Darcio
Genicolo-Martins

Public health procurement is shaped not only by administrative choices but also by judicial decisions that enforce the law on public buyer units. Judicial enforcement is costly for two reasons. First, as mandatory purchases are invariably urgent, judicial enforcement undermines procurement planning. Second, as judicial sanctions for noncompliance are severe, auctioneers have higher incentives to maximize tender success at the expense of higher prices, which we call the “under the gun” effect. Unique data on health litigation and procurement of prescription and nonprescription drugs allow us to estimate the overall enforcement costs and the “under the gun” effect. Judicial enforcement implies (i) higher negotiated prices (from 30.73% to 44.37%), (ii) fewer participant firms (from 28.63% to 32.21%), (iii) fewer bids (from 39.40% to 45.93%), and (iv) a lower probability of success (from 38.56% to 48.66%) in urgent tenders in comparison with ordinary tenders. To estimate the “under the gun” effect, we utilize urgent administrative tenders that are not subject to judicial sanctions. We estimate that judicial sanctions increase prices between 8.83% and 9.97%. Thus, judicial enforcement compels the executive branch to carry out the purchases, which generates high costs to the public budget. These results suggest that judges should consider the social costs associated with the enforcement of court decisions when the judiciary acts as a policymaker.

Insper Knowledge
National Dissertation Award - CAPES

While there are numerous examples of policies that benefit small and medium-sized enterprises (SMEs) worldwide, research offers little direct evidence on the benefits of such policies for the economy. Additionally, assessments of the costs of implementing such policies are practically ignored in the literature. This paper exploits a quasi-experimental variation from a program incentivizing the restriction of public tenders to SMEs in Sao Paulo, Brazil, to estimate this policy’s costs. The way that this institutional change occurred allows me to assess those costs only indirectly. Using detailed data on public procurement and a variation of the standard DiD method (difference-in-differences in reverse), I estimate the pre-intervention effects of the policy shift. I find that before the policy shift, for group 65 (the ‘switched’ group) in comparison with other groups (the ‘always treated’ group): (i) the negotiated prices were lower (between 4.58% and 8.08%); (ii) the number of participants was approximately 22% higher; and (iii) the number of valid bids was approximately 25% higher. These results suggest that the policy of incentivizing the restriction of public tenders to SMEs may severely undermine the quality and efficiency of the public procurement process. Finally, before the policy change, sellers who won tenders for group 65 were more distant from the public buyer units (approximately 4 km). This result may indicate that the policy change has successfully induced more local suppliers to win more bids for this group of items.

National Dissertation Award - CAPES

The systematic loss by firms of public tenders in which they participate for long periods and several times may indicate a cartel. This paper proposes a screening method to detect cartels in public tenders by considering frequent losers. Using data on public procurement in Sao Paulo State, Brazil, from 2009 and 2019, we estimate that frequent losers are associated with 10% higher prices, 32% more participants, and 29% more bids. These results are consistent with the behavior of a cartel (higher prices) that tries to avoid detection by manipulating variables that signal competition (number of players and bids). The proposed method can address two limitations of traditional screening methods: (i) the ability to distinguish between tacit collusion and explicit collusion and (ii) the identification of a possible cartel before the conclusion of public tender processes.

National Dissertation Award - CAPES

A Machine Learning Approach to Analyze and Support Anti-Corruption Policy