Date of Award


Degree Type


Degree Name

Doctor of Philosophy (PhD)



First Advisor

Prof. WEI Xiangdong

Second Advisor

Prof. WONG Ho Lun Alex


This dissertation contains four empirical studies on related topics in development economics. The four studies are all randomized field experiments and all of them use original data obtained from rural households in Northern Ethiopia for empirical analysis. The first study examines the medium term effects of commercial weather index insurance. I examine a randomized controlled trial in which commercial weather index insurance was offered to rural households who are highly vulnerable to drought. The results show that weather index insurance can significantly improve agricultural investments, agricultural yield and also household finance.

The second study evaluates the effects of weather index insurance on rural households' spending on children's education. I examine whether insured households can better pay for their children's education expenses due to better risk management. I also examine whether insured households can maintain their children's education expenses when they experienced rainfall shocks. The results of these two questions are both positive.

The third study examines the level of risk aversion of individuals and of groups. Specifically, I examine whether risk preference among poor people with low level of education can be framed by interactions among peers. The results of a randomized field experiment show that groups not only are more risk averse than individuals but also that the risk preference of individuals can be largely affected by prior interactions in groups.

The final study examines how individuals provide responses to survey questions of sensitive nature. I conducted a randomized field experiment to examine whether survey design methods and survey incentives affect how individuals provide responses to sensitive questions. The results show that individuals who receive high survey incentives will more likely provide positive responses to questions that are moderately sensitive in nature (but not highly sensitive). I find no evidence that the way survey questions are being asked affects the way individuals respond to sensitive questions

Recommended Citation

Kahsay, H. B. (2016). Four empirical essays in development economics (Doctor's thesis, Lingnan University, Hong Kong). Retrieved from

Included in

Economics Commons