Collusion Suspicion Among Bidders of Construction Supervision Projects in China: Based on Bidding Big Data and Complex Networks

Collusion is unethical and anti-competitive behavior, which has become a prominent issue in the construction industry. However, existing studies were unable to identify the degree of collusion suspected for each bidder, which has little significance to regulators. Therefore, we develop a bidder network for the supervision project in Shaanxi Province, China, and analyze the behavioral characteristics of bidders from three levels. The results indicate that the bidding network is low-density, scale-free, and small-world at the macro level. The possibility of potential collusion is extremely high among supervisory enterprises. At the meso level, the bidder network can be divided into multiple communities and subgraphs with a high degree of cohesion, and the level of conspiracy suspicion was divided into four categories. At the micro level, the number of co-bidding among the bidders with the highest suspicion was extremely high, with higher winning bid times and significantly abnormal regularity of the bidding price, which is typical of bid-winning rotation practice. These results provide an effective method for Chinese government regulators, which are conducive to the formation of a good collusion regulatory system worldwide.