437	104,437	104,158	103,949	103,699	103,444	103,157	102,837	102,726	102,601	102,430	102,280
No. of observations	602,916	602,323	601,458	600,846	600,165	599,403	598,250	597,607	597,227	596,763	596,209	595,688
R²	0.032	0.028	0.026	0.024	0.022	0.021	0.021	0.022	0.022	0.022	0.021	0.020
Post-reform placebo test (F-test, p-value)	0.58	0.73	0.81	0.62	0.84	0.88	0.64	0.96	0.93	0.78	0.32	0.37
Effect on the short-term unemployed (p-value)	0.91	0.85	0.75	0.94	0.94	0.77	0.89	0.87	0.73	0.50	0.44	0.39

The Breusch–Godfrey LM test on grouped errors

	Weighted OLS
Lagged grouped error	0.0526^*** (0.0109)

Number of observations	9,742
R²	0.021

DiD exploiting differences across unemployment duration among older jobseekers and placebo tests

	DiD: short-term vs. long-term unemployed, among jobseekers aged 45–48 years	Placebo test 1: post-reform parallel trends	Placebo test 2: jobseekers aged 40–43 years
Post-reform ## long-term unemployed	−0.0352^*** (0.00264)		−0.0230^*** (0.00248)
Long-term unemployed ## year (base level: 2015)
2012		0.00626 (0.00444)
2013		0.0162^*** (0.00384)
2014		0.00572 (0.00403)
2016 Q1–Q2		−0.000136 (0.00548)
2016 Q3–Q4		−0.0194^*** (0.00523)
2017		−0.0283^*** (0.00448)
2018		−0.0377^*** (0.00522)

No. of observations	283,469	283,469	319,447
R²	0.034	0.034	0.031
Post-reform parallel trend (F-test, p-value)		0.012^**

The impact of the abolition of the hiring subsidy on the job-finding rate: triple-difference estimates

	Benchmark	Ruling out:			More stringent definition of “work” (excluding short-lived temporary work)

		Substitution between age groups	Substitution between the short-term and long-term unemployed	Anticipation effects
Age 45–48 years ## post-reform ## long-term unemployed (base level: 40–43 years)
	−0.0118^*** (0.00358)		−0.0119^*** (0.00377)	−0.0133^*** (0.00408)	−0.00636^** (0.00273)
Post-reform ## long-term unemployed (base level: 35–37 years)
## Age 38 years		0.00779^** (0.00389)
## Age 39 years		0.000466 (0.00394)
## Age 40 years		0.00170 (0.00399)
## Age 41 years		−0.00158 (0.00404)
## Age 42 years		0.00222 (0.00413)
## Age 43 years		−0.000159 (0.00410)
## Age 44–44.5 years		0.000781 (0.00558)
Age 45–48 years ## post-reform ## unemployed for 5 months
			−0.000717 (0.00667)

No. of unemployment spells	104,437	281,701	104,437	89,291	104,437
No. of observations	602,916	1,686,890	602,916	516,196	753,119
R²	0.032	0.033	0.032	0.032	0.013

Placebo tests
Post-reform parallel trend test (F-test, p-value)	0.59	0.93	0.75	0.63	0.20
Effect on the short-term unemployed (p-value)	0.84	–	–	0.86	0.91

The effect of the subsidy by gender and level of education

	Job-finding rate		Job stability

	Benchmark specification	More stringent definition of “work” (excluding short-lived temporary work)	Remaining employed for 6 consecutive months
Sex
Men	−0.0076 (0.0118)	−0.0048 (0.0039)	−0.0044 (0.0039)
Women	−0.0168^*** (0.0052)	−0.0083^** (0.0038)	−0.0041 (0.0042)
Education
Less than a bachelor's degree	−0.0074^*, ^a (0.004)	−0.0051^* (0.003)	−0.0005^a (0.0032)
At least a bachelor's degree	−0.0264^***, ^a (0.0077)	−0.0113^* (0.0062)	−0.0163^**, ^a (0.0064)

DiD exploiting differences across age groups for the long-term unemployed and placebo tests

	DiD: young vs. older long-term unemployed jobseekers	Placebo test 1: post-reform parallel trends	Placebo test 2: short-term unemployed
Post-reform ## age 45–48 years	−0.00138 (0.00245)		0.0101^*** (0.00261)
Age 45–48 years ## year (base level: 2015)
2012		0.00769^* (0.00394)
2013		0.00220 (0.00346)
2014		0.00463 (0.00363)
2016 Q1–Q2		0.00444 (0.00495)
2016 Q3–Q4		0.00652 (0.00473)
2017		0.00102 (0.00407)
2018		−0.00195 (0.00538)

No. of observations	241,611	241,611	361,305
R²	0.005	0.005	0.020
Post-reform parallel trend (F-test, p-value)		0.34

Studies on the effectiveness of private employment programs for the long-term unemployed included in the review by Card et al_ (2018) (studies not included in the bibliography)

Title	Authors	Year of publication	Journal	Country
Active labor market policies in Denmark: A comparative analysis of postprogram effects	Blanche	2011	U	Denmark
Evaluating the impact of a well-targeted wage subsidy using administrative data	Cseres-Gergely, Scharle, Földessy	2012	U	Hungary
The effectiveness of targeted wage subsidies for hard-to-place workers	Jaenichen, Stephan	2011	Applied Economics	Germany
Evaluation of subsidized employment programs for long-term unemployed in Bulgaria	Mihaylov	2011	Economic Studies	Bulgaria
Do long-term unemployed workers benefit from targeted wage subsidies?	Schünemann, Lechner, Wunsch	2013	German Econ Rev	Germany
Employment subsidies: A fast lane from unemployment to work?	Forslund, Johansson, Lindquist	2004	U	Sweden
Income support policies for part-time workers: A stepping-stone to regular jobs?	Cockx, Robin, Goebel	2010	U	Belgium
An application to young long-term unemployed women in Belgium
Dynamic treatment assignment: The consequences for evaluations using observational data	Fredriksson, Johansson	2008	U	Sweden
Do active labor market policies help unemployed workers to find and keep regular jobs?	Van Ours	2000	Book	Slovakia
The locking-in effect of subsidized jobs	Van Ours	2004	Journal of Comparative Economics	Slovakia

Characteristics of jobseekers by age group and unemployment duration: Age group; Unemployment duration, before and after the reform

	Jobseekers aged 40–43 years				Jobseekers aged 45–48 years					Triple difference²

	Short-term unemployed		Long-term unemployed		Short-term unemployed		Long-term unemployed

	Before	After	Before	After	Before	After	Before	After	Hired with a subsidy¹
Age, years	41.48	41.44	41.49	41.44	46.47	46.47	46.48	46.46	47.13	0.77
Men, %	54.12	54.77	54.76	54.10	53.80	54.84	54.83	54.13	56.84	0.43

Education, %

Low	38.74	37.66	40.99	39.50	42.73	39.37	44.15	41.56	43.33	−1.18
Medium	39.69	39.01	38.72	37.59	37.47	38.53	36.41	37.10	43.73	−0.08
High	21.57	23.34	20.30	22.91	19.80	22.11	19.44	21.34	12.94	1.25

Proficiency in Dutch, %

No	1.26	2.80	1.35	2.83	1.07	2.58	1.20	2.71	2.22	−0.05
Limited	11.22	14.17	13.25	16.06	10.06	13.08	11.79	15.48	9.61	−0.81
Good	37.40	32.47	37.70	32.32	39.04	35.13	39.50	35.32	37.60	−0.18
Excellent	50.13	50.57	47.70	48.79	49.83	49.21	47.51	46.50	50.57	1.04

Origin, %

EU-15	74.46	68.08	70.33	63.86	80.16	73.80	76.42	70.15	77.97	−0.17
EU-13	2.16	3.46	2.44	3.58	1.76	2.97	1.95	3.35	3.11	−0.34
Other	23.38	28.47	27.24	32.56	18.08	23.23	21.63	26.50	18.92	0.52

No. of individuals	34,318	21,127	18,259	8,332	30,405	18,587	16,381	7,192	1,221	–

The five most important sectors in terms of new hires with a subsidy in each quarter (2015Q2–2016Q4)

Sector	Share of new hires with a subsidy, 2015Q2–2016Q4
Temporary employment agency activities	61%
Cleaning activities	4%
Restaurants and mobile food service activities	2%
Retail sale in nonspecialized stores	2%
Other social work activities without accommodation	2%

The effect of the subsidy on the job-finding rate: robustness checks

	Censor spells when jobseekers leave the labor market	Treatment group includes jobseekers aged 44.5–48 years	Sensitivity to classification of jobseekers into pre-reform and post-reform periods			Effect by month

			Using July 1, 2016, as cutoff date	Excluding jobseekers registering in June 2016	Excluding jobseekers registering in the period April–June 2016
Effect	−0.0125^*** (0.00394)	−0.0113^*** (0.00347)	−0.0113^*** (0.00363)	−0.0116^*** (0.00364)	−0.0115^*** (0.00365)
Age 45–48 years ## post-reform
## Unemployed for 6 months						−0.0168^** (0.00664)
## Unemployed for 7 months						−0.00731 (0.00678)
## Unemployed for 8 months						−0.00979 (0.00684)
## Unemployed for 9–12 months						−0.0262^*** (0.00292)

No. of unemployment spells	104,437	110,930	104,437	103,322	101,296	104,437
No. of observations	545,580	640,138	602,916	596,171	583,026	602,916
R²	0.028	0.032	0.032	0.032	0.033	0.032

How effective are hiring subsidies in reducing long-term unemployment among prime-aged jobseekers? Evidence from Belgium

Figures & Tables

Figure 1

Figure 2

Figure 3

Figure 4

Figure A1

Figure A2

Figure A3

The impact of abolition of the hiring subsidy on the job-finding rate: full results of the benchmark specification

Placebo tests for the benchmark analysis

The effect of the subsidy for different populations

The effect of the abolition of the hiring subsidy on finding a job and remaining employed for 1–12 consecutive months (Figure 4, left panel in the main text plots the results)

The Breusch–Godfrey LM test on grouped errors

DiD exploiting differences across unemployment duration among older jobseekers and placebo tests

The impact of the abolition of the hiring subsidy on the job-finding rate: triple-difference estimates

The effect of the subsidy by gender and level of education

DiD exploiting differences across age groups for the long-term unemployed and placebo tests

Studies on the effectiveness of private employment programs for the long-term unemployed included in the review by Card et al_ (2018) (studies not included in the bibliography)

Characteristics of jobseekers by age group and unemployment duration: Age group; Unemployment duration, before and after the reform

The five most important sectors in terms of new hires with a subsidy in each quarter (2015Q2–2016Q4)

The effect of the subsidy on the job-finding rate: robustness checks

Paradigm

My account