Wedeveloptestsforcommonvaluesatfirst-price sealed-bid auctions. Our tests are nonparametric, require observation only of the bids submitted at each auction, and are based on the fact that the “winner’s curse ” arises only in common values auctions. The tests build on recently developed methods for using observed bids to estimate each bidder’s conditional expectation of the value of winning the auction. Equilibrium behavior implies that in a private values auction these expectations are invariant to the number of opponents each bidder faces, while with common values they are decreasing in the number of opponents. This distinction forms the basis of our tests. We consider both exogenous and endogenous variation in the number of bidders. Monte Carlo experiments show that our tests can perform well in samples of moderate sizes. We apply our tests to two different types of U.S. Forest Service timber auctions. For unit-price (“scaled”) sales often argued to fit a private values model, our tests consistently fail to find evidence of common values. For “lumpsum ” sales, where aprioriarguments for common values appear stronger, our tests yield mixed evidence against the private values hypothesis.

Abstract not found

This paper considers the structural estimation of the affiliated private value (APV) model in first-price sealed-bid auctions. The model allows for bidders' individual efficiencies and opportunity costs, while permiting dependence among bidders' private values through affiliation. We establish the nonparametric identification of the APV model, characterize its theoretical restrictions, and propose a computationally convenient and consistent two-step nonparametric estimation procedure for estimating the joint private value distribution from observed bids. Using simulated bid data we provide a step by step guide on how to implement our procedure and show the good behavior of our estimator in small samples.

We develop ecopyBB#LL models of ascending (English) auctions which allow for both bidder asymmetries as well ascyBDV and/or private valuecueyV49#; in bidders' underlying valuations. We show that the equilibrium inverse bidfuncDVBD ineac round of theaucALC are implicyxV4D#Vyc (pointwise) bya system of nonlinear equations, so thatctyA;L4Ay for theexistenc and uniqueness of an incVAVyxVD;#DycxVD;#DycyB areessentiallyidenticc to those whic ensure a unique and incL###yx solution to the system of equations. We exploit the ceyCB;4#yxV; trac4#yxV;CB thiscsyB4D9VyxVCABC in order to develop anecVVCLyxVC model, thus extending the literature onstrucDyxV estimation ofaucCVB models. Finally, an empiricx example illustrates how equilibrium learning affects bidding during theceyVL of the auction.

When bids do not represent binding commitments, the use of a first price sealed bid auction favors those bidders who are less penalized from reneging on their bids. These bidders are the most likely to win but also the most likely to default on their bid. In this paper I study theoretically two methods often used in public procurement to deal with this problem: (1) augmenting the first price auction with an ex-post verification of the responsiveness of the bids and (2) using an average bid auction in which the winner is the bidder whose bid is closest to the simple average of all the bids. The average bid auction is new to economics but has been proposed in civil engineering literature. I show that when penalties for defaulting are asymmetric across bidders and when their valuations are characterized by a predominant common component, the average bid auction is preferred over the standard first price by an auctioneer when the costs due to the winner’s bankruptcy are high enough. Depending on the cost of the ex-post verification, the average bid auction can be dominated by the first price with monitoring. I use a new dataset of Italian public procurement auctions, run alternately using a form of the average bid auction or the augmented first price, to structurally estimate the bids’ verification cost, the firms’mark up and the inefficiency generated by the average bid auctions.

In this paper, we consider nonparametric identification and estimation of first-price auction models when N ∗ , the number of potential bidders, is unknown to the researcher, but observed by bidders. Exploiting results from the recent econometric literature on models with misclassification error, we develop a nonparametric procedure for recovering the distribution of bids conditional on the unknown N ∗. Monte Carlo results illustrate that the procedure works well in practice. We present illustrative evidence from a dataset of procurement auctions, which shows that accounting for the unobservability of N ∗ can lead to economically meaningful differences in the estimates of bidders ’ profit margins. In many auction applications, researchers do not observe N ∗ , the number of bidders in the auction. (In the parlance of the literature, N ∗ is the “number of potential bidders”, a terminology we adopt in the remainder of the paper.) The most common scenario obtains under binding reserve prices. When reserve prices bind, the number of potential bidders N ∗ , which is observed by auction participants and influences their bidding behavior, differs from the observed number of bidders A ( ≤ N ∗), which is the number of auction participants whose bids exceed the reserve price. Other scenarios which would cause N ∗ to be unknown to the researcher include bidding or participation costs. In other cases, the number of auction participants may simply not be recorded in the researcher’s dataset.

This paper empirically studies the consequences of unobserved heterogeneity on auction design. Unobserved heterogeneity in the objects for sale induces correlation among bidders’ valuations, which violates the standard modeling assumption of independence. We show that standard data from ascending auctions partially identifies a model of correlated private values, in a way that is useful for mechanism design purposes. In particular, if larger auctions are ones where bidders on average have higher valuations (a condition which follows naturally from models of endogenous entry), we get an upper bound on the seller’s expected profit at a given reserve price. If bidders ’ valuations are independent of auction size, a stronger assumption, then we also get a lower bound. We then show that the stronger identifying assumption of independence implies nonparametrically testable restrictions, and we develop a precise asymptotic test of these restrictions. We apply both our identification and testing results to data from the United States Forest Service. We fail to reject independence between valuations and auction size in the data. Furthermore, we find that unobserved heterogeneity has significant implications for USFS reserve price policy, suggesting substantially lower reserve prices than standard analysis and rationalizing changes made to the selling mechanism in the

Abstract not found

Programs that encourage the participation of disadvantaged business enterprises (DBE) as subcontractors have been a part of government procurement auctions for over three decades. This paper examines the impact of DBE programs on highway procurement auctions. We study how DBE subcontracting requirements affect bidding behavior in federally funded projects. Specifically, these rules require prime contractors to subcontract out a portion of a project to a DBE. Within a symmetric independent private value framework, we use the equilibrium bidding function to obtain the cost distribution of firms undertaking projects either with or without subcontracting goals. We then use nonparametric estimation methods to uncover and compare the cost of firms bidding on a class of asphalt projects related to surface treatment in Texas. The analysis shows little differences in the cost structure between auctions that have subcontracting goals and those that do not.

Recall private signals are normalized to be equal to bids observed in data, i.e. x = b0(x) for all x. Hence by de…nition,; l are normalized versions of vh; vl de…ned over the support of bids [b 0 L; b0 U]. For clarity, I use notations vh; vl for the original, un-normalized conditional expectations de…ned over [xL; xU]. This distinction is more than just notational, as regularity conditions in S1-S2 are only imposed on the distribution of the original, "un-normalized" private signals only. The consistency of the estimator follows from a sequence of results, most of which were shown in Li, Perrigne and Vuong (2002). First, I prove some smoothness properties of the bid distributions in equilibrium. Second, I show the kernel estimators ^ l and ^ converge in probability to l and ^ 1 br