Empirical studies of vulnerability to weather catastrophes

How vulnerable are poor households to weather catastrophes? Why are some households and communities more vulnerable than others? How can government policies help poor households prepare for and cope with weather shocks? While much has been written about vulnerability, few empirical studies show how households, particularly poor households in developing societies, respond to and recover from external stresses. I will use longitudinal household surveys to study household responses to the 1997 El Niño event in Peru and Hurricane Mitch in Nicaragua. In both cases household surveys were conducted before and after the weather shock. Preliminary analysis of both datasets shows that they contain information relevant to the research questions posed below. In the case of Peru, I have already augmented the household survey data with remote sensing and agricultural census data. The Nicaragua data are also geographically explicit, and I am currently working to obtain additional variables.

In first phase of this work I will formulate a conceptually valid empirical measure of household vulnerability that can be estimated from household survey data. I anticipate that this measure will draw on recent work by Pamela Matson’s research group at Stanford, which has been developing a quantitative measure of ecosystem vulnerability using remotely sensed data from Mexico. The essence of their approach is to model vulnerability as a function of sensitivity to perturbations and proximity to some damage threshold.

A second phase will explore underlying causal relationships between vulnerability and household and community characteristics, and will explore what kind public policy can mitigate the social impacts of weather shocks. Characteristics of interest include education, income, access to credit, and agricultural practices. In addition to improving the analytical foundation for policies related to El Niño and other sporadic weather shocks, this work will contribute to ongoing efforts to predict the impacts of climate change in developing countries.


Learning and the diffusion of clean household energy technology

What factors affect the diffusion of technologies that promote both environmental quality and human well being? In particular, how do households learn about the benefits and costs of new technologies, and how does this learning process affect diffusion dynamics? I ask these questions in the context of household energy technology. According to recent World Health Organization analysis, indoor air pollution from cooking with solid fuels in poorly ventilated indoor spaces kills approximately 1.5 million people each year and adds hugely to the burden of disease. Yet households with alternatives continue to cook with traditional fuels.

My dissertation research analyzes the spread of clean household fuels in Peru. I use longitudinal household survey data to study the extent to which poor households learn from their neighbors about the health benefits of clean household energy technologies. I show that household’s beliefs regarding the health benefits of switching to clean-burning fuels reflect neighbors’ experiences in a manner consistent with a Bayesian learning model. This establishes that information about health effects plays a role in fuel choice.

As an EI Postdoctoral Fellow I hope to explore how much “expert” information affects behavior. I plan to study this question by carrying out experiments in which communities in the central Sierra of Peru are exposed to different kinds of information about the risks of indoor air pollution. I am currently working with The Mountain Institute, a non-governmental organization that works in Peru, and the Peruvian Ministry of Social Affairs to lay the groundwork and obtain funding for these experiments.


Ecosystems services for the poor

Efforts to quantify the services rendered by ecosystems have largely focused on developing countries, despite the greater dependence of non-urban poor on ecosystems. Household survey data contains potentially useful information about the nature and magnitude of the benefits that poor households derive from ecosystems (water, food, fuel, waste disposal, etc.). This research investigates how standard methods for ecosystem valuation using household production function models can be adapted to household survey data from developing countries. To test these new methods I will study variation in societal reliance on ecosystem services in several Latin American countries.

A central challenge here is to frame valuation questions in terms of changes in levels of ecosystem services resulting from human activity (e.g. the loss attributable to a mining project that disrupts a village water supply). Another challenge is that approaches to ecosystem valuation that rely on willingness to pay for ecosystem services are extremely sensitive to income. In wealthy countries where these tools evolved analysts have handled these distributional concerns on an ad hoc basis. Large variations in household income typical in developing countries underscore the need for a more systematic way of comparing the value of ecosystem services across populations with disparate levels of income.