Employment effects of the Danish rehabilitation benefit

Aims: Social benefits aim to bring marginalised citizens back into the labour force. As benefits constitute a burden for tax payers, attention has been given to measure the effect. We used register data to assess the employment effect of rehabilitation benefit; the most liberal social benefit in Denmark. Methods: We used data on rehabilitation benefits paid during 1994–2004. In the study design, special care was given to overcome selection bias. We identified municipalities with most frequent (generous) and least frequent (stingy) use of rehabilitation benefit, controlling for confounders (step 1). From the generous municipalities, population subgroups where ≥10% received rehabilitation benefit (vulnerable group) were identified based on age, sex, education, cohabitation, children, and health (step 2). Outcome measure was employment rate in 2005 for vulnerable groups in generous municipalities compared with vulnerable groups in stingy municipalities, controlled for municipality employment rate, i.e. a ratio of ratios (step 3). Results: Use of rehabilitation benefit varied 10-fold across the 271 Danish municipalities. The employment rate in 2005 for vulnerable groups was lower in generous than in stingy municipalities. The ratio of ratios was 0.915 (95% CI 0.888–0.945) for men and 0.919 (95% CI 0.896–0.942) for women. No positive employment effect of generous use of rehabilitation benefit was found. Although residual confounding cannot be excluded, it is nevertheless remarkable that the possibly positive effect of rehabilitation benefit was not sufficiently large to become visible, even in a study particularly designed to eliminate selection biases and to control for confounding.