Dynamic Targeting: Experimental Evidence from Energy Rebate Programs

Economic policies often involve dynamic interventions, in which individuals receive repeated treatments over multiple periods. However, existing economic studies typically focus on static targeting, potentially overlooking valuable information generated by dynamic interventions. We develop a framework for designing optimal dynamic targeting policies that maximize social welfare. We show that dynamic targeting can outperform static targeting through several key mechanisms—learning, habit formation, and adaptive targeting effects—and develop a method to empirically identify these effects. Using a sequential randomized controlled trial (RCT) of a residential energy rebate program, we find that dynamic targeting nearly doubles the social welfare gains relative to static targeting and produces more than five times larger welfare gains than non-targeting policies. We also investigate how households’ anticipatory responses to adaptive assignment rules affect the welfare gains from dynamic targeting.