Working Papers¶
Bitter Pills to Swallow: the Enforcement Costs of Health Litigation¶
with Paulo Furquim de Azevedo (updated April, 2026) | PDF
Abstract
Judicial enforcement is essential for the rule of law, yet the form of enforcement can generate economic inefficiency. We study this tension using bid-level data on court-mandated pharmaceutical procurement in Sao Paulo, Brazil (2009--2019). With item, time, and buyer fixed effects, court orders raise prices by 5.4% and reduce competition by 5.4%. Exploiting an institutional distinction between urgent purchases that share planning constraints but differ in penalty exposure, we isolate the "under the gun" effect: judicial sanctions alone raise costs by 23--30%. This premium operates entirely through demand fragmentation---sanctions prevent order aggregation---not through officials accepting higher unit prices. The finding speaks to a general asymmetry in accountability design: when non-completion is punished more severely than overpayment, compliance crowds out cost-efficiency.
Media coverage
The Price of Protecting Small Firms: Evidence from Public Procurement¶
(updated April, 2026) | PDF
Abstract
Governments worldwide restrict public tenders to small and medium-sized enterprises (SMEs), yet the fiscal cost of this policy is virtually unknown. I exploit a natural experiment in Sao Paulo, Brazil, where the policy operates through opt-out costs: public buyers that wish to conduct open tenders must justify their choice to audit courts, effectively penalizing the efficient procurement decision. Until March 2018, medical supplies were uniquely exempt from these costs. When a legal reinterpretation---unrelated to procurement outcomes---ended this exemption, it created a clean difference-in-differences in reverse (DiDiR) design. Open tenders yielded prices 7--13% lower (12--13% nominal, 7--10% in real terms) and attracted 19% more firms, but a Gelbach decomposition reveals that increased participation explains only 6% of the price reduction---most of the savings come from fiercer competition among bidders. The fiscal cost for one product group alone is R$50--85 million over 18 months, depending on the inflation adjustment. These costs are concentrated among standardized goods; for specialized items, restrictions appear less distortionary. The one clear benefit: SME-only tenders successfully promote local sourcing.
Media coverage
Screening for Bid Rigging with Frequent Losers¶
with Paulo Furquim de Azevedo (updated April, 2026) | PDF
Abstract
Minimum-bidder rules in public procurement give cartels a reason to deploy cover bidders---firms that show up repeatedly with no intention of winning. We exploit this behavioral footprint to build a participation-based screen: frequent losers (FL), firms that never win yet bid abnormally often. The screen requires only win/loss records, making it deployable in settings where bid microdata are unavailable. In Sao Paulo's electronic procurement (4.5 million tender-items, 2009--2019), it achieves AUC = 0.94 against competition-authority co-participation (0.54 within participation-count bands, confirming that volume contributes substantially to the unconditional figure), complements bid-level tools (correlation 0.06), and flags environments with 3.6--7.7% higher conditional prices. The price association is concentrated in competitive markets and where cover bidding is voluntary rather than forced by the minimum-bidder constraint---suggesting strategic deployment, not mere rule compliance. We propose a three-stage enforcement pathway (screen, triage, investigate) that allocates investigative resources toward the procurement environments where the association is strongest and oversight weakest.