Jonathan Old

University of California, Berkeley · Economics · 2026–27 job market

Fields

Political Economy, Development

Job market paper

Measuring Bid Rigging at Scale: Detection and Prices in Indian Procurement

Competition agencies have long recognized collusion in public procurement as a priority, but little is known about how common it is or how much it raises prices, especially in contexts with poor data and little enforcement. This paper develops an empirical model that detects collusion among bidders in public procurement auctions when cartels use different coordination strategies, and applies it to the universe of tenders on India's government-developed e-procurement portals. Starting from twelve behavioral markers, such as bids submitted minutes apart or rotation of the winning bidder, the model identifies colluding pairs of bidders from the excess frequency of combinations of markers relative to a benchmark constructed under the assumption of no collusion. It estimates the prevalence of collusion and of each coordination strategy, and yields a ranking of bidder pairs by how likely they are to collude. Estimated on a new dataset of Indian public procurement covering 42.3 million bids on 15.7 million tenders since 2008, the model finds that collusion is widespread and most common under rural local bodies and in markets with few firms. Tenders won by the pairs most likely to collude settle 3.2 percent above comparable tenders in the same market on average. Linked rural road contracts show no difference in quality. The cost of collusion is therefore primarily a price increase, which, extrapolated across India, implies an annual transfer of approximately USD 4.0 billion from the government to colluding firms.