Evaluating Small Businesses' Performance in Public e‐Procurement: Evidence from the Italian Government's e‐Marketplace

Abstract

e‐Procurement is widely advocated as an effective tool to promote the participation of small and medium enterprises () as well as micro enterprises in the public procurement market. However, little evidence exists as to what factors may influence the allocation of contracts to differently sized suppliers. We address this issue by analyzing a rather rich and unique data set, namely the “direct award” (low‐value) transactions that took place on the Italian e‐marketplace during the period 2005–2010.

In this paper, we exploit data from low‐value transactions to test the impact of different structural dimensions of the national public e‐marketplace on the probability that firms of different size are awarded public contracts, particularly focusing on micro and small firms. We find that both the nature of the public buyer and the characteristics of the traded goods/services do have an impact on different firms' ability to be awarded public contracts. We also test to what extent the geographical distance between public buyers and private suppliers explains the propensity of different size class of firms to be awarded public contracts, proving, at least to some extent, that some features of “physical” procurement markets are mirrored in the “virtual” market.

Key words:

• H57
• L11
• L25
• L81

Notes

1 Namely 20 million of SMEs, except financial business. Figures on SMEs retrieved from “Key figures on European business with a special focus feature on SMEs” by Eurostat (2011) (http://epp.eurostat.ec.europa.eu/statistics_explained/index.php/Small_and_medium‐sized_enterprises) and from the European Commission (2004a, 2004b) (http://ec.europa.eu/enterprise/policies/sme/index_en.htm). Essential figures on SMEs in Europe and in Italy are reported in the Appendix.

2 See Meyer (2011).

3 On January 1, 2010, the European Commission Regulation (EC) n. 1177/2009—published on the Official Journal of the European Union of January 1, 2009 (L314/1964)—for goods and services, introduces new relevant thresholds for public contracts values above which the European Directives apply, namely €125,000 for central government bodies and €193,000 for all other entities. For below‐the‐threshold transactions, on the other hand, the provision in the art. 238 of the Italian Legislative Decree 163/2006 (Public Contracts Code) provides for looser selection and awarding mechanisms.

4 AVCP Annual Report 2010 (Autorità di Vigilanza sui Contratti Pubblici (AVCP) 2011).

5 The Italian National Institute for Statistics (ISTAT), in its Annual Report 2010, reported that in Italy, enterprises with fewer than 20 employees represented the 98.1 percent of the total number of firms, occupying 58.8 percent of the overall workforce, and realizing 38.4 percent of the total revenue and 44.1 percent of the added value.

6 MePA is the Italian acronym for “Mercato elettronico della Pubblica Amministrazione.”

7 The MePA counted for a turnover of €700 million in the period 2005–2010, respectively €234 and €465 million for DPs and RFQs. Despite the fact that MePA was started in 2003, we analyze the data starting from the year 2005, when the implementation of the instrument reached its consolidation phase.

8 See, among others, Gaudeul and Jullien (2001), Rochet and Tirole (2003), Jullien (2005, 2006), Armstrong (2006), Galbreth et al. (2007). In particular, Mithas and Jones (2007), exploring empirically some set of B2G transactions, confirm that electronic marketplaces stimulate the economic competition, giving an increase in the surplus of the purchasing party (increase of direct and process share of savings).

9 See, among the others, Jagoda (2010).

10 GHK Final Report (2010), “Evaluation of SMEs' access to public procurement markets in the EU.”

11 Beside the large diffusion in EU, digital procedures of public procurement are implemented also in the United States. Many Latin American countries are realizing modern electronic platform to ensure transparency in most of the different stages of the awarding of purchasing public contracts. South Korea still represents the most established reality of electronic public procurement, since it was one of the first countries to introduce it.

12 After having concentrated its main activities on the electronic management of the public authorities' financial accountability (Legislative Decree 414/1997), with particular focus on the management and development of the IT activities connected to the Ministry of Finance (Ministry Decree of the 22nd of December 1997 and of the 17th of June 1998), Consip core business moved to the rationalization on the purchasing activities of public authorities, as specified in the 2000 Italian Financial Law. In conclusion, with the Ministry Decree of 24th February 2000, Consip shifted its effort to implement the public procurement e‐marketplace.

13 AVCP Annual Report 2010 (Autorità di Vigilanza sui Contratti Pubblici (AVCP) 2011).

14 Qualification requirements are quite loose, consisting mainly of not having been convicted for major criminal offences.

15 The type of DP (say, ICT, stationery, etc.) indicates the prevailing nature of the goods/services included in the basket. A DP may well include heterogeneous items from different catalogues.

16 The number of DPs from 2006 and 2007 rose roughly by 2.4 times.

17 Since there is a lack of reliable data concerning the turnover of the firms involved in the MePA, we classified the size of the firms relating to the number of employees as depicted in the EUROSTAT scheme: micro enterprises (0–9 employees), small (10–49 employees), medium (50–249 employees), and large (≥250 employees).

18 Differences with respect to data presented in Table 2 are due to the exclusion—from the data depicted in the descriptive statistics—of 347 transactions that did not include the dimension of the suppliers and 130 transactions where the value per contract was negligible (less to 1€).

19 As for usual binary models, the OLM is a nonlinear model. Despite the classical OLM (where the estimated coefficients affect the estimated probability that a certain phenomenon happens as a result of the level of all the independent variables included in the model), we relax progressively those assumptions on the strong proportionality of estimated coefficients of the classical model, estimating some generalized ordered logit models (GLM) with nonparallel‐lines (NPL) and partial‐parallel‐lines (PPL) assumptions. See Table A1 in the Appendix.

20 HCPI stands for Harmonized Consumer Price Index.

21 The odds ratio indicates the ratio between the probability that the “higher” event (Y = 2, or 3) happens (p) and the probability that it does not happen (1‐p), where (1‐p) is the probability of the “lower” event (or the reference base to which the effects of explanatory variables are evaluated, say, Y = 1 or the awarded supplier is a micro firm). If p/1‐p < 1, the probability that the “lower” event happens is higher, vice versa p/1‐p > 1 suggests a higher probability that the “higher” event happens.

22 See the seminal work of Jan Tinbergen (1961).

23 There are many examples of empirical studies framed on the gravity equation: For instance, Harrigan (1993) evaluate the effect of trade protection in OECD countries, whereas Frankel, Stein, and Wei (1995) analyze the effects of regional trade agreements: Through the gravity equation model, they identify trading blocks in the Western hemisphere and arrive to an interesting welfare analysis. Similarly, Frankel and Wei (1993) search for a stable relation between trading blocs and currency blocs, moving the analysis on the exchange rate variability. Conversely, Alesina, Barro, and Tenreyro (2003) analyze the possibility that optimal currency unions may benefit countries that have frequent bilateral trade exchanges.

24 In the classical literature of empirical international trade, the distance in the gravity equation is a complex variable which consists of a number of specific import–export factors explaining not only the physical distance between parties but also the ease of trade or multilateral trade resistance (policy factors, tariff barriers, common borders, cultural language religious, and legal order affinities). In our model, we concentrate only on the physical distance between players. Moreover, the dependent variable—the value of each DP—has a disaggregated nature, thus in stark contrast with the aggregated approach used in the classical gravity model applied to international trade.

25 The validity of this procedure is strictly connected to the assumption of independence of the error term and the regressors involved. If this is not true, as in many empirical cases, the estimators may be inconsistent when the error terms are heteroskedastic (Silva and Tenreyro 2006). Given that the OLS model may be risky for the previous reason, we implement other models: First of all, we estimate weighted least squares (WLS) for reducing the distributional problems related to the residual issue and generalized least squares (GLS) or heteroskedasticity‐consistent White's estimator. Finally, the Pseudo‐Poisson Maximum Likelihood (PPML) estimator provided in Silva and Tenreyro (2006) deals effectively with the problem of heteroskedasticity, displaying major properties of robustness and efficiency. The PPML estimator is applied to the levels of variables, thus estimating directly the nonlinear form of the gravity model.

Additional information

Notes on contributors

Gian Luigi Albano

Gian Luigi Albano is head of research at Consip SpA and professor at the National School of Administration (SNA).

Federico Antellini russo

Federico Antellini Russo is senior economist at Cassa Depositi e Prestiti SpA.

Gionata Castaldi

Gionata Castaldi is PhD student of economics at University of Tor Vergata.

Roberto Zampino

Roberto Zampino is senior economist at Consip SpA.