Working Paper: NBER ID: w30271
Authors: Kei Kawai; Jun Nakabayashi; Daichi Shimamoto
Abstract: We apply a Regression Discontinuity based approach to screen for collusion developed in Kawai et al. (2022) to public procurement data from five countries. We find that bidders who win by a very small margin have significantly lower backlog than those who lose by a very small margin in the sample of procurement auctions from Indonesia, suggesting that bidders collude by bid rotation. Our results suggest that the proportion of noncompetitive auctions is at least about 5% for all E-procurement auctions and about 3% for all auctions in Indonesia. We cannot reject the null of competition in other countries.
Keywords: Bid-rigging; Procurement Auctions; Collusion; Regression Discontinuity; Public Procurement
JEL Codes: L41; O52; O53
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| collusion (D74) | lower backlog (D25) |
| winning by a small margin (D72) | lower backlog (D25) |
| losing by a small margin (D72) | higher backlog (D25) |
| collusion (D74) | differences in backlog (C69) |
| bid rotation (D44) | collusion (D74) |
| noncompetitive auctions (D44) | collusion (D74) |