Forbearance, Interest Rates, and Present-Value Effects in a Randomized Debt Relief Experiment

I design a debt relief experiment that randomizes forbearance term and interest rates independently in a 2-by-2-by-2 design for delinquent borrowers. Forbearance take-up prevents one in three defaults in the first month with no long-run effects beyond expiration. Before expiration forbearance reduces payments by twice as much compared with rate reductions but reduces defaults by less. Using the experimental assignment as an instrument to decompose the effect of future payments entailed by a rate reduction from current liquidity a dollar change in the present value of future payments has a similar effect on defaults as a 30 cent change in current payments. Whether forbearance or rate reductions are relatively more effective the relative sensitivity of behavior to future payments and whether interest rates affect behavior to the extent that they affect current payments are all tightly linked to balance sheets. Influencing interest rates has benefits through the present value of future payments which is difficult to replicate using rescheduling policies