In 1999, the year before Ghana's 2000 election, the country experienced a large, unexpected decline in aid. The incumbent National Democratic Congress (NDC) lost the election. Did the decline in aid hurt the NDC at the polls, or was it simply incidental? Using data from a national, World Bank-funded electrification project, this article shows that the NDC was able to allocate aid according to explicitly political criteria. The article also exploits a quasi-experiment in aid disbursements to show that electrification caused NDC voting to increase in the constituencies that received electrification. Pre-electoral aid fluctuations exert a modest but measurable force on voting patterns. These findings add weight to calls for donors to coordinate to reduce aid volatility. They also show that incumbent governments can allocate aid strategically to secure votes, even under the best-case scenario of strict donor monitoring in an established democracy.