Collusion risk in corporate networks

Abstract

Abstract Collusion among economic operators increases prices, reduces product quality, and hinders innovation. Structural links can affect the incentive and ability of firms to behave competitively by facilitating collusion. We use a network-based approach to study the relationship between ownership links and bidding behavior in procurement markets. We build temporal multiplex networks based on firms’ ownership and co-bidding ties to find network measures that may signal collusion risk. We test four network measures, two at market-level (density and average harmonic closeness centrality) and two at firm-level (degree centrality and harmonic closeness centrality). Using data on public procurement contracts awarded in Sweden from 2010 to 2015, we found higher incidence of single bidding in markets that are more closely related through ownership links. Missing bidders are also more likely in these markets. Single bidding and missing bidders may indicate the presence of collusive arrangements such as of bid suppression or rotation. For the firm-level analyses, our results showed a positive relationship between winning probability and centrality in the ownership network. A similar result was obtained for cut-point position, indicating that firms that are more closely connected to other firms through ownership links have a more important position in the co-bidding network and are also more likely to win contracts.