The (large) impact of buyer competence on the cost and timing of procurement

Francesco Decarolis, Leonardo M. Giuffrida, Elisabetta Iossa, Vincenzo Mollisi, Giancarlo Spagnolo 31 May 2018



Every year about 12% of the world’s GDP is spent by governments in procuring a wide range of works, services, and supplies (World Bank 2017). Items of very different types and complexity – from standard office supplies, computers, and cleaning services to specialised engineering consulting, hospitals, and high-speed trains – are purchased regularly by public administrations, typically by tender. 

Being under fiscal pressure, governments are promoting reforms in their procurement processes, by adopting new practices aimed at better managing this component of public spending. Some examples include capacity development strategies, the digitalisation and automation of acquisition processes, and strategic aggregation of demands through central purchasing bodies. 

Yet, a large aspect of wastage in public procurement is that which occurs when the performance acquired is delayed, contractual terms are renegotiated, and costs increase. How can this waste be reduced? A lot of media attention has typically been placed on possible episodes of corruption as a major source of waste, but in fact a lack of competence among public buyers could make an even bigger impact. Being able to make purchases of quality at fair cost is not an easy task. A suitable supplier must be selected and induced to execute the contract within the budget, and to respect deadlines and agreed quality standards. To achieve this, contracting authorities must know market characteristics and design, manage efficient award mechanisms, balance risks and incentives in the contract, and effectively monitor the contractor in the execution phase. These are not simple tasks. When problems arise, so do questions. Was the contractor efficiently selected at the award stage? Was the contract appropriately designed? Was the contractor effectively monitored during contract execution? 

In 2015 the World Bank began to release its Benchmarking Public Procurementseries. In 2017, this report presented data from 180 countries on the legal and regulatory environments affecting the ability of private companies to do business with governments. It revealed the existence of great heterogeneity in the quality of the procurement process across countries. Furthermore, questions have been raised as to whether procurement offices are well organised and managed. Saussier and Tirole (2015) show that 63% of the staff of French contracting authorities do not have a legal profile and only 39% have qualifications specific for managing public purchases. Yet, one of the central objectives of the new European Directives on Public Purchases (Directives 2014/25/EU, 2014/24/EU) is to grant greater margins of flexibility and discretion to contracting authorities, making the valorisation of their competence the cornerstone of the new European public procurement legislation. The transposition of these directives into national laws has even led Italy to envisage a system of qualification for contracting authorities based on criteria of quality, efficiency, and professionalisation (Article 38, Legislative Decree 50/2016). 

In a recent paper (Decarolis et al. 2018), we adopt an organisational perspective and focus on the competence of buyers, its drivers, and its impacts on procurement outcomes. We analyse data on all service and construction contracts worth over $25,000 awarded by US Federal Agencies from 2010 to 2015 (Federal Procurement Data System, FPDS), and a survey conducted by the Office of Personnel Management since 2002 on federal employees (Federal Employee Viewpoint Survey, FEVS). From this dataset, we build measures of federal procurement offices’ competence and investigate the working place features that underlie this competence. We focus on three different competence components: cooperation among employees, incentives for employees, and their skills.

To quantify the extent to which the government-bureau-level competencies determine procurement outcomes, we construct two proxies of performance based on time delays and cost overruns that represent our main outcome variables. As more complex contracts may be managed by more competent bureaus, a simple correlation between competence and cost or time performance provides paradoxical results. As shown in the figure below, the performance of the agencies that are worst in terms of competence (DVA and DOJ) is superior to that of the two most competent agencies (NASA and NRC) in terms of both time and cost performance. This striking inversion of the relative ranking is a key feature of the economic environment that we analyse and around which we construct our empirical strategy. More competence is associated with more complex contracts, which are intrinsically associated with higher levels of delays and cost overruns. 

Figure 1 Cost and time competence, overall quality, and bureau and agency level

To deal with this endogeneity issue, we construct instruments for bureaus' competencies based on death occurrences of specific types of employees, starting from a third dataset (FedScope) containing characteristics of the public workforce.

None of these three datasets has ever been combined before, and the evidence they reveal is particularly interesting. Three main facts emerge. 

  • First, contractual performance is a persistent characteristic of a bureau. 
  • Second, relevant variation in performance occurs at the bureau level and not at the department level. 
  • Finally, the simple association between our measures from the FEVS and the procurement performance proxies is likely to underestimate the benefits of higher competence in procurement, as indeed more complex contracts are managed by more competent bureaus. 

Our instrumental variable estimates indicate a strong, positive effect of competence on both cost and time performance. The magnitude of these effects is substantially higher than the corresponding OLS estimates. We find that the effect of lifting the level of competence of all bureaus to that of the bureau at the 90th percentile of the competence distribution – which happens to be NASA Glenn Research Center5 – would imply a reduction in cost overrun of $102,619 on average per contract or around $13.5 billion in total across all contracts in the dataset. Moreover, this implies a saving of 54.4 days in the effective execution time, corresponding to 7.2 million days across all the contracts in the dataset. We assess the robustness of these estimates to a broad set of checks involving the econometric model used, sample selection, and measurement of both competencies and performance.

Our study also suggests that heterogeneity in the reported degree of cooperation within the office is a more critical driver of differences in competence across bureaus than the skills of the employees and the incentive system in place. The prominence of cooperation conforms to the view that a successful procurement process requires buyers to appropriately handle and coordinate a multiplicity of complex tasks. Involving offices and collaboration among employees with different skills is essential to choosing the tender and contract design that is most appropriate in each circumstance. Thus, whilst we think that incentives and skills still matter, our study suggests that they may matter the most through their impact on cooperation.

Furthermore, we investigate the extent to which the role played by cooperation can be ascribed to the presence of effective managers. While we lack direct measures to draw definitive answers, in the spirit of the recent work by Jäger (2017), we explore the heterogenous effects obtained through instruments considering the deaths of different subgroups of employees.  We show that the deaths that matter most are those of relatively young and best-paid white-collar employees. Moving along the age and salary dimensions, the estimates change in an intuitive way, with the death of older and less well-paid employees being less consequential in terms of changes in the overall bureau competencies.

The results obtained are qualitatively generalisable and confirm the importance of improving decision-making within procurement organisations. They point at the large potential savings and improvement in the performance of public contracts that could be achieved by investing more resources in increasing the competency of contracting authorities that handle large budgets, even in a developed country with long-established institutions such as the US. In Europe, recent policy initiatives see the introduction of qualification systems for public procurers as a necessary response to the greater discretion granted to them by the 2014 Procurement Directives. Existing certification programmes, however, have mainly targeted individual contracting officers. Our results on the crucial role of cooperation suggest that certification programs would be also useful at the level of the procuring office and should include features such as the organisation of the acquisition process and the prevailing management practices, as is often done for private firms.


Decarolis, F, L Giuffrida, E Iossa, V Mollisi, and G Spagnolo (2018), “Bureaucratic Competence and Procurement Outcomes”, NBER Working Paper 24201.

Saussier, S, and J Tirole (2015), “Strengthening the Efficiency of Public Procurement”, French Council of Economic Analysis, April.

Jäger, S (2017), “How Substitutable Are Workers? Evidence from Worker Deaths”, mimeo.

World Bank (2017), Benchmarking Public Procurement 2017, The World Bank.



Topics:  Competition policy Industrial organisation

Tags:  government, efficiency, procurement, competence

Associate Professor, Università Bocconi

Postdoctoral researcher, University of Tor Vergata

Professor of Economics, University of Rome Tor Vergata; CEPR Research Fellow

Postdoctoral researcher, Free University of Bolzano-Bozen

Professor of Economics SITE - Stockholm School of Economics and University of Rome Tor Vergata, Research Fellow EIEF and CEPR


CEPR Policy Research