Job market paper
The Misallocation Costs of Bid Rigging: Evidence from Montreal Asphalt Procurement Auctions
This paper examines the misallocation costs of collusion in public procurement auctions using cartel-period data from 2007–2009 on the Montreal asphalt cartel. I estimate firm-specific cost distributions nonparametrically from the post-cartel period and simulate competitive counterfactual outcomes for each cartel-period auction. First, I find that the cartel inflated average procurement prices by 24 percent. Second, the inefficient allocation of contracts in the cartel period increased total production costs by 14.1 percent. Third, decomposing the price overcharge, 58.4 percent is attributable to misallocation of contracts toward higher-cost firms and the remaining 41.6 percent to higher collusive markups. These findings suggest that in markets with substantial dispersion in firm productivity, inefficient allocation may be the primary channel through which bid-rigging harms public buyers.
Heterogeneity, Productivity and Reallocation: Evidence from the U.S. Cement Industry
This paper develops a structural model of spatial competition to estimate plant-level total factor productivity in the U.S. Portland cement industry, where market power makes standard production function approaches inconsistent. Using plant-level data from 1995 to 2007, I implement a simulation-based estimation framework that allows for both observable and unobservable sources of productivity heterogeneity. I find that there is notable productivity dispersion: plants in the top decile are on average 25% more productive than those in the bottom decile. Decomposing the sources of this dispersion, I find that observable factors, particularly kiln technology, account for approximately 95% of the variance in total factor productivity, leaving little role for unobserved plant-specific differences. A counterfactual carbon tax analysis exploits this technology heterogeneity to show that environmental regulation reallocates output toward cleaner plants, while disproportionately burdening carbon-intensive producers.
Should We Share Our Cookies? The Effect of Data on Match Quality in Online Advertising
With Mehdi Akhbari
User data plays a central role in online advertising by enabling platforms to match ads to users more effectively. This paper quantifies how access to additional data affects match quality, measured by user click-through rates, and evaluates how data-driven mergers between platforms with complementary datasets influence market outcomes. Mergers improve match quality through two distinct mechanisms: (i) a precision effect, whereby a larger number of observations improves the accuracy of predicted click probabilities, and (ii) a targeting effect, whereby broader tracking of user behavior enhances the platform’s ability to tailor ads to user preferences. The analysis utilizes the conditionally random allocation of ads in the data provided by Iran’s largest online advertising platform and employs causal forest, a novel causal machine learning method, to estimate heterogeneous user click rates based on extensive browsing and ad exposure data from more than 1.6 million internet users. Simulations of user and platform behavior reveal that hypothetical mergers could boost match quality by 14–25%. In addition, this research finds that the majority of gains arise from an increase in the number of observations while expanding the set of observable user characteristics provides comparatively smaller gains.
Heterogeneous Effects of Community-Based Interventions on Mental Health: Evidence from Catalonia
This paper evaluates the impact of an intersectoral public health policy implemented in Catalonia, aimed at improving population well-being through enhanced community engagement and the promotion of social activities. Leveraging regional variation in the intensity of policy implementation over time, I examine both the overall and heterogeneous relationship between policy exposure and a validated measure of mental health. The results find no significant association at the population level. However, I find that higher policy intensity is associated with improved mental health outcomes among immigrants and, to a lesser extent, young adults. The results have policy implications that can help increase the efficacy of similar interventions.