VoxEU Column Competition Policy EU policies Industrial organisation

Discrimination in public procurements: A cost-reducing guide

Bid preferences and set-asides are popular discriminatory practices in US public procurement, but are prohibited in the EU. This column argues that discrimination can be cost-reducing provided it is targeted to favour those firms whose participation is more responsive to the auction procedure. Situations when set-asides may be cost-reducing are also discussed.

The US and EU approaches to public procurement legislation are very different. The paradigm in the EU is to select the most advantageous tender without discrimination. Thus, if competition is only over price, it means that the firm offering the lowest bid should win. Starting in 2014, a new EU directive on public procurements permits the exclusion of a participant due to that participant having caused some disappointment in the past (DIRECTIVE 2014/24/EU, Article 57). This evolution, praised by Saussier and Tirole (2015), can be viewed as providing a refinement of the ‘no discrimination’ paradigm insofar as the information about tenderers’ past performances would be used to better rank tenders, and the most advantageous tender based on this refined ranking would still be selected without discrimination. The legislation in the US is more flexible – bid preferences and set-asides are used extensively in public procurements and resource sales such as spectrum auctions. Bid preference programmes consist of giving a bonus to some bidders, allowing them to win the tender even in cases when their offer was not the best. Set-asides consist either in excluding some specific bidders (e.g. large firms or previous auction winners) or equivalently in limiting access to some well-chosen bidders (e.g. SMEs or new participants). Those instruments are used for various motives: protectionism, affirmative action, for local employment purposes, or to promote SMEs.

In this column, we focus on whether discrimination and set-asides allow for a reduction in procurement costs.1 Presumably, the benefit of discriminatory policies is that they can boost competition, and the question is to understand whether and when such a boost in competition would dominate the efficiency losses attached to discrimination. Such a perspective on discrimination is also the one adopted a few years ago by the Select Committee on Trade and Industry in the UK, which pleaded for reforms allowing for more flexibility to help SMEs as in the US legislation. Indeed, the Federal Acquisition Regulation (FAR) leaves some discretion to contracting officers to exclude one or more firms, in particular under the motives that it would “increase or maintain competition and likely result in reduced overall costs for the acquisition, or for any anticipated acquisition”. Kang and Miller (2016) report that among those (large) contracts that are not automatically reserved to SMEs, the exclusion clause is used in one third.2

An endogenous participation perspective

Discriminatory policies, by tilting the competition in favour of some firms, not only affect who is awarded the contracts within the pool of participants, but also the composition of the pool of participants, thus inducing complex trade-offs. A recent and still flourishing empirical literature has investigated the impact of those policies on procurement costs taking into account the effect on participation.

In particular, Marion (2007) establishes that in California the 5% subsidy that accrues to SMEs in auctions for road construction projects using only state funds increases the procurement costs by 3.8% as compared with projects using federal aid, where there is no such bid preference programme. He observes that the main channel for this detrimental effect of discrimination comes from the reduced participation of large more efficient firms in those auctions in which they are handicapped. In another context, Athey et al. (2013) find a positive effect of bid subsidies to SMEs up to a 20% subsidy. A notable difference between the two studies is that in Athey et al. (2013), the number of large firms that are potential participants is often very small so that the handicap is likely to have little effect on their participation decisions.

Set-asides may also have a beneficial effect on participation and be good overall from a budgetary perspective. This is clearly the case when a given firm is known to be more efficient than other firms so that the participation of the former would deter any other participation in non-discriminatory formats. Such a situation arose in the 1994 US spectrum auctions, where a license covering all of southern California was offered and where it was publicly known that Pacific Bell, the incumbent company, had a higher valuation than its rivals, and thus would win for sure (Milgrom 2004). Relatedly, in the UK spectrum auctions in the 2000s, the fact that one licence was reserved to a new entrant is considered to be an important source of its success (Jehiel and Moldovanu 2003). However, in general, the empirical evidence about the benefits of pro-SME set-asides is mixed – Denes (1997) finds no significant impact; Nakabayashi (2013) finds that set-asides reduce government costs in procurement auctions for civil engineering projects in Japan; and Athey et al. (2013) find that set-asides in US forest service timber auctions substantially reduce revenue.

Our results

In recent papers, we develop models in which bidders’ participation response to the auction procedure is of two types (Jehiel and Lamy 2015, 2016). For a subset of bidders called ‘potential entrants’, participation is perfectly elastic insofar as their entry rates adjust so that bidders’ payoffs are completely determined by their (exogenously set) participation costs. For the complementary set of bidders called ‘incumbents’, participation is perfectly inelastic – they participate for sure. Such models allow us to capture jointly two kinds of asymmetries: those relative to cost distributions, and those relative to participation elasticities.

When considering policies that favour SMEs, one has in mind that SMEs are potential entrants. Concerning large firms, whether they should be considered as potential entrants or as incumbents depends on the market conditions (as illustrated in Athey et al. 2013, where the number of large firms turns out to depend on the geographical location).

In Jehiel and Lamy (2015), we characterise the optimal mechanism. When there are no incumbents, the optimum can be implemented by assigning the contract to the most efficient firm, and thus without discrimination independently of the pre-entry characteristics. When there are incumbents, the optimum involves some discrimination against the incumbents in an attempt to reduce their informational rents. Interestingly, the optimal form of discrimination does not depend on the characteristics of the non-incumbents (in particular their cost distributions). It can be implemented with (non-linear) bid-penalties to be imposed to the incumbents.

In Jehiel and Lamy (2016), we analyse the incidence of set-asides taking the auction format as given. In efficient auctions, when the set of incumbents consists of a single firm, it is always beneficial to exclude the incumbent irrespective of how costs are distributed. Moreover, the desirability of such a set-aside policy carries over to inefficient auction formats where the incumbent is more often the winner compared to what efficiency would dictate. An example is the first-price auction, where the incumbent benefits from a right of first refusal allowing the incumbent to submit his bid after seeing the entrants’ bids, as observed in London in some public transport procurements (Amaral et al. 2009). When there are several incumbents, the question of whether set-asides reduce procurement costs is less clear-cut and typically requires finer knowledge about the cost distributions of the tenderers.


The above discussion has abstracted from moral hazard and adverse selection – for example, ex post renegotiation (Bajari et al 2014) or bankruptcy risks. This would plead for incorporating estimates of these (e.g. based on firms’ past performance) to redefine an efficiency ranking that takes the corresponding asymmetries between tenderers into account.3 Such a rescaling would be similar in spirit to the way the new EU directive enlarges exclusionary motives.

In terms of results from our studies, the EU ‘no discrimination’ paradigm (augmented by its refinements to take into account the possible asymmetries as just discussed) is the best one for procurement cost purposes when there are no incumbents. In contrast, when some large firms participate for sure, discrimination in favour of SMEs (and also in favour of some large firms if these can be considered as potential entrants) would allow to reduce the procurement costs, thereby nuancing the EU view on discrimination. A striking case is when there is a single incumbent firm participating for sure in the procurement –excluding the incumbent would reduce procurement costs independently of how strong (or weak) the incumbent is relative to his potential competitors in terms of cost distribution. Such an insight can be viewed as providing some justification to the Veolia Transport and Transdev merger case in 2010, in which the French antitrust authorities imposed that ‘Transdev Group’, the merged entity, commits to not bidding in a bunch of cities in the south-east of France for which the merged entity was the incumbent as the previous holder of the contract.4


Amaral, M, S Saussier and A Yvrande-Billon (2009) “Auction procedures and competition in public services: The case of urban public transport in France and London”, Utilities Policy, 17(2): 166-175.

Athey, S, D Coey and J Levin (2013) “Set-asides and subsidies in auctions’’, American Economic Journal: Microeconomics, 5(1): 1-27.

Bajari, P, S Houghton and S Tadelis (2014) “Bidding for incomplete contracts: An empirical analysis of adaptation costs’’, American Economic Review, 104(4):1288-1319.

Coviello, D, A Guglielmo and G Spagnolo (2016) “The effect of discretion on procurement performance’’, Management Science, forthcoming.

Denes, T A (1997) “Do small business set-asides increase the cost of government contracting?’’, Public Administration Review, 57: 441-444.

European Parliament (2014) "DIRECTIVE 2014/24/EU on public procurement”.

Federal Acquisition Regulation (FAR) (2016) “Subpart 6.2—Full and Open Competition After Exclusion of Source”.

Federal Acquisition Regulation (FAR) (2016) “Part 19—Small Business Programs”.

Jehiel, P and L Lamy (2015) “On discrimination in auctions with endogenous entry’’, American Economic Review, 105(8): 2595-2643.

Jehiel, P and L Lamy (2016) “On the benefits of set-asides”, CEPR Discussion Paper No DP11564.

Jehiel, P and B Moldovanu (2003) “An economic perspective on auctions”, Economic Policy, 18(36): 269-308.

Kang, K and R Miller (2016) “Winning by default: Why is there so little competition in government procurement?”, mimeo, Carnegie Mellon University.

Marion, J (2007) “Are bid preferences benign? The effect of small business subsidies in highway procurement auctions”, Journal of Public Economics, 91: 1591-1624.

Milgrom, P (2004) Putting auction theory to work, Cambridge University Press, Cambridge.

Nakabayashi, J (2013) “Small business set-asides in procurement auctions: An empirical analysis”, Journal of Public Economics, 100: 28-44.

Saussier, S and J Tirole (2015) “Renforcer l’efficacité de la commande publique’’, Les notes du conseil d’analyse économique, 22.

Trade and Industry Committee Minutes of Evidence (2007) “Appendix 25: Memorandum submitted by the Forum of Private Business”, House of Commons – UK parliament.


[1] The other motives, such as local employment, provide more immediate arguments in favour of explicit discrimination.

[2] Federal procurement contracts billing between $3,000 and $150,000 are automatically reserved to SMEs. In Japan, Nakabayashi (2013) reports that approximately two thirds of civil engineering contracts are subject to set-asides.

[3] Coviello et al (2016) argues that broadening public buyer discretion is a way to solve those issues through reputation mechanisms. In public procurements in Italy, they estimate that less discretion is associated with longer delays in realisation. 

[4] See paragraphs 446-450 in ‘Décision n°10-DCC-198 du 30 décembre 2010’ and also paragraph 43 in ‘Décision n°13-DCC-137 du 1er octobre 2013’ that confirms that Transdev Group has met his commitments.

1,365 Reads