Pursuing Deficit Reduction Through Diversity:
How Affirmative Action at the FCC Increased Auction Competition
by Ian Ayres* and Peter Cramton**
*William K. Townsend Professor of Law, Yale Law School.
**Associate Professor, University of Maryland, Deptartment of Economics.
Steve Bainbridge and Jennifer Brown provided helpful comments. Franklin Parlamis provided excellent research assistance.
In 1993, Congress for the first time authorized the Federal Communications Commision (FCC) to auction licenses to use slices of the radio spectrum. Since then, the FCC has been busy conducting a series of auctions which have raised $9 billion. In authorizing these auctions, Congress required the FCC to “ensure that … businesses owned by members of minority groups and women are given the opportunity to participate in the provision of spectrum-based services, and for such purposes, consider the use of tax certificates, bidding preferences, and other procedures.” Relying on this statutory mandate, the FCC granted firms controlled by women or minorities (so called “designated entities” or “DEs”) substantial bidding preferences.
This article focuses on the “Regional Narrowband” auction of 30 licenses for advanced paging services (which, for example, might both transmit and receive messages). Designated entities in the regional narrowband auction were (1) allowed to pay for any of the licenses in installments over 10 years at a favorable interest rate and (2) granted a 40 percent bidding credit on 10 of the 30 narrowband licenses. The combined effect of these preferences was that favored bidders only had to pay the Government 50 percent of a winning bid.
The FCC’s affirmative action has been criticized as being a huge giveaway, but this article will show that the bidding preferences increased the government’s revenue by more than 15 percent – an increase in total revenues of more than $50 million. Although at first blush it would seem that allowing DEs to pay fifty cents on the dollar would necessarily reduce the government’s revenue, we will show that subsidizing DEs created extra competition in the auctions and induced the non-subsidized firms to bid higher.
The non-subsidized firms bid more both because they had to compete against the subsidized DE bids and because they had fewer licenses to compete for (once the substantial DE preferences effectively set aside 10 of the 30 licenses). The regional narrowband auction provided a vivid example of how subsidized bids by a minority-controlled firm could substantially increase the price that the government received from a non-minority firm. Early in the auction, a non-DE firm (PageMart) attempting to aggregate a national license (by bidding for all 5 regional licenses on a particular frequency block) had succeeded in outbidding all of its non-DE rivals by bidding a total of $76 million. However, a minority-controlled bidder, PCS Development, entered the fray, upping the ante more than a dozen times and forcing PageMart to ultimately bid $93 million to win the licenses. The additional competition from the minority-controlled firm increased the government’s revenue by $17 million. The extra revenue the Government earned from unsubsidized winning bidders – such as PageMart – more than offset the subsidy to the Designated Entities. Far from being a giveaway, affirmative action bidding preferences induced competition that prevented established firms from buying the airwaves on the cheap.
Our positive thesis then is that affirmative action can enhance bidding competition. Of course, affirmative action can only enhance competition in the limited contexts where competition among unsubsidized bidders would otherwise fail – for example, if there was a shortage of serious unsubsidized bidders or if these bidders would (explicitly or tacitly) collude. But the ability of affirmative action to enhance competition is not limited to situations where the government is a seller. Indeed, the government buys far more then it sells – and affirmative action bidding preferences can reduce the cost of government acquisitions for the same reasons. Unsubsidized suppliers competing against subsidized bidders may lower their bids to increase their chances of winning the contract. Moreover, this article reveals a potential profit motive for private affirmative action. If competition among the strongest job applicants is not sufficient for the employer to extract all the gains of trade from the employment relationship, then employers may have an incentive to subsidize weaker candidates and thus induce stronger applicants to work harder (or for a lower wage).
Before proceeding, it is useful to say a few words about terminology. A “weak” bidder is simply a bidder who has a lower expected reservation price. In the regional narrowband auction, the affirmative action subsidies were premised on the FCC’s belief that firms controlled by women and minorities had a lower ability to pay for licenses (in part because of discrimination in credit markets). Our description of DEs as relatively weak bidders is intended only to connote that these bidders may have lower expected reservation prices.
While we show that affirmative action at FCC auctions decreased the budget deficit (and might plausibly be used to reduce government procurement costs or increase private profits), we are not arguing that this revenue-enhancing effect is normatively sufficient to justify race- or gender- conscious decisionmaking. Indeed, using affirmative action to reduce the budget deficit would not satisfy either prong of Adarand’s strict scrutiny analysis: we doubt that raising additional revenues would qualify as a “compelling governmental purpose” and the race- and gender-conscious means would not be “narrowly tailored” to further that goal – because race- and gender-neutral subsidies of small bidders would likely be able to similarly enhance the government fisc.
The revenue-enhancing effect of affirmative action, however, may establish a necessary condition for justifying the subsidy. Affirmatively subsidizing the participation of women and minorities (even to remedy past discrimination) might be ill advised if the subsidy were too costly. Showing that even substantial subsidies need not drain the treasury might accordingly be a prerequisite for garnering legislative support, even if the revenue effect is not constitutionally sufficient by itself.
The relevance of showing that affirmative action subsidies do not burden the treasury is manifest in current debates concerning California ballot initiatives to end state-sponsored affirmative action. Proponents of these initiatives are trumpeting estimates of the nonpartisan legislative analyst’s office which claims that the state could save tens of millions of dollars annually if affirmative action were eliminated. These estimates assume that affirmative action increases the state’s procurement costs whenever the state rejects a low bid to contract with an historically disadvantaged firm. But the take-home lesson of this article is that affirmative action may not be nearly as costly as implied by crudely estimating the size of the subsidy. In the procurement context, there is anecdotal evidence that affirmative action bidding subsidies have destabilized tacit collusion among the non-subsidized bidders and reduced the average cost of procurement. The all-too-familiar story of a few government suppliers entering inflated, collusive bids can be rewritten by affirmative action which subsidizes new entrants to spur more competitive bidding. An unidentified source at Caltran reports that affirmative action has reduced the price of winning construction bids closer to independent estimates of construction costs. As with the government sale of communication licenses, affirmative action subsidies may be much less costly (or may even aid the treasury) if the subsidies enhance bidder competition.
But in emphasizing the normative relevance of enhanced bidder competition, it is also important to emphasize that increasing the government’s revenue (or decreasing the government’s cost) does not mean that affirmative action subsidies promote efficiency. Indeed, the beneficial impact on the government’s revenue from bidding subsidies will often come at the cost of some economic inefficiency – because in equilibrium some contracts will be awarded to lower-valuing buyers (or higher-cost producers). While enhancing market competition usually increases efficiency, the enhanced bidding competition caused by affirmative action merely allows the government to capture more of the gains of trade – usually at the cost of some inefficiency. These inefficiencies, however, may be short term if the affirmative action promotes new entry that stimulates subsequent market competition. Moreover, there may be no efficiency loss if the smaller expected DE bids reflect simply a lower inability to pay (caused possibly by discrimination in credit markets) instead of a lower prospective ability to supply paging services.
This paper is divided into four parts. Part I analyzes a series of game-theoretic examples to show how bidding preferences could enhance government revenue. In Part II, we argue that this theoretical possibility occurred in the FCC’s regional narrowband auction. Part III identifies a limited set of other contexts – including some government procurement and private employment settings – where affirmative action might be profitable. Finally, Part IV explores the normative implications of the revenue-enhancing effect.