Anti‐poverty programs increasingly target multiple outcomes to address current and future poverty. Conventional evaluation exercises, however, estimate programs’ impact on outcomes separately. We present a framework, drawing from the counting approach, that captures the joint distribution of outcomes and allows evaluating program impact on the distribution of multiple outcomes. We apply the framework to analyze the Philippine conditional cash transfer program using an embedded randomized control trial survey. We observe that the program induced targeted behavioral changes by reducing overall noncompliances to conditionalities, but did not necessarily benefit the poorer families experiencing a larger number of simultaneous noncompliances. The overall positive impact is achieved by leaving the neediest behind. Our results show the valuable contribution of considering the joint distribution of outcomes in evaluating poverty alleviation programs.