The law of small numbers: investigating the benefits of restricted auctions for public procurement

ABSTRACT

A commonly accepted view in the academic literature is that dispensing with competition may only be beneficial when tendering complex contracts. However, restricted auctions are frequently used among EU member states to procure small contracts. In this article, we investigate this paradox. Using an original data set of 180 contracts used by a local public buyer of social housing between 2006 and 2009, we show that limiting competition may enable economies to be made on transaction costs while the most efficient bidders still come forward, and that abuses such as corruption or favouritism do not result. To our knowledge, this article is the first to shed light on the advantages of using restricted auctions when tendering small simple contracts.

KEYWORDS:

• Auctions
• competition
• public procurement

JEL CLASSIFICATION:

• D44
• L22
• D21

Acknowledgements

The authors are grateful to Giancarlo Spagnolo and Steve Tadelis for helpful comments on previous versions. The authors would also like to thank Paris Habitat-OPH for providing data and useful information. Any remaining errors are the authors’ responsibility.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

¹ See the OECD report, entitled ‘Performance Measurement’ (2011), dedicated to public procurement.

² Coviello, Guglielmo, and Giancarlo (2017) used Italian procurement data on public works to compare the outcomes of restricted and open auctions, finding that open auctions decrease the probability that the contract of an incumbent firm will be renewed.

³ These thresholds vary regularly; over the period of study, they were around 200,000€.

⁴ See, for instance, this extract from European Commission (2010): ‘study confirms that the higher the value of the contract, the less the likelihood that an SME will win the contract. The threshold above which SMEs are seemingly disadvantaged is around 300,000 €’.

⁵ The contracts we study are small (43,234 € on average) and rarely renegotiated.

⁶ See, for instance, the Green Paper related to the revision of EU directives. On page 11 of the synthesis of replies, it is reported that ‘a broad majority of respondents from all stakeholder groups consider that the Directive should explicitly allow contracting authorities to take into account their previous experience with one or several bidders’.

⁷ For instance, according to the annual report of Paris Habitat-OPH in 2008, the main local public operator in social housing in Paris and the focus of the empirical part of our study, contracts below the EU thresholds account for 55.7% of the total number of contracts but only 3.6% of the total value

⁸ This decreasing rate is apparent in several empirical studies evaluating the impact of the number of bidders on received bids, which conclude that there is an optimal number of bidders (see, e.g. Amaral et al., [2013])

⁹ See, for instance, the works of Demsetz (1968) or Bulow and Klemperer (1996), which formalize the benefits of competition for the market.

¹⁰ It should be noted that small contracts usually attract more bidders than more complex ones, strengthening the arguments concerning the cost of organizing calls for tenders and selecting offers as related to the value of the contract.

¹¹ For example, woodworking, isolation.

¹² The pools are generally renewed every 2 years.

¹³ 10 categories of activity are identified by Paris Habitat-OPH. We analysed 24 different pools; on average there are 9 contracts per pool and 18 contracts per category of activity.

¹⁴ For each project, the buyer makes his own estimate.

¹⁵ We only have this information for the most recently constituted pools.

¹⁶ Hong and Shum (2002) are focusing on winner’s curse effect associated with common value auctions that often lead contracts to be renegotiated ex post. We are concerned in this article with private value simple auctions. The main reason why the public authority (in our case) is restricting the number of bidders is not because of winner’s curse effect but because they want to reduce ex ante transaction costs (ex post renegotiations are not an issue in our simple contracts data set).

¹⁷ Some of these data were obtained from the internal database of the buyer concerned, while other data were obtained from on-line sources, thanks to websites that gather information on firms’ characteristics (such as societe.com, manageo.com for instance).

¹⁸ See Evaluation of SMEs’ Access to Public Procurement Markets in the EU (2010).

¹⁹ $\tilde{E}mploye{e}_i$=$Employe{e}_i-\overline{Employee}$. See pages 68–69 of Chapter 4 of Wooldridge (2001) for an explanation of the need to use demeaned variables when using interaction terms.

²⁰ Table 2 also contains information about the characteristics of the contracts. Given that our first concern is to assess the impact of firm characteristics on the probability of being invited to post a bid, characteristics of contracts are excluded from the invitation phase estimates. Moreover, integrating these characteristics in the selection phase has no significant effect either on the invitation phase or on our main findings regarding its impact on the bidding phase.

²¹ Additional specifications were run and are explained below regarding the quality of low bids.

²² We do not include firm fixed effects in our second stage regressions. Firms’ unobserved heterogeneity is left in the residuals and can be interpreted with Heckman models: These models provide statistical information regarding the relationship between unobservable variables of both stages. Thus, results show that firms’ unobserved heterogeneity is negatively/positively correlated between both stages and we can conclude that the firms which are the more likely to be invited are those who post the lower/highest bids.

²³ See, for instance, De Silva, Kosmopoulou, and Lamarche (2009), which use both measures.

²⁴ We thank an anonymous referee for pointing this out.