Research Interests: Development Economics, Health Economics, Political Economy


“An innovative public-private partnership to target subsidized antimalarials: a study protocol for a cluster randomized controlled trial to evaluate a community intervention in Western Kenya” (with Jeremiah Laktabai, Adriane Lesser, Alyssa Platt, Manoj Mohanan, Diana Menya, Wendy Prudhomme OMeara, Elizabeth L Turner) BMJ Open, March 2017; 7(3):e013972. [article]

“Is Traditional Male Circumcision Effective as HIV Prevention Strategy? Evidence from Lesotho” PLoS ONE, May 2017, 12 (5): e0177076. [article]

“Assessing the independent and combined effects of subsidies for antimalarials and rapid diagnostic testing on fever management decisions in the retail sector: Results from a factorial randomized trial in western Kenya” (with Wendy Prudhomme OMeara, Jeremiah Laktabai, Manoj Mohanan, Adriane Lesser, Alyssa Platt, Elizabeth Turner, Diana Menya) BMJ Global Health, September 2016; 1(2):e000101. [article]

Working Papers

“The Political Economy of Health Epidemics: Evidence from the Ebola Outbreak” [pdf]

Abstract. Health epidemics represent a unique test of governmental accountability. Compared to other disasters, the way the government responds is crucial to stop the contagion and limit the ultimate costs to citizens. However, political motives may distort the allocation of governmental resources. In this paper, I study the response of the Liberian government to the 2014 West Africa Ebola outbreak, and the subsequent effects on citizens’ voting behavior and political perceptions. I combine proprietary data on Ebola cases, novel surveys, and publicly available data on the government’s relief effort and post- outbreak Senatorial election. I first provide evidence of limited governmental response before the arrival of foreign aid, in contrast to improved assistance after aid was provided. I then build a spatio-temporal epidemiological model to estimate the ex-ante optimal allocation of relief effort. I find that the govern- ment misallocated resources towards swing villages affected by the contagion. Voters in turn reacted to the differential response: the incumbent party lost political support in areas hit in the first part of the epidemic, while it did not lose votes in areas hit in the second part and in swing villages. I conclude by discussing the costs to citizens of the political misallocation of resources. The findings provide novel evidence on the presence of strategic electoral motives in the allocation of resources in health epidemics.

“The Local Health Impacts of Natural Resources’ Booms”

Abstract. Do booms in production of natural resources benefit or harm the health of local commu- nities? This paper uses administrative micro-data on local natural resources and health outcomes to study how increases in the production of minerals affect birth outcomes in Brazilian municipalities. I address the endogeneity of production exploiting fluctuations in international commodity prices and the pre-existing spatial variation in mineral endowments. Difference in difference estimates suggest that, following booms in mineral production, local communities experience lower fetal and infant mortality and number of low-weight births. I discuss mechanisms in light of the standard trade-off between the benefits – more wealth from taxes and job opportunities – and costs – corruption or pollution – of natural resources.

“Can individuals beliefs help us understand non-adherence to malaria test results? Evidence from rural Kenya” (with Manoj Mohanan, Wendy Prudhomme OMeara, and Elizabeth L Turner), Economic Research Initiatives at Duke (ERID) Working Paper No. 243. – submitted  [pdf]

Abstract. Malaria testing remains relatively uncommon in endemic countries, individuals often purchase antimalarial drugs without diagnostically-confirmed malaria. More than a quarter of those testing negative still purchase first-line antimalarial therapy (artemisinin-based combination therapies, ACTs). ACT overuse depletes scarce resources available for subsidies, and contributes to parasite resistance. As part of a randomized experiment in Western Kenya that provided a subsidized rapid diagnostic test (RDT) and/or a conditionally-subsidized ACT for febrile individuals with diagnostically-confirmed malaria, we examine whether treatment decisions are associated with beliefs that the illness is malaria. We conceptualize treatment-seeking as two sequential decisions: whether (i) to get tested and (ii) to purchase ACT. First, we study whether prior beliefs are associated with each decision. Second, we analyze how these behaviors are associated with posterior beliefs (one week after the intervention). Prior beliefs are not associated with the decision to get tested, instead RDT price is the driving force. For malaria-negative individuals, prior beliefs do not explain ACT purchasing behavior and those who purchased ACT reported a higher posterior belief than those who did not. In order to improve utilization of information from malaria testing, further research is needed to understand how beliefs are shaped and how actions shape beliefs.

“Effect of diagnosis-dependent subsidies on rational use of artemisinin combination therapy (ACT) for malaria: evidence from a cluster randomized controlled trial in western Kenya” (with Wendy Prudhomme OMeara, Diana Menya, Jeremiah Laktabai, Alyssa Platt, Indrani Saran, Joseph Kipkoech, Manoj Mohanan, Elizabeth L. Turner) – submitted [available upon request]

Abstract. Background: More than half of all publicly-subsidized ACT is dispensed in the retail sector where diagnostic testing is uncommon, leading to overconsumption and poor targeting. We tested an innovative test and treat strategy that targeted ACT subsidies to confirmed malaria cases by coupling community-based diagnostic testing with a diagnosis-dependent ACT discount voucher. Methods: We conducted a cluster-randomized controlled trial in 32 community units (CUs) in western Kenya. Eligible CUs had existing community health worker (CHW) programs and were randomly assigned to control and intervention arms in equal proportions. In intervention areas, CHWs were trained to perform malaria rapid diagnostic tests for any individual >1 year who was experiencing a malaria-like illness. Individuals with a positive test received a voucher for a discount (subsidy) on a quality-assured ACT. Retail medicine outlets were enrolled to redeem vouchers and dispense ACT at a discounted price. We conducted four surveys (baseline, 6, 12 and 18 months) of a random sample of all households with fever in the last 4 weeks to evaluate the predefined individual-level outcomes. Primary outcome was uptake of testing at 12 months. Main secondary outcome was targeting of ACT to individuals with a positive test. Findings: Between July 2015 and May 2017, 7,417 randomly-selected participants from 32 clusters were surveyed. At 12 months, 50.5% in the intervention arm and 43.4% in the control arm had a malaria diagnostic test (Adjusted Risk Ratio=1.20, 95%CI:1.05-1.38). By the 18-month survey, the ARR had increased to 1.25 (95%CI:1.09-1.44, Adjusted Risk Difference=0.11, 95%CI:0.05-0.18). Targeting of ACTs improved substantially (46.0% intervention v. 32.5% control; ARR=1.47, 95%CI: 1.20-1.80). Interpretation: Diagnosis-dependent ACT subsidies and community-based interventions that include the private sector can have an important impact on targeting and rational use of ACTs.

“Development of Touchscreen based methods for elicitation of subjective expectations in India” (with Manoj Mohanan) – submitted [pdf]

Abstract. When making decisions under uncertainty, individuals may form subjective expectations about probabilities of events relevant for their choice. Accurate measurement of subjective expectations is critical for high-quality data needed to analyze individual behavior. This paper reports the development and validity of a new method of eliciting subjective expectations. We developed a touchscreen-based application that combines an animated slider along with dynamic images that change relative sizes based on the probability indicated by the respondent. We compare our method to the more traditional approach of using beans as visual aids. First, we find that respondents have a sound understanding of basic concepts of probability. Second, we test for equality of the distributions elicited with the different methods and find them highly comparable. Our findings suggest that the slider could be a viable elicitation method for empirical researchers who aim to collect data on subjective expectations.

Work in Progress

The Health Impacts and Effective Delivery of Maternal Cash Transfers in Myanmar” (with Erica Field) – Baseline and Midline completed, Endline data collection in Fall 2018

“The Impact of Formal Finance on the Moneylender Market: Evidence from Rural India” (with Isabelle Cohen, Erica Field and Rohini Pande) – Empirical analysis

“Is the Phone Mightier than the Virus? Cell Phone Access and Epidemic Containment Efforts” (with Erica Field and Robert Gonzalez) – Data gathering and analysis

“Eliciting Beliefs and Subjective Expectations in the Context of Chronic Disease, Tamil Nadu, India” (with Manoj Mohanan and Harsha Thirumurthy) – Exploratory pilot completed

Other research (Protocols, Policy Reports and Op-Ed Contributions)

“An innovative public-private partnership to target subsidized antimalarials: a study protocol for a cluster randomized controlled trial to evaluate a community intervention in Western Kenya” (with Jeremiah Laktabai, Adriane Lesser, Alyssa Platt, Manoj Mohanan, Diana Menya, Wendy Prudhomme OMeara, Elizabeth L Turner) BMJ Open, 2017; 7(3):e013972. [article]

“Feasibility Study: Technical options to implement a universal maternal and child cash transfer program in Myanmar”, IGC, UNICEF, Save the Children, 2015 (with Jenny Hilton) [pdf]

“Leveraging Technology-Enabled Banking Agent Distribution Networks”, Fino Fintech Foundation, Interim Baseline Report, CIRM (Center for Insurance and Risk Management), 2014 (with Saurabh Sharma and Rupalee Ruchismita)