I am a Ph.D. candidate in the Department of Agricultural and Resource Economics at the University of California, Davis. My areas of research include development economics, firm productivity, trade, and migration. I also currently work as a consultant with the World Bank’s East Asia and Pacific Chief Economist’s Office and have previously worked with the Poverty and Equity Global Practice.

I am interested in the drivers of private sector development, particularly in developing countries, including how skills, trade, supply chains, and technology adoption affect firm productivity and resilience to shocks. My work applies quasi-experimental strategies and machine learning methods to study these themes across a range of contexts. I am also interested in cross-border and rural-to-urban migration, especially Mexico–US labor flows, and how climate, economic, and policy-related shocks affect migration patterns.

I am on the 2025-26 job market.

Working Papers

Chawla, P. 2026. “Local Human Capital Development and Firm Resilience: Evidence from the 1997 Indonesian Crisis.” Working paper. (Job Market Paper) SSRN Link

Presentations: Japan Economic Policy Association 2025

Summary

Do returns to human capital rise during crises? This paper examines whether human capital accumulation created by Indonesia's INPRES school construction program in the 1970s improved firm resilience in the 1997 Asian Financial Crisis. I find that each additional school per 1,000 children increased real labor productivity and output during the crisis period by 2.7 and 3.3 percent, respectively. INPRES induced changes in the local pre-crisis workforce composition that shifted workers toward more skill-intensive production work. I argue that these skills became particularly valuable during the crisis, when firms faced disruptions. Consistent with this mechanism, INPRES had larger effects in sectors with higher import intensity, where adjustment demands were greater. I show that the local abundance of skilled workers in high-INPRES districts kept wages relatively lower, enabling plants to retain more of them when they were most valuable.

Publications

Barriga-Cabanillas, O., Chawla, P., Redaelli, S. and Yoshida, N. 2025. “Estimating Poverty in Afghanistan Without Consumption Data: An Imputation-Based Approach.” The Journal of Development Studies, 1–22. Link

Earlier version circulated as World Bank Policy Research Working Paper No. 10616.

Summary

This paper uses a machine learning-based survey-to-survey imputation method to estimate poverty in Afghanistan following the Taliban's return to power in August 2021. A model trained on the 2019/20 Expenditure and Labor Force Survey is used to predict household consumption in the 2023 Afghanistan Welfare Monitoring Survey, a phone survey drawn from the same sampling frame. The results show that 48.3 percent of the population was poor as of April-June 2023, a 4 percentage point decline since the same months in 2020. This decline was driven by falling rural poverty, while urban poverty remained unchanged.

Research in Progress

“Predicting Mexico-to-US Migration with Machine Learning for Counterfactual Analysis,” with J. Edward Taylor

Presentations: AAEA & WAEA Joint Annual Meeting 2025 (Best Poster Award Recipient) Poster

Summary

Reliable tools to predict migration are increasingly important amid rising climate and economic risks, and demographic shifts. Tree-based machine learning models can uncover complex, nonlinear relationships that conventional models often miss and can be used to simulate responses to shocks. Migration data are costly to collect, so models must perform well with readily available data. We first train a LightGBM model on an ideal dataset, a panel tracking the employment locations of 10,739 individuals from 1980 to 2007, and achieve high predictive accuracy. Using this as a benchmark, we then train a model on just four years of data without migration histories. By adding public weather data, this restricted model approaches benchmark performance (within 0.1 F1 score). Counterfactual shocks show that a 10% rise in temperature reduces migration by 13% the following year, a 10% increase in age lowers it by 17%, and a 10% drop in income by 18%.

“Firm Networks, Risk Sharing and Resilience to Shocks Among Small Firms in Tanzania,” with Daniel Putman and Jess Rudder

Presentations: World Bank-LSMS Conference "Better Data for Better Jobs and Lives" (2025)

Summary

We examine the role of formal and informal networks among small firms in helping them cope with shocks. Using novel survey data that we collected in rural Tanzania, we estimate complete firm networks and analyze how network characteristics, such as centrality and clustering, shape firms’ exposure to shocks and their responses, including access to credit, performance, productivity, and entry and exit.

“Is Technology Inappropriate for Developing East Asia?,” with Francesca de Nicola, Aaditya Mattoo and Jonathan Timmis (World Bank East Asia Pacific)

Summary

We test the "inappropriate technology" hypothesis in Vietnam by examining whether the productivity premium from imported technologies adopted by manufacturing firms is lower when those technologies are not well suited to local conditions, exploiting tariff variation as an instrument for technology adoption.

“Financial Literacy and Small Firm Performance in Uganda”, with Ester Agasha, Andrew Hobbs, Travis Lybbert, Nathalie Nyanga, and Bruce Wydick

“Local Economic Impacts of Cash Transfers to Refugees and Asylum Seekers in Mexico, Mauritania, and Moldova,” with Justin Kagin and J. Edward Taylor