We estimate the impact of a large anti-poverty program â€“ the Uruguayan PANES â€“ on political support for the government that implemented it. The program mainly consisted of a monthly cash transfer for a period of roughly two and half years. Using the discontinuity in program assignment based on a pre-treatment score, we find that beneficiary households are 21 to 28 percentage points more likely to favor the current government (relative to the previous government). Impacts on political support are larger among poorer households and for those near the center of the political spectrum, consistent with the probabilistic voting model in political economy. Effects persist after the cash transfer program ends. We estimate that the annual cost of increasing government political support by 1 percentage point is roughly 0.9% of annual government social expenditures.