These loan provider procedures are excluded while they offer no improvement in the likelihood of therapy during the boundary. The treatment effect is estimated as the jump in the outcome at the boundary divided by the jump in the probability of treatment at the boundary in the fuzzy RD. For those loan provider procedures, the latter is undefined; thus these examples are excluded from subsequent analysis. First-stage discontinuity plots are shown for time horizons of 10 times, 1 month, and 24 months in on line Appendix Figure A1. These quotes aren’t responsive to variation within the estimation bandwidth, illustrated in Online Appendix Figure A2.