Home Work: Exploring the Labor Market Effects of Subsidizing Domestic Services (with Ilan Tojerow), Labour Economics, 2024, vol. 90.
This paper analyzes subsidies for the domestic services sector, an increasingly popular policy to create employment opportunities for low-skilled workers. Using Belgian administrative data, a differences-in-differences approach, and a shift-share instrumental variable, we estimate the local effects of the policy in targeted industries as well as overall effects on the labor market. We find that domestic service subsidies can increase female employment in the subsidized industries as well as the overall employment rate. This increase in employment is primarily driven by an increase in (formal) labor market participation and, to a lesser extent, a reduction in the rate of participation in unemployment insurance and in other social welfare programs. We also find that these subsidies can lead to an increase in the rate of work incapacity, likely due to the fact they broaden the population that can access the social safety net.
[Featured in La Libre and this short video]
Previously circulated as IZA Discussion Paper No. 13544, entitled "Subsidizing Domestic Services as a Tool to Fight Unemployment: Effectiveness and Hidden Costs"
The Consequences of Job Search Monitoring for the Long-Term Unemployed: Disability Instead of Employment? (with Octave De Brouwer and Ilan Tojerow), Journal of Public Economics, 2023, volume 224.
We study the effect of job search monitoring (JSM) on individual labor market outcomes of the long-term unemployed. Exploiting the implementation of a JSM program targeted at jobseekers under the age of 49, we set up a regression discontinuity design that credibly identifies the program’s causal effect on unemployment, employment, and disability insurance (DI) participation and participation in other social welfare programs within a three-year period. We find that JSM increases exits from unemployment to DI without affecting transitions into employment or other social welfare programs. We further find that the effect of JSM on DI materializes before any sanction can be imposed and monitored individuals are still 10 percentage points more likely to be on DI three years after the start of monitoring. Ultimately, exploring fiscal implications reveals that the decrease in unemployment transfers as a result of JSM is entirely offset by the increase in DI transfers.
Previously circulated as IZA Discussion Paper No. 12304, entitled "The Unexpected Consequences of Job Search Monitoring: Disability Instead of Employment?"
In a field experiment involving 100,000 recently unemployed individuals, we investigate how jobseekers' training decisions react to information about labor shortages. The intervention increased training participation by 6%, driven by an increased take-up of transversal programs (+10%) rather than specialized occupational training. Further analyses show that the intervention also reorients job search toward shortage occupations. At the same time, the increase in transversal training is not disproportionately concentrated among jobseekers who also shift their search toward shortage occupations. These findings suggest that jobseekers respond to information on labor shortages both by investing in relatively broad skills and by redirecting their occupational search toward jobs in higher demand.
Previously circulated as IZA Disucssion Paper No. 17731, entitled "Closing the Mismatch: Encouraging Jobseekers to Reskill for Shortage Occupations"
"Staying to Opportunity: The Effects of Exposure to Better Neighborhoods on Residents", with Anne Gielen, Olivier Marie and Tommaso Tulkens
We exploit a Dutch policy that exogenously changed the socioeconomic composition in targeted disadvantaged neighborhoods, in order to estimate the effects of neighborhoods on adults' and children's outcomes. We use administrative data and compare the outcomes of individuals living in targeted neighborhoods to those in observably similar control neighborhoods. We find that the policy improved the socioeconomic composition of treated neighborhoods: over the six years following implementation, the share of residents employed rose by 2.8 pp, the share receiving welfare benefits fell by 3.3 pp, and criminal activity fell by 0.5 pp. These socioeconomic changes in neighborhood composition had little impact on adults, but children who were exposed before the age of 12 were more likely to obtain a tertiary education degree. These findings confirm previous evidence from mover designs that neighborhood effects are limited for adults, but matter for children's long-term opportunities.
"Unemployed Jobseekers and the Decision to Re-skill: Evidence from a Discrete Choice Experiment", with Benjamine Dejardin, Alexia Delfino, Raffaella Sadun, and Ilan Tojerow
Rapidly evolving labor markets have led to skill mismatches and a growing need for workforce reskilling. This study explores the drivers of jobseekers' reskilling decisions through a discrete choice experiment conducted with a Public Employment Service. We find that the unemployed hold strong occupational preferences in the context of training decisions, but that these preferences can be shifted under some conditions. Training program attributes, such as shorter waiting times and commuting distances, the teaching format, and involving employers in the organization of the training program, can increase willingness to reskill. Working conditions in the target occupation such as salary, career evolution prospects, work demands and schedule flexibility, also matter for reskilling decisions. We discuss implications for policy-makers and employers struggling to find adequately skilled workers.
"Finding the Perfect Fit: Supporting Unemployed Jobseekers in their Choice of Occupational Training", with Benjamine Dejardin and Ilan Tojerow
We study whether persistently low training take up can be explained by the difficulty of choosing the right program. We develop a tool aimed at helping prospective training participants to identify the occupational trainings that best fit their professional and educational profile. The tool is based on the same philosophy as the transition matrix used in Belot et al. (2019, 2022 and 2025), but provides training recommendations (based on the education, past professional experience, or occupational preferences) rather than occupational job search recommendations. We partner with a Belgian Public Employment Service and its partner Public Training Provider to evaluate the effects of the tool in multiple RCTs. The tool is currently being tested in the field.