Essays on the economics of public procurement

Public procurement is the acquisition of goods, services, and works from private contractors by the government and related entities. It represents a strategic policy instrument that can have critical externalities on society. Cumbersome rules, abuses of discretion, excessive prices, and post-award quality degradation arguably pose some of the main challenges to procurement effectiveness. This thesis consists of three related chapters in which I explore these issues using advanced empirical techniques. The first chapter – which is a joint work with Decio Coviello (HEC Montréal), Andrea Guglielmo (Wayfair), and Giancarlo Spagnolo (SITE-Stockholm School of Economics, University of Rome Tor Vergata, and EIEF) – addresses rules vs discretion in public procurement and the corresponding trade-off between accountability and performance. Rules constraining bureaucratic discretion may limit the misuse of public funds, but also hinder government performance. Using data from Italian public works, we study whether and which procuring administrations manipulate contracts’ value to avoid crossing regulatory thresholds that limit discretion, and how this impacts procurement outcomes. We use bunching estimators to document substantial manipulation just below these thresholds, performed by administrations led by appointed officials but not by elected ones. We estimate the effects of manipulation, finding that it increases the use of discretionary procedures (restricted auctions) thereby reducing the number of bidders, with mixed effects on rebates and positive effects on other outcomes: it reduces works’ length, delays and cost overruns; and it increases repeated awards to financially less risky suppliers. We cross-validate our estimates using a reform that shifted the discretion threshold, finding that it reduced manipulation and the use of discretion, worsening procurement outcomes, especially for administrations with appointed officials. This evidence is consistent with appointed administrators circumventing excessively strict rules mainly to improve procurement outcomes, and with electoral incentives preventing other administrators from doing so. A simple procurement model where administrations may choose to manipulate at different cost provides structure to this interpretation. iii The second chapter – which is a joint work with Giancarlo Spagnolo (SITE-Stockholm School of Economics, University of Rome Tor Vergata, and EIEF) – explores one organizational lever that can be used to improve the efficiency of public procurement, namely centralization. An influential paper by Bandiera, Prat and Valletti (2009) (BPV) exploits the introduction of a central purchasing agency in Italy as a quasi-experiment to identify the amount and sources of public waste. Among other things, it estimates that public bodies purchasing through the central agency save on average 28% of price. We argue that centralized prices may also exert important indirect effects on non-centralized ones. In the subsample of contracts in BPV’s data relating to goods for which it is possible to observe prices both before and after the central agency entered the market, we find that the existence of a centralized price induces a 19% reduction in the price of non-centralized purchases. Accounting for this, the estimated direct savings from centralized purchases increase to 47%. These indirect effects appear driven by informational externalities – rather than improved outside option – on less competent public buyers purchasing more complex goods. The third chapter studies the effect of political competence on state capacity in the context of public procurement in Italian municipalities. Using a regression discontinuity design with close elections, I find that mayoral competence causes an increase in the use of discretion in public procurement and greater efficiency. In detail, it reduces administrative delays, the number of renegotiations, and the timing of contract execution. Furthermore, I explore whether mayoral competence impacts the composition of the bureaucracy using an event study analysis and find that it leads to an increase in the number of discretionally appointed managers. The joint interpretation of these results suggests that the hiring of new managers might be a channel through which the competent mayor improves the administration’s performance.