Detecting collusion in public procurement: lessons from data-driven screening tools

Abstract Collusion in public procurement remains a major challenge for governments. Even when direct evidence of communication is unavailable, coordinated bidding often leaves identifiable traces in auction data. Across several studies, several coauthors and I have developed screening tools using observed bids as inputs. These tools exploit the fact that colluding firms must coordinate their behavior and therefore generate bidding patterns that diverge from those produced under competitive forces. The tools are based on three principles: persistence of designated winners across rebid rounds, the absence of close losing bids, and the presence of rotation or market division patterns. This survey synthesizes the findings of my previous work emphasizing intuitive reasoning over technical detail.