Reported Corruption vs. Experience of Corruption in Public Procurement Contracts

This paper examines the accuracy of estimates of corruption reported in business surveys by comparing reported experience of corruption in public procurement from Malagasy firms having won public contracts with a more objective measure of corruption in the sector using a red flag indicator of corruption risk. This red flag indicator of corruption identifies contracts that failed to comply (or circumvented) public procurement regulations. We find that about 68 percent of public contracts in Madagascar in 2013 and 2014 were awarded with a method not complying with the Public Procurement Code and classified according to our methodology at risk of corruption, with 85% of contracting firms having won at least one corruption-prone contract. Matching public procurement data in Madagascar with a firm-level survey in 2015 among firms awarded public contracts in 2013-2014, we find that experience of corruption has no influence on firms' survey participation or propensity to answer questions about corruption.