Date of Award


Degree Name

Doctor of Philosophy



First Advisor

Dr. Eskander Alvi


Households in poor economies often strive to protect consumption from fluctuations in income through various mechanisms. Due to the absence or scarcity of formal insurance, however, they are often forced to rely on informal means to achieve this objective. This dissertation consists of three essays that analyze the nature, performance and instruments of risk sharing and consumption insurance in poor urban areas.

The first essay identifies the peculiar problems poor urban households face in their attempt to pool risk from idiosyncratic income swings, and derives unique results pertaining to the level of insurance and effort. Two specific problems are identified: lack of commitment due to the unavailability of institutions that enforce contracts, and moral hazard arising from unobservable effort in production. The premise is that information problems of the latter type cannot be ignored in poor urban areas because of the relatively loose societal structure. This is in addition to the ubiquitous enforcement problem. A model of risk sharing under double moral hazard and lack of enforcement is proposed and the inefficiencies that occur relative to the first-best are analyzed. Results show that the deviation from the first best insurance is greatest in the presence of both commitment and moral hazard problems. Interestingly, optimal effort levels under moral hazard and limited commitment are higher than those under moral hazard only.

The second and third essays use extensive survey data to empirically examine risk sharing and consumption insurance in poor urban areas. The second essay adopts consumption based tests to assess the hypothesis, as implied by the theoretical model, that there is significant deviation from complete insurance. Inaddition to least squares, Generalized Method of Moments estimation that controls for endogeneity and measurement error problems is employed. The estimates for various consumption types and income source groups forcefully reject the full insurance null. Further tests reveal that partial insurance occurs among the most vulnerable, and that households may better insure consumption against unemployment than sickness induced shocks.

Consumption based tests indicate the extent to which households are able to insure consumption without being informative about the means adopted. The third essay focuses on examining the instruments of insurance. In particular, it investigates whether households in poor urban areas use private transfers and informal loans as mechanisms of risk sharing and insurance. Pooled and random effects ordered probit estimates show that, despite households' reported use of both private transfers and informal credit mostly for augmenting consumption, only transfer receipts respond to proxies of income shock and vulnerability. Loans appear to serve a different purpose, presumably not insurance, where repayment considerations, not income risks, matter.

Access Setting

Dissertation-Open Access

Included in

Economics Commons