Abstract:

This dissertation consists of three chapters based on original data collected through fieldwork in Western Kenya. The dissertation is focused on whether poor individuals in developing countries are able to cope with the considerable risk that they face, and on what constrains their adoption of new technologies which appear to have the potential to considerably improve their incomes.

The first chapter, entitled "Limited Insurance within the Household: Evidence from a Field Experiment in Western Kenya," employs a field experiment to test whether married couples are able to efficiently insure income risk between themselves. The experiment followed 143 daily income earners and their spouses for 8 weeks. Every week, each individual had a 50% chance of receiving a 150 Kenyan shilling (US $2) income shock. This chapter has 2 main results. First, households do not efficiently pool risk, as male private consumption increases in weeks in which the husband receives the experimental payment but does not change in weeks in which the wife receives the shock. Second, I explore whether limited commitment due to contract unenforceability is a possible explanation for this observed failure of full insurance. I split the sample into 3 Treatment Groups with varying levels of intra-household correlation in the experimental payments. If insurance is constrained by limited commitment, transfers should be highest when spousal incomes are uncorrelated or negatively correlated. I find evidence that women transfer more of the shock to their husbands when the shocks are independent or negatively correlated, a result which suggest that limited commitment is an important constraint on even within-household risk sharing.

The second chapter (co-authored with Ethan Yeh), entitled "Sex Work as a Response to Risk in Western Kenya," is also concerned with how poor individuals cope with risk. This paper studies whether commercial sex workers in Kenya choose to increase their supply of better compensated but riskier unprotected sex in order to cope with unexpected income shocks. Using a panel dataset constructed from 234 self-reported sex worker diaries, we find that sex workers are more likely to see clients and have anal or unprotected sex on days in which a household member falls ill, or on days just after recovering from the symptoms of an STI. Since the HIV prevalence is 9.8% in this part of Kenya, and since health shocks are extremely common among this sample of sex workers, these choices entail significant health risks and have important implications for the spread of HIV/AIDS.

The third paper (co-authored with Esther Duflo and Michael Kremer), entitled "Understanding Technology Adoption: Fertilizer in Western Kenya," studies why it is that only 31.1% of farmers in Western Kenya have ever used fertilizer, even though it appears to be a profitable investment. This paper reports the results from a series of field experiments designed to determine which barriers are most important in constraining technology adoption. The results suggest that learning demonstrations on farmers' plots have relatively large impacts on fertilizer adoption, but that programs which provide farmers an opportunity to commit their harvest income towards future fertilizer use may have even bigger effects. These savings programs suggest that the inability of farmers to save for even a short period of time is a significant impediment to technology adoption.