Research

Publications


Book chapters

​


Working Papers


Work in Progress

Targeting subsidies through price menus: Menu design and evidence from clean fuels (with KP Asante, M. Daouda, D. Jack, F. Malagutti, P. Oliva, I. Seidu and A. Sulemana) ​

Abstract: Targeting subsidies to maximize social objectives often requires screening potential recipients on unobservable characteristics. We consider the case where a social planner would like to subsidize a good that produces positive environmental benefits, while keeping expenditures low by targeting the subsidy to more price sensitive individuals. Our empirical setting is clean fuels (LPG) for household cooking in Ghana, where the use of biomass for energy remains common and generates substantial health and environmental costs. Households purchase LPG in cylinders of different sizes, which allows for size-differentiated pricing, but not limits on the frequency of purchase. We start by adapting the theory of second degree price discrimination to a setting where it is not feasible to impose limits on the frequency of purchase, but households face heterogeneous saving costs and transaction costs.  We show that a separating equilibrium with non-linear pricing can maximize a social planner's objective function in this setting.  Next, we demonstrate implementation through a two stage design that first gathers the empirical inputs for a sufficient statistic based approach to menu design, and then tests the optimal price menu against counterfactual linear pricing. We find that relative to the counterfactual, non-linear pricing achieves the same level of LPG demand with 30% less in public expenditure.


Collateral accounts: Debt recovery through prepaid electricity metering (with J. Cross and A. Jensen)

Abstract: Much of the literature on tax compliance has focused on evasion. Delinquency is an equally pervasive problem in developing countries, yet less is known about the effectiveness or distributional impacts of efforts to address delinquency. We study a program in Cape Town, South Africa that enrolls household with delinquent municipal tax and service accounts in an aggressive debt recovery program that leverages linkages across municipal accounts. Specifically, households in Cape Town receive electricity through prepaid meters, that require advance purchase of electricity, thereby eliminating arrears in electricity payments. When households reach high levels of billing debts, these debts are recovered through electricity purchases. We exploit program rules surrounding eligibility and targeting to recover a price elasticity of demand for electricity, cross-price elasticity of water demand, and the overall revenue impacts of the program, inclusive of demand responses. We also assess the welfare consequences of debt recovery, both for targeted households and for the population as a whole. Our findings highlight the importance of digitization and linking of government databases for state capacity.


Paying for power: Prepaid electricity and the spending patterns of the poor (with K. McDermott, J. Romine and A. Sautmann)​

Abstract: Compared to high-income households, low-income households make very small, high-frequency purchases of a wide variety of regularly consumed goods, such as fuel and staple supplies. Proposed explanations are often good- and setting- specific or require income-dependent preferences, for example that poor households have higher storage costs, low or no transaction costs, or greater challenges resisting temptation. Here, we ask whether the observed patterns by income can be generated by an intertemporal optimization model with unified preferences, savings and borrowing constraints, and transaction costs. We use detailed information on pre-paid electricity purchases by South African households to calibrate a two-good model of spending and consumption with one ``illiquid'' good that incurs transaction costs to stock up, and validate it with experimental data on a surprise electricity transfer. We then use the calibrated model to examine purchasing patterns, consumption distortions, and welfare losses as a function of income.

​

Dodging Day Zero: Drought, Adaptation, and Inequality in Cape Town (with C. Cole, K. Meng and M. Visser)

Abstract: We illustrate the potential for climate shocks to destablize the revenue model for public utilities, using the case of an extreme drought in Cape Town, South Africa. To curtail demand in the face of impending "Day Zero" -- the day taps would run dry -- the utility raised prices. We use detailed municipal billing data to show a dramatic decline in residential water use, driven by richer residents. Most strikingly, by the peak of the drought, top deciles of the city's (highly unequal) wealth distribution were using less municipal water than their more disadvantaged neighbors, a pattern that persisted for two years following the return to normal water prices. We then show that some of this reduction can be explained by differential substitution to groundwater by richer households, corroborated by groundwater drilling applications, groundwater levels, and satellite-based measures of vegetative water content. While the shift toward groundwater helped alleviate the water crisis in the short run, it also undermined the effectiveness of increasing block tariffs for both revenue and redistribution goals. In addition, the induced groundwater extraction may be unsustainable in the long run.


Health insurance as tool to smooth seasonal incomes: A randomized controlled trial (with D. Daouda, G. Fink, D. Schwarzkopf and R. Strobl)

Abstract: The majority of rural households in developing countries continue to depend on small-scale agriculture as their primary source of income. In most settings, incomes are generated only once or twice a year, leading to significant liquidity constraints during the “hungry” periods between harvests. While health insurance programs have historically focused on providing financial protection and access to care for all, we hypothesize that well-designed health insurance can also help households smooth income and consumption in rural agricultural settings: if insurance premiums can be paid upfront and households use covered health services on a regular basis, insurance can absorb cash reserves after harvest, like a savings account, and free up liquidity during the rest of the year. As several recent papers show, adoption is a major challenge for health insurance programs in low income settings, especially where liquidity is limited. However, very little is known about how seasonal income affects the demand for, and impact of, insurance programs among rural farming populations. In a field experiment in Cote d'Ivoire, we test i) whether customized payment and enrollment schedules can increase the adoption of health insurance schemes in rural agricultural settings; and ii) the extent to which health insurance schemes with optimized payment schedules can reduce seasonality in consumption, investment and care-seeking.