We study the short-run effect of elections on monetary aggregates in a sample of 85 low and middle income democracies (1975-2009). We find an increase in the growth rate of M1 during election months of about one tenth of a standard deviation. A similar effect can neither be detected in established OECD democracies nor in other months. The effect is larger in democracies with many poor and uneducated voters, and in Sub-Saharan Africa, Latin America, and in East-Asia and the Pacific. We argue that the election month monetary expansion is related to systemic vote buying which requires significant amounts of cash to be disbursed right before elections. The finely timed increase in M1 is consistent with this; is inconsistent with a monetary cycle aimed at creating an election time boom; and it cannot be, fully, accounted for by alternative explanations.