Bidding Frictions in Ascending Auctions∗

This paper develops a strategy for identifying and estimating the valuation distribution in ascending auctions where bidders have an unknown number of bidding opportunities. Our model leads to equilibrium jump bidding and potentially inefficient allocations. Our identification approach requires few assumptions on the types of equilibria played. In particular, we need not assume that a unique equilibrium exist or that it be separating. We apply the model to data drawn from a monthly financial market, where the treasurers of state governments purchase from local banks savings vehicles, called certificates of deposit (CD), for unallocated state funds via a procurement auction mechanism based on the offered interest rate. In this market, the state acts as a liquidity provider that allows banks to invest in opportunities that are private to the bank. The data exhibit features that are inconsistent with standard auction formats. For example, the distribution of final bids first order stochastically dominates the distribution of winning bids, which are sometimes made early in this open ascending auction. These features motivate our model of bidding frictions. We prove that the distribution of valuations, but not the individual valuations, are identified and estimation is feasible even in the presence of multiple and mixed strategy equilibria. When bidding frictions are ignored, the monotonicity property of the sealed bid framework is rejected, and the estimated valuations are orders of magnitude greater than the corresponding final bids, creating the impression that banks have a large amount of market power. The model with frictions generates more sensible valuation estimates and provides insight into a local banking market. Since our data cover the 2008 credit crisis, we can analyze the local financial effects of the recession. The estimates of frictions before and after the crisis suggest that banks face inertia in the re-organization of their activities even after large negative macroeconomic shocks have hit. We show that sealed bid mechanisms perform better than the current mechanism and that these benefits are greater in the post-2008 period.