Research

I am an applied microeconomist working on topics related to risk and investment. By focusing on both classical and behavioral components of decision-making, my research helps make sense of observed behavior and guide the design of policies seeking to foster resilience in the face of economic and environmental uncertainty. I believe that these questions are as relevant for my home community in Maine as for the communities in Mozambique and Nicaragua, where I have spent much of the previous ten years studying the effects of drought on smallholding farmers.

My extensive fieldwork in Mozambique, Guatemala, and Nicaragua has also left me with an abiding interest in the formation of agricultural value chains and the role of information in determining prices and distributing gains from production.

Full Research Statement

Research Interests
· Risk, Insurance, and Investment
· Behavioral Influences of Economic Behavior
· Agricultural Value Chains

Working Papers and Ongoing Research


“The Farmer and The Fates: Locus of Control and Investment in Rainfed Agriculture”.

Abstract
Economists and policy makers alike often find themselves struggling to explain low levels of adoption for seemingly profitable agricultural technologies. This easily extends to technologies that stabilize yields in the face of weather and pest pressure. In this paper, I take a behavioral approach to explaining adoption and investment rooted in the psychological notion of locus of control, which captures the extent to which a decision-maker believes that her actions affect outcomes relative to forces outside of her control. I consider how locus of control affects neoclassical models of decision-making and test various pathways using primary data from smallholding farmers in Eastern Africa. In addition to a standard locus of control instrument from psychology, I develop a measure of locus of control specific to maize production. I show that farmers with more external locus of control are significantly less likely to adopt improved maize varieties. Using data on subjective weather expectations and risk attitudes, I present suggestive evidence that locus of control is not reflecting differences in preferences or subjective probabilities, but rather affecting beliefs about the production function itself.

“Salient Risks and Contract Quality: Estimating Demand for Index Insurance in Mozambique and Tanzania.” With Laura Paul, Jon Einar Flatnes, Stephen Boucher, and Michael Carter.

Abstract
In this paper, we study perceptions of index insurance quality and demand for an input-replacement contract marketed to maize farmers in Mozambique and Tanzania. The product, which bundled index insurance with improved maize seed, was marketed beginning in 2016 as part of a randomized controlled trial (RCT) in partnership with local maize seed and insurance companies. The unique combination of market purchases, survey responses, and experimental variation present in the data allow us to gain insight into how perceptions are formed by farmer experiences and translate into demand for risk-reducing technologies. Nearly half of farmers in areas where the bundled product was offered purchased some quantity of insurance. Similar to Cole and Giné (2013), we offered households randomly chosen discounts to create experimental price variation in order to estimate the elasticity of demand. We find that demand for both the bundled product and the seed, by itself, are highly sensitive to price (nearly unit elastic). Comparison of the two demand curves, however, indicates that farmers value the insurance component more than commensurate with its cost. We generate further insight using survey-based willingness-to-pay experiments to extend our analysis to non-purchasing households, which yields similar results.

Poster

“Bundling Stress Tolerant Seeds and Insurance for Resilient Small-scale Agriculture: Impacts and the Challenge of Learning about Technologies with Stochastic Benefits”. With S. Boucher, M.R. Carter, T. Lybbert,  J.E. Flatnes, P. Marenya, and L. Paul

Abstract
Risk has long been hypothesized to inhibit small farm investment in new technologies. Recent evidence indicates that index insurance and stress tolerant seeds can separately boost small farm investment. In this study, we explore whether the complementarities between these risk management technologies can be harnessed to underwrite a resilient, high productivity small farm sector. Utilizing a multi-year, randomized control trial that spanned two countries and also exploits natural variation in weather shocks, we find that drought tolerant maize seeds mitigate the impact of the single peril of mid-season drought. We also find that shocks have long-lasting effects, depressing future investment and productivity for the control group. Index insurance, which covers the full spectrum of shocks, makes small farms resilient as they immediately bounce back from shocks. Importantly, we find that experiential learning is key. Farmers who experienced shocks intensify their use of the technologies in the future, exhibiting what we call resilience-plus. Those who did not experience shocks, begin to move away immediately from the technologies and their additional costs. Together these findings put a premium on the creation of policies that allow small-scale farmers to learn about and trust technologies that offer stochastic benefits.

“Determinants of Participation and Impact Heterogeneity in Nicaragua’s Rural Business Development Program.” With Michael R. Carter and Stephen Boucher.

Abstract
As evidence mounts for heterogeneous treatment effects in entrepreneurial interventions, program implementers look for ways to tilt the balance in favor of cost effectiveness. One such strategy is targeting on observable characteristics believed to be associated with positive program benefits. In this paper we show that such a strategy carries significant risks and may exclude households who stand to gain significant benefit from the intervention. The risks of sharp targeting are especially salient for multifaceted entrepreneurial programs, which benefit participants through multiple channels.