Article dans une revue: We provide evidence of corruption in allocation of public procurement and assess its efficiency. Firms with procurement revenue increase tunneling around regional elections, whereas neither tunneling of firms without procurement revenue, nor legitimate business of firms with procurement exhibits a political cycle. Data are consistent with the corruption channel—cash is tunneled to politicians in exchange for procurement contracts—and inconsistent with alternative channels. Using the strength of correlation between procurement revenue and tunneling around elections as a proxy for local corruption, we reject the "efficient grease" hypothesis: in more corrupt localities, procurement contracts go to unproductive firms.