Research

Peer-Reviewed Publications

• “A Taste for Taxes: Minimizing Distortions Using Political Preferences.” 2019. Journal of Public Economics, vol. 180, p. 104055. With Andrea Robbett and Emiliano Huet-Vaughn. doi.org/10.1016/j.jpubeco.2019.104055
  Show Abstract

  We conduct an experiment with online workers to assess whether the distortionary effect of a tax is sensitive to the ideological match between taxpayer and tax expenditures. We find that, among self-identified political moderates, the labor supply elasticity with respect to the net of tax wage is significantly smaller when individuals pay taxes to a favored government agency as compared to an unfavored one. While the tax has a significant distortionary effect in the latter case, with a point estimate for the labor supply elasticity of approximately 0.77, the elasticity point estimate is close to zero when taxes go to a favored agency. There is also an increase in total output for the matched population among moderates. There is no evidence that these effects hold for self-identified liberals or conservatives.

Research in Progress

• The Pollution–Productivity Curve: Non-linear Effects and Adaptation in High-pollution Environments (with Faraz Usmani) Draft available upon request
  Show Abstract

  Air pollution harms labor productivity, yet little is known about whether workers adapt to chronic exposure. We address this question using 14 years of individual-level performance data from India's premier cricket league, a setting characterized by some of the highest levels of particulate matter air pollution (PM2.5) and whose schedule and geography result in variation in both acute and chronic exposure histories. We pair these granular performance metrics with an India-specific machine learning data product that incorporates remotely sensed and ground-monitor measures of PM2.5. Our results reveal that both chronic and acute exposure to pollution are costly, but in different ways. A 10 microgram-per-cubic-meter increase in same-day PM2.5 concentration (half a standard deviation in our sample) reduces productivity by about 1 percent, with effects concentrated at the highest pollution levels, implying a nonlinear dose-response. The dose-response also exhibits surprising heterogeneity: same-day shocks harm those chronically exposed at the highest levels approximately 40 percent less than those with median exposure histories, indicating adaptation. Nevertheless, chronic exposure itself results in performance declines that, though smaller in magnitude than the declines resulting from same-day shocks, far outweigh any protective effect from adaptation. Our findings suggest that standard estimates from low-pollution environments do not capture the dynamics between acute and chronic exposure in high-pollution settings.

  Invited Seminars: Indian Statistical Institute, Delhi; University of Massachusetts, Amherst
  Conference Presentations: Association of Environmental and Resource Economists (AERE) 2025 Annual Conference; Occasional Workshop in Environmental & Resource Economics at UC Santa Barbara; University of Colorado Boulder Environmental & Resource Economics Workshop

• Got Goat? The Effects of a Digital Inventory Tool on Livestock Market Outcomes in Rural Nepal (with Travis Lybbert, Conner Mullally, Nick Magnan)
  Conference Presentations: Agricultural and Applied Economics Association (AAEA) 2025 Annual Meeting
  Funding: J-PAL’s Digital Agricultural Innovations and Services Initiative (DAISI)
  Pre-Analysis Plan: AEA RCT registry

• Environmental Justice and the Costs of Accessing Clean Air: Experimental Evidence from California (with Shotaro Nakamura and Collin Weigel)
  Conference Presentations: University of Chicago Advances with Field Experiments (AFE) 2025 Conference; UC Berkeley/Davis Giannini Foundation of Agricultural and Resource Economics Student Conference
  Funding: California Air Resources Board
  Pre-Analysis Plan: AEA RCT registry

• Global Spillovers in Agricultural Technology Development (with Ashish Shenoy)

Policy Publications

• “Does An Innovation’s Reach Reveal Anything About Its Impact? Under The Right Conditions: Possibly.” 2025. Rome, Italy: Standing Panel on Impact Assessment. With Travis Lybbert and Madeleine Walker. Link to PDF
  Show Abstract

  Impact—the extent to which a specific innovation causally affects a target outcome—is an ideal measure of a program’s success. However, impact evaluations are time- and resource-intensive, making them unavailable in many circumstances. By contrast, measuring the reach of an innovation, i.e., the number of adopters, is more straightforward. As a result, measures of reach are more prevalent than causal impact estimates. This note discusses the relationship between reach and impact. It is motivated by the mandate of the Standing Panel on Impact Assessment (SPIA) of the CGIAR and draws on SPIA country studies, which explicitly focus on measuring the reach of CGIAR innovations, to illustrate the possibilities and pitfalls of using reach to infer impact. Whereas the reach of an innovation in a given population may reveal something about its impact, the conditions under which reach may be a useful proxy for impact often deviate sharply from on-the-ground realities. In most settings, understanding the total benefits of an innovation requires reliable evidence of both reach and impact.

• “Evaluation of the Liberia Compact’s Mt. Coffee Hydropower Plant Rehabilitation and Capacity Building and Sector Reform: Findings from the Final Round.” 2024. Washington, DC: Mathematica. With Candace Miller, et al. Link to PDF
  Show Abstract

  Customer connections grew quickly as the Mount Coffee Hydropower Plant and donor-funded distribution projects were completed. However, plant sustainability is jeopardized by inadequate funding for operations and maintenance. Despite challenges, the management services contractor, which the firm hired to support the utility, increased connections, reduced outages, and improved management practices. Power theft remains high and threatens sustainability. The utility requires investments in meters and the low voltage network to meet customer demand. The Liberia Electricity Regulatory Commission has developed a regulatory framework for electricity generation, transmission, distribution, and sales. However, the regulator approved a tariff that is too low to cover costs, private operators are unlicensed, and its independence is threatened without adequate income.

• “Liberia Energy Project: Findings from the Final Evaluation of the Pipeline Sub-Activity.” 2024. Washington, DC: Mathematica. With Poonam Ravindranath, Paolo Abarcar, Cullen Seaton, and Candace Miller. Link to PDF
  Show Abstract

  The pipeline was successfully completed and has been transmitting raw water from the Mt. Coffee Hydropower Plant to the Liberia Water and Sewer Corporation’s (LWSC) water treatment plant since December 2020. Overall, the pipeline did not improve the quantity or quality of the treated water supply. LWSC staff reported reduced electricity use and costs given the gravity-fed design replaced an expensive system that pumped the water. LWSC is not conducting pipeline maintenance, risking long-term sustainability. LWSC staff said there was a shortage of spare parts and insufficient management support for maintenance.

• “Maternal Health Care Quality Improvement in Rajasthan, India: Insights from a Development Impact Bond Verification Agent.” 2021. Washington, DC: Mathematica. With So O’Neil, Divya Vohra, Shveta Kalyanwala, and Dana Rotz. Link to PDF
  Show Abstract

  At the end of the Development Impact Bond (DIB), modeling using the Lives Saved Tool (LiST) predicts that facilities meeting DIB quality standards will avert an estimated 13,449 maternal and neonatal deaths across 405 private healthcare facilities in Rajasthan by 2023. With an estimated cost of $21,800 per private facility needed to improve quality, further studies on potential returns could provide more insights into whether a case could be made to implement and sustain related quality improvement efforts. The Utkrisht DIB has demonstrated that broader context changes and learning related to partner needs can spur innovation in the approach to verification and clarify thinking around whether and how verification processes can answer broader questions about the DIB’s contribution to social change.

• “Jordan Refugee Livelihoods Development Impact Bond Evaluation Framework.” 2021. Washington, DC: Mathematica. With Evan Borkum, Paolo Abarcar, and Laura Meyer. Link to PDF
  Show Abstract

  The IKEA Foundation has contracted with Mathematica to conduct an independent evaluation of the program funded by the DIB. Mathematica’s evaluation seeks to both measure the metrics to determine payments to investors and generate broader learning about the program’s impacts on participants' economic and social wellbeing, to support future adaptation and scale-up. This evaluation framework report describes Mathematica’s proposed design for the evaluation. Our rigorous mixed-methods evaluation comprises three components: (1) income-generating activity (IGA) validation, which will seek to estimate the percent of microenterprise grant recipients who are actively engaged in IGAs about 10 months after grants are disbursed; (2) an impact evaluation, which will assess the impacts of the program on household consumption and other outcomes related to social and economic wellbeing, using a matched design to compare the outcomes of first cohort participants 24 months after grant disbursements with those of matched third cohort participants shortly after program enrollment; and (3) a process evaluation, which will seek to summarize the programmatic context, explore participants’ experiences with and perceptions of the program, and identify facilitators and barriers to achieving the outcomes envisaged in the program logic.

• “Maternal Health Care Quality and Outcomes Under the Utkrisht Impact Bond: Midline Findings and Insights.” 2020. Washington, DC: Mathematica. With So O’Neil, Divya Vohra, Dana Rotz, and Shveta Kalyanwala. Link to PDF
  Show Abstract

  Quality improvement represents a journey that a point-in-time verification design might not capture. Outcome payments under the Utkrisht DIB exceed implementation costs for facilities that have met quality improvement standards. At the midpoint of the Utkrisht DIB, it is too soon to tell whether improved quality at participating SHCOs has led to meaningful improvements in key health status outcomes. Verifying DIB outcomes requires that the outcomes triggering repayments are clearly defined, appropriately measured, and agreed upon by all parties.

• “Evaluation of the Liberia Power Compact’s Mt. Coffee Hydropower Plant Rehabilitation and Capacity Building and Sector Reform: Baseline and Interim Findings.” 2020. Washington, DC: Mathematica. With Candace Miller, et al. Link to PDF
  Show Abstract

  The rehabilitation of MCHPP was successful. The hydropower plant is Liberia’s largest electricity asset. However, ongoing operations and maintenance is underfunded increasing the risk of turbine or plant failure and possible consequences including performance losses, extended outages, higher rehabilitation costs and potential emergency situations such as the loss of life or property. A careful analysis indicates that ESBI has been successful in diagnosing and beginning to solve critical problems. Although performance has not met stakeholders’ expectations, ESBI assumed responsibility when LEC was in a grave financial situation. The utility requires increased funding for operations and capital expenses, a systematic response to theft and corruption, and support from the Government of Liberia and donors to implement the reform needed to sustain the utility. LERC has made progress in establishing the regulatory commission in 2019 however it lacks resources beyond January 2021. Commissioners believe that donor financing is essential to LERC being an independent, transparent, accountable, and sustainable agency. Liberians frequently say, “Electricity Is Life” indicating how much they value and demand “LEC current”. Liberians report that they prefer to pay for a legal connection. However, they will collect illegally if they feel there is no other option. And while electricity improves quality of life and feelings of security, insufficient education and low-quality infrastructure present important dangers and safety risks, including fires and electrocution.