This paper investigates the impact of survivors insurance on marital behavior. We study the 1996 Dutch reform which considerably tightened eligibility rules to survivors' benefits. Exploiting a discontinuity in date of birth eligibility to survivors insurance and using a rich and exhaustive of the Dutch population administrative dataset, we carry out a regression discontinuity design and we find no evidence of the reform on divorce probability. Exploring possible explanations for our zero-effect result, we study how labor supply responses can compensate the income drop the reform induced. We find a strong increase in the labor force participation of widows after the reform. However this response does not completely offset the decrease in income generated form the cut in survivors benefits.