Detecting Corruption in Single-Bidder Auctions via Positive-Unlabelled Learning

Corruption risks of Hungarian municipalities: Quantitative and Qualitative Analyses

“…However, the relevance of the existing red flags has to be critically investigated and their revision should be a continuous duty of researchers and practitioners referring to them (Tátrai and Németh, 2018); a red flag tool may be excessively sensitive, signalling corruption risks in every situation and vice versa, and the extent of such problems may vary due to regulatory and cultural differences and changeseven if countries have to combat similar institutional problems related to corruption and face with the necessity of the implementation of systematic detection of risk patterns in public procurement data, implying that public tendering presents several procedural similarities across countries and subnational units (Velasco et al, 2021). Comparison of single-and multi-bidder contracts in Russia with machine learning techniques also challenges the common assumption that the single-bidder rate can serve as a good proxy of corruption in public procurement (Goryunova, Baklanov and Ianovski, 2021), even if there is empirical evidence, that this indicator serves as a good basis for corruption research in dozens of countries around the world (Dávid-Barrett et al, 2020); meanwhile, more and more emphasis is being put on data mining techniques in the identification of the riskiest tenders 13 (Bratsas et al, 2021). It is also debated whether the features of public procurement understood as corruption riskmostly the single-bidder indicatorshould be measured on the level of contracts or procedures (Rigó and Kugler, 2021).…”