Beneath the Surface: Investigating Anti-Competitive Signals in Brazilian Public Procurement

Abstract

This paper examines **anti-competitive behavior in public procurement** using a **Regression Discontinuity Design (RDD)** applied to **3.8 million sealed-bid auctions** conducted in **São Paulo, Brazil**, between **2009 and 2019**. Following *Kawai et al. (2023)*, the analysis exploits **near-tie auctions**, comparing firms that **narrowly win and narrowly lose** contracts. Under competitive conditions, these firms should be observationally similar; thus, any **discontinuous differences at the winning threshold** indicate non-competitive forces. The study focuses on two key outcomes: **incumbency**, defined as having won the previous auction in the same market, and **bid timing**, measured by whether the firm submitted the **final bid**. The results reveal significant discontinuities at the cutoff. Firms that narrowly win are **1.8 percentage points more likely to be incumbents** and **6.4 percentage points more likely to place the last bid** than firms that narrowly lose. These effects are statistically significant at the 1% level and robust to **market and year fixed effects**. The findings are consistent with **strategic bidding, incumbency advantages, and coordination mechanisms** such as bid rotation or bid leakage. Importantly, these patterns arise **even without direct evidence of corruption**, highlighting that anti-competitive behavior can persist through subtler channels. Overall, the paper shows that in **developing-country procurement systems**, anti-competitive practices can substantially undermine efficiency and competition. The proposed RDD approach provides a **scalable and practical screening tool** for procurement oversight and competition authorities.