Improved Targeting of Social Programs: An Application to a State Job Coaching Program for Adults with Intellectual Disabilities

In a climate of flat or shrinking budgets, can programs reallocate existing resources to improve efficiency? We illustrate the potential for gains from redirecting resources using data from a state job coaching program that is designed to increase employment among adults with intellectual disabilities (ID). We model selection into the program and employment outcomes for participants and non-participants allowing for potentially heterogeneous response among observationally equivalent individuals. In our simulations, we find that state ID population employment can be increased from 10.7 percent to an upper bound of 16.7 percent by a program administrator who can allocate the job coaches to those with the most to gain. This is a 56 percent increase in the overall employment rate. While we assume that program administrators know more about individual program participants than we do, we can consider an administrator who has only the information available to the econometrician. In this case, targeting gains based only on observable characteristics would lead to 11.8 percent employment, which is an 11 percent increase in the overall employment rate. Surprisingly, a simple rule that only requires administrators to predict employment success when treated (based on observables) will achieve almost the same results.