A major determinant of input subsidy programs’ effects on the achievement of national policy goals is the extent to which they raise total fertilizer use. This study synthesizes recent literature on how the new generation of targeted input subsidy programs has affected national fertilizer use after accounting for crowding effects, and derives benefit–cost (BC) estimates of the fertilizer subsidy programs for Kenya, Malawi, and Zambia after accounting for crowding out and diversion. We highlight two major findings. First, accounting for the illicit diversion of program fertilizer can profoundly influence estimates of how fertilizer subsidy programs affect total fertilizer use and program impacts. Given recent evidence that 33% or more of total program fertilizer may be diverted before being received by intended beneficiary farmers, the failure to account for program diversion is shown to overestimate the contribution of the subsidy programs to national fertilizer use by 67.3% in the case of Malawi, by 61.6% for Zambia, and by 138.0% for Kenya. The second major finding is that the incremental value of maize output produced from these subsidy programs is considerably less than their costs in most years, except under unusually high maize price assumptions. Conventional BC analyses that do not account for crowding out and diversion may seriously overestimate the benefits of input subsidy programs. Greater attention to program design and implementation details to reduce problems of crowding out and diversion can substantially raise the returns to such programs.