ترجیح پیشنهاد در مزایده های جواز: اقدام مثبت می تواند به بهره وری اقتصادی دست یابد

If allocative externalities are present among bidders such as when they interact subsequent to the auction, their valuations for the item may differ from their contributions to the social welfare. This paper shows that bid preference in auctions given to those bidders who can contribute more to the social welfare relative to their valuations is an effective measure to achieve efficiency, that is, social welfare maximization. This paper therefore provides a rationale in terms of efficiency for the practice of granting affirmative action bid preferences to minorities or other designated groups. This insight may be applicable to the broader issue of affirmative action programs in general as well.

It is common in many government auctions that a designated group of bidders are treated preferentially. A famous example is the Federal Communications Commission's license auctions for radio spectrum. In these auctions, the FCC has granted businesses owned by minorities and women substantial bidding credits, tax certificates, and other preferential treatments. In particular, favored bidders in the “regional narrowband” license auction for paging services were given a 40% bid preference or bidding credit, so that they had to pay only 60% of a winning bid.1 This practice of preferential treatment is not special to the sales of government properties. Various governmental agencies such as federal, state, city, and administrative bodies use minority price-preferences in their procurement programs as well. In international trade, many national governments favor domestic firms by explicitly giving bid preferences. Thus, a foreign firm can win a contract only when its bid is lower than the lowest domestic bid by more than a specified bid preference. The United States Government, for instance, has maintained the “Buy-American” program, in which 6% up to 50% bid preferences are granted to domestic suppliers in procurement contracts.2 These programs are controversial. The bid preference programs in international trade are recognized as nontariff barriers, and their validity has been challenged. The government auctions and procurement programs, which can be broadly classified as the affirmative action programs, are under serious debate regarding their rationale and effectiveness.3 One of the grounds against these programs is that they increase the cost of government. While the origin of these programs is probably political, some recent studies provide an economic justification for them. In particular, McAfee and McMillan (1989) show that the domestic preferences in international procurements can lower the expected price paid by the government for the item. Similarly, Ayres and Cramton (1996) argue that the affirmative action bid preferences in the FCC auctions increased the government's revenue from the sale of the licenses. Corns and Schotter (1999) also prove both theoretically and experimentally that preferential treatment in government procurements can decrease government's cost of purchasing. What the studies have established is that these programs are not as costly as they appear. On the contrary, bid preferences can increase the government's revenues in the sales of government properties and decrease the government's costs in procurements by creating effective competition among bidders. Due to the preferences given to a subset of bidders, the unfavored bidders have to compete more fiercely among themselves (intragroup competition) as well as with the favored bidders (intergroup competition). This leads to an increase in government's net revenue compared to the situation where no such preferences are given at all. Note that this conclusion is in fact based on Myerson's (1981) theorem on optimal auction design: The seller of an item can increase his revenue by giving bid preferences to weak bidders whose expected willingness to pay for the item is lower. A similar logic applies to the procurement programs in that the buyer of an item can decrease her purchasing cost by giving bid preferences to weak bidders whose expected cost of provision is higher. These studies, however, miss a more fundamental economic issue in these programs: The issue of whether they improve or impede efficiency.4 This paper aims to provide an answer to this question. We show, with a stylized license auction model, that bid preferences are effective in achieving efficiency when allocative externalities are present among bidders. If an allocation of the license to a bidder may affect other bidders such as when bidders in an auction interact subsequent to the auction, it may happen that a bidder's valuation for the license may differ from the social welfare he can contribute with the license.5 Hence, traditional auctions which assign the license according to the valuations may produce an inefficient outcome. The main contribution of the present paper is to show that auctions with properly-set bid preferences can eliminate this discrepancy between valuation and social welfare. By granting bid preferences to weak bidders whose valuations are relatively lower than the social welfare they can contribute, the government can ensure that the license be assigned to a bidder with the highest social contribution. The rest of the paper is organized as follows. We set up a simple Cournot duopoly and present the key observations of the present paper in the next section, and then generalize the model in the following section. Although we present the idea in license auction models in which government properties or public projects are sold, it is clear that the insight can be symmetrically applied to procurement models. The final section contains a fuller discussion of this and related issues.

We have shown that bid preference in auctions is an effective measure to achieve efficiency when bidders' valuations are different from their respective contributions to the social welfare. We observed that this discrepancy may occur if allocative externalities are present among bidders, such as when they interact subsequent to the auction. We have found that bigger bid preferences should be granted to those bidders who can contribute more to the social welfare relative to their valuations. We have demonstrated this result in Section 2 with a stylized Cournot duopoly in which firms have different marginal costs of production. As the analysis in Section 3 implies, however, this insight can be equally applied to more general situations. For instance, if firms have differing capital costs or budget constraints, then bid preferences given to financially weak bidders can fill the difference between the social welfare and valuation to achieve efficiency. The same insight may also shed some light on the broader issue of affirmative action programs in general. In particular, the fact that a group of people or firms have disadvantageous opportunities alone is not sufficient for affirmative actions. Another condition is allocative externalities, so that this group interacts with others and contributes more to the social welfare. I believe this condition holds in many situations. Bidders in our bid-preference auction for the incomplete information case report their private information rather than monetary bids. That is, we have constructed a direct revelation mechanism. While this faithfully follows the common practice in the literature,13 we still admit that the auction may not be applicable to real-world situations. Note however that the first-price auction with bid preference for the complete information case is real enough to be applicable. Moreover, to correctly set the magnitude of bid preferences, the government or the auctioneer needs to know the structure, i.e., the functional form, of bidders' valuations and the social welfare they contribute. This might be a hard task, which is probably why we often observe bid preferences that are fixed to a pre-specified level across many runs of the same auction.14 On the other hand, the main purpose of the present paper is to provide a new perspective on affirmative action bid preference, not its practical implementation.