Gender-targeted transfers by default? - Evidence from a child allowance reform in Sweden

We exploit a sharp birthday discontinuity in a large and universal Swedish cash transfer program, creating plausibly exogenous variation in the default disbursement option, while holding entitlements and other financial incentives constant. When the cash transfer is paid out to the mother by default, instead of a 50/50 default, it has a large effect (55 percentage points) on the probability that the transfer is deposited in the mother's bank account also in the long run. Surprisingly, we find that the default policy redistributes resources to separated low-income mothers. We find no indications that the 100%-to-the-mother default induces mothers to work less or to take more responsibility for the children.