Microcredit – joint-liability loans to the poorest of the poor – has been touted as a powerful approach for combatting global poverty, but sustainability varies dramatically across banks. Efforts to improve the sustainability of microcredit have assumed defaults are caused by free-riding. Here, we point out that the response of other group members to delinquent groupmates also plays an important role in defaults. Even in the absence of any free-rider problem, some people will be unable to make their payments due to bad luck. It is other group members’ unwillingness to pitch in extra – due to, among other things, not wanting to have less than other group members – that leads to default. To support this argument, we utilize the Ultimatum Game (UG), a standard paradigm from behavioral economics for measuring one's aversion to inequitable outcomes. First, we show that country-level variation in microloan default rates is strongly correlated (overall r = 0.81) with country-level UG rejection rates, but not free-riding measures. We then introduce a laboratory model ‘Microloan Game’ and present evidence that defaults arise from inequity-averse individuals refusing to make up the difference when others fail to pay their fair share. This perspective suggests a suite of new approaches for combatting defaults that leverage findings on reducing UG rejections.