Introduction From January 1st 2006 onwards, an early retirement reform was implemented in the Netherlands, in which workers born in 1950 or later were no longer fiscally rewarded to retire early. This reform did not apply to those born in 1949 or earlier.

Objectives This study investigated the effects of a national early retirement reform, which was implemented in 2006 and penalized early retirement, on paid employment and different exit pathways and examined whether these effects differ by gender, income level and health status.

Methods This study included all Dutch individuals in paid employment born six months before (control group) and six months after (intervention group) the cut-off date of the reform (1 January 1950) that fiscally penalized early retirement. A regression discontinuity design combined with restricted mean survival time analysis was applied to evaluate the effect of penalizing early retirement on labor force participation from age 60 until workers reached the retirement age of 65 years. This means that the effects between the intervention and control group were compared, while accounting for secular trends.

Results The intervention group postponed early retirement by 7.41 months (95% confidence interval (CI) 6.11–8.72), and partly replaced this by remaining 4.87 months (95% CI 3.60–6.24) longer in paid employment. Workers born after the threshold, annually earning €25 000–40 000, spent 1.24 months (95% CI 0.31–2.18) more in economic inactivity than those born before. The working months lost to unemployment increased by 1.50 months (95% CI 0.30–2.71) for female workers and 1.99 months (95% CI 0.06–3.92) for workers reporting multiple chronic diseases.

Discussion The national reform successfully prolonged working lives of older workers. However, workers with a middle income, female workers, and workers with chronic diseases were more vulnerable to premature exit from the labor market through unemployment or being without any income or benefit.