Research

Working Papers

• Metric Visibility and Uneven Social Performance

When performance is multidimensional but unevenly observable, monitors concentrate on what can be verified at low cost. I study this mechanism through foreign institutional ownership and social performance. Using a global panel of 12,351 firms across 58 countries, I find that higher foreign ownership widens the gap between visible and opaque social outcomes: diversity, the most standardized metric, remains stable while worker safety, customer responsibility, and supply-chain ethics deteriorate. Instrumental variable estimates confirm the causal direction and rule out selection. I formalize this divergence with a Visibility Gap measure and show that domestic ownership narrows it, consistent with local informational advantages. The distortion intensifies under investor distraction and is concentrated among short-term and distant investors; it attenuates with stronger legal institutions and concentrated ownership. These patterns are consistent with information frictions, not preferences, as the driver. Incident-level evidence from RepRisk confirms the divergence reflects real behavioral change, not reporting artifacts.

• Socially Responsible Investing and Multinationals' Environmental Harm: Evidence From Global Remote Sensing Data

(With Virginia Gianinazzi, Victoire Girard and Melissa Prado)

This paper studies how Socially Responsible Investment capital shapes the environmental footprint of multinational firms. We exploit the inverse relationship between local pollution and high-frequency satellite measures of vegetation health captured by the normalized difference vegetation index (NDVI). Merging NDVI with SRI ownership for 52,806 facilities of 911 multinationals in 124 countries from 2006 to 2020 allows us to trace changes in SRI exposure within and across facilities. Higher SRI ownership is associated with improved nearby vegetation, and using mergers as a plausibly exogenous source of variation in SRI ownership corroborates these findings. Analysis of global production networks reveals a marked asymmetry: improvements near facilities in OECD countries coincide with deterioration near the same firms’ non-OECD sites, consistent with pollution offshoring. This asymmetry strengthens with more active investor oversight, indicating that engagement alone cannot counter weak domestic regulation or limited global monitoring.

• Procuring Sustainability: The Impact of Green Government Contracts [New version coming soon]
  
  This paper examines the impact of integrating sustainability criteria in federal procurement on firm behavior and environmental performance. Using comprehensive U.S. government contract and firm-level emissions data, I find that firms awarded green contracts significantly reduce their GHG emission intensity, though these reductions are not consistently greater than those from standard contracts. Despite the higher costs of green contracts, the evidence suggests that their environmental benefits may not always justify the cost premium. The effect on emissions is more pronounced for firms that are brown and financially constrained. I address endogeneity and selection biases by applying a matching technique, a Bartik instrument and exploiting an exogenous increase in public spending following census revisions. Green procurement also fosters innovation, particularly in green technologies, and generates positive but largely unilateral spillover effects in supply chains. Overall, these findings indicate that while government procurement can promote sustainability, the environmental benefits of green contracts may not justify their higher costs.
  
  

• Extreme Events and ESG Commitment: Evidence from Mutual Funds

I study how personal experiences shape mutual fund managers’ ESG commitment. Using a difference-in-differences design, I show that managers located near extreme disaster areas increase the ESG score of their portfolios in the following quarters, with stronger effects for environmental scores. The adjustment comes mainly from divestment of low-ESG stocks rather than acquisition of new high-ESG assets. Social connectedness to affected counties, prior beliefs about climate change, and periods of high media attention amplify the effect. Returns, flows and volatility remain unchanged, ruling out performance or risk aversion explanations. The findings support the salience hypothesis and show that personal experience is an independent channel through which ESG integration occurs.