An empirical analysis of the impacts of competition on procurement costs

ABSTRACT

This paper empirically analyses the effect of a competitive environment on the cost of procurement. We exploit a dataset based on the Spanish procurement award notices for the years 2012–2017. After controlling for possible endogeneity, we show that a more competitive environment significantly reduces the cost of Spanish procurement. In particular, the use of the open procedure, the most competitive one, would reduce the procurement price by 9.9%, while the increase by one unit of the number of participants would on average lead to a 2.1% decrease in the procurement price. Consequently, a more extensive use of the open procedure would generate savings of €354 millions per year. These findings can have an important effect on the public deficit of a country.

KEYWORDS:

• Procurement
• public procurement
• competition effects
• auctions

JEL CLASSIFICATION:

• H57
• D44
• D23

Disclosure statement

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

Notes

¹ See ‘Government at a glance (2017)’.

² EU law sets minimum harmonized rules for tenders whose monetary value exceeds a certain amount. According to the Directive 2014/24/EU, for the central government authorities these thresholds are: €5,548,000 for works contracts and €144,000 for supplies and services contracts (with some exceptions). All contracts whose reserve price exceeds these thresholds have to be published in the European platform dedicated to public procurement (TED). Until now TED has been the main data source, however in Spain, around the 80% of tenders are below the EU thresholds (Ferrando-Gamir 2015) and they have particular characteristics, among others, a more extensive use of the negotiated procedure and a smaller number of participants.

³ The Revised text of the Law on Public Sector Contracts approved by Royal Decree 3/2011 of 14 November, the Law on public procurement in the water, energy, transport and postal services sectors 31/2007 of 30 October and the Law on Public Procurement in the defense and security sectors 24/2001 of 1 August.

⁴ Spanish regulations stipulate that for negotiated procedures without call for competition the tender announcement on the electronic platform does not need to be published, however, certain minimum information must be provided, including the specific administrative clauses document, the technical specifications document, and the award and contract formalization announcements.

⁵ These contracts have a value below EUR 18,000 for services and suppliers and EUR 50,000 for public works (these thresholds have changed with the new Spanish public procurement law that came into force in March 2019).

⁶ The reserve price of an auction represents the maximum price the contracting authority is willing to pay for the contract. Following the article 87 of the Revised text of the Law on Public Sector Contracts, the contracting bodies will estimate correctly this cost of the contract ‘considering the general price of the market at the moment of fixing the tender budget’.

⁷ The evaluation team is appointed by the contracting authority. It has different members, one of them belongs to the department promoting the tender.

⁸ EU law sets minimum harmonized rules for tenders whose monetary value exceeds a certain amount. According to the Directive 2014/24/EU, for the central government authorities these thresholds are: €5,548,000 for works contracts and €144,000 for supplies and services contracts (with some exceptions). All contracts whose reserve price exceeds these thresholds have to be published in TED. More details in https://ec.europa.eu/growth/single-market/public-procurement/rules-implementation/thresholds_is.

⁹ Tenders Electronic Daily, the online version of the Supplement of the Official Journal of the EU.

¹⁰ The national law allows the division of contracts into lots whenever these lots are suitable to use separately and constitute a functional unit. In those cases where the contract is not divided into lots, we equate the tender with a lot.

¹¹ More information about these variables in the Appendix A of the document.:

¹² See Appendix A of the document.

¹³ The contracting authorities are responsible for introducing in the platforms the data relating to the awarded tenders. We have excluded from the analysis winning bids that are higher than the reserve price of the contract, as the procurement platform used to generate the dataset does not include cost overruns in contracts, at least in the period under review.

¹⁴ The winning rebate is the complementary of the relative winning bid: the discount offered by the winning bidder expressed as the percentage of the reserve price.

¹⁵ There could be some manipulation around these thresholds in certain contracting authorities, but it is not observed in the overall set of contracts. The normalized estimated costs (Figure B2 and B3) are declining similarly to the distribution of estimated costs in the Czech Republic presented in Palguta and Pertold (2017). To identify potential contracting authorities with fraudulent behaviour, a more detailed analysis would be necessary, in line with Tas (2023)..

¹⁶ According to the sample, the contracting authorities of the ministries use the open procedures with a frequency of 57.2% while the contracting bodies not belonging to ministries use these procedures with a frequency of 49.9%. This difference is statistically significant with a 99% confidence level based on a Welch’s t-test, which does not assume equal variances and adjusts the degrees of freedom, making it more robust and suitable for comparisons between samples of different sizes and variances.

¹⁷ Generally, without distinguishing by type of contract, the median number of participants in Ministry tenders is 3, while in non-Ministry organizations it is 2. The difference in medians is significant at a 99% confidence level with the Wilcox test.

¹⁸ What underlies the selection of LAGNP as an instrument for the number of participants is that within the same agency, for the same type of contract and procedure, the visibility and ability to attract bidders of a contracting authority is relatively similar over time. Specifically, in a linear regression of the number of participants in a lot on LAGNP, an adjusted R² of 0.557 is obtained, and the LAGNP variable appears as statistically significant at the 99% level. Notice also that it does not seem economically reasonable that the number of previous participants might directly affect in the current relative bid.

¹⁹ The Register releases its data with some delay, when the estimation was made the last available data correspond to 2016.

²⁰ We have to highlight that this calculation is a very conservative one: following the OECD the public procurement spending supposed the 10.5% of the Spanish GDP in 2015, however, according to the Register, this public procurement spending supposed approximately the 2% of the Spanish GDP in 2015.

²¹ More information about the CPV in https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:32008R0213.

²² In detail, the percentile 50 is €92,389..

²³ After deleting errors, missings ….

Additional information

Funding

This work was supported by the Ministerio de Ciencia, Innovación y Universidades MICIU/AEI/10.13039/501100011033 under grant PID2022-136547NB-I00.