The added worker effect: Evidence from adisability insurance reform

Mario Bernasconi, Tunga Kantarci, Arthur van Soest, Jan-Maarten van Sonsbeek

Research output: Contribution to journalArticleScientificpeer-review

2 Citations (Scopus)

Abstract

The Netherlands reformed its disability insurance (DI) scheme in 2006. Reintegration incentives for employers became stronger, access to DI benefits became more difficult, and benefits became less generous. Using administrative data on all individuals who fell sick shortly before and after the reform, we study the impact of the reform on labor participation of individuals who fell sick and their spouses. Difference-in-differences estimates show, among other things, that the reform led to an increase of labor participation of the individuals who fell sick only if these individuals had a permanent job, whereas spouses responded to the DI reform in other cases, where the individuals reporting sick had a temporary job or were unemployed. More generally, the spouses respond when the sick individual’s labor market position is weak and the individual him- or herself has trouble finding or retaining employment. The effects are persistent during the 10 years after the reform. The effect on the spouse can be seen as an “added worker effect,” where additional earnings of the spouse compensate for the sick individual’s income loss so that both partners share the burden of a more stringent DI scheme. Comparing individuals reporting sick with and without partner provides further support for the notion that the responses of couples to the reform are joint decisions of the two partners.
Original languageEnglish
JournalReview of Economics of the Household
DOIs
Publication statusE-pub ahead of print - Jan 2024

Keywords

  • added worker effect
  • disability insurance
  • program evaluation
  • household behaviour
  • differences-in-differences

Fingerprint

Dive into the research topics of 'The added worker effect: Evidence from adisability insurance reform'. Together they form a unique fingerprint.

Cite this