Looking behind mortgage delinquencies

Delinquency rates for mortgages originated before and after the financial crisis are examined using a novel and large panel obtained by merging data from tax records and credit registers. First, we estimate the selection into the mortgage market using an exogenous index of local credit supply as exclusion restriction. Second, controlling for selection we estimate the impact of income shocks on the delinquency rate. We find that since 2008 the selection process has led to the halving of the delinquency rate. Conditional on the creation of a new mortgage, job losses nearly double the delinquency risk; estimates uncorrected for selection are severely downward biased.