We estimate the effects of a large programme of public investment subsidies granted to Italian firms in disadvantaged areas. Projects were given numerical scores according to objective criteria and local politicians' preferences, and funded in rank order until the funds were fully allocated. We estimate that subsidies increased investment by marginal firms near the cutoff by 39 per cent and employment by 17 per cent over a 6-year period. Building on recent advancements in the econometrics of regression discontinuity designs, we characterize heterogeneity of treatment effects and cost-per-new-job across inframarginal firms away from the cutoff. Firms ranking high on objective criteria and firms preferred by local politicians generated larger employment growth on average, but the latter did so at a higher cost per job. Under a policy invariance assumption, we estimate that eliminating political discretion and relying only on objective criteria would reduce the cost per job by 11 per cent, while relying only on political discretion would increase the cost by 47 per cent. The effect of political discretion is larger in southern regions, which received the largest share of funds and exhibited the highest cost-per-job under the actual allocation criteria
