Essays on Programme Impacts, Vulnerability and Welfare Measurement

This dissertation consists of three essays on programme impacts poverty and vulnerability. The first essay examines the impacts of a conditional cash transfer program on vulnerability to poverty. The second essay studies the effects of development aid program on the poor in Mozambique, while the third essay develops an alternative vulnerability measure that incorporates aspects of Prospect Theory and in this setting, brings forth a new way to measure and compare social welfare empirically. The first essay explores for the first time, the impacts of the Mexican conditional cash transfer programme 'Progresa' on vulnerability to expected ultra-poverty (less than US$0.5 per day) using consumption and income. Different from traditional ex-post poverty measures, the paper explores the Vulnerability as Expected Poverty model (VEP) ex-ante that yields probability statements regarding vulnerability, the risk that an individual will suffer from poverty in the future. The results indicate that the treated population has lower levels of vulnerability to ultra-poverty (20 percentage points) and higher levels of expected consumption (17 log-percent) and income in the future (over 50 log-percent, when including transfers) compared to their untreated counterparts in the control group. Treatment effects on disadvantaged groups are even larger. This paper suggests that the expectations for better future livelihood of the treated people have clearly changed after the launch of the cash transfer programme. The second essay evaluates the impact of an intervention to improve farming techniques and food security in the Gaza area of rural Mozambique. We examine the impact of a group-based approach to technology adoption in subsistence agriculture, using panel data collected by our research team on over 200 households from treatment and control villages from 2008-10. The intervention was successful in encouraging vulnerable households to participate in farmers' groups, and the impacts on farming techniques, such as fertilizer use, are significant in the first treatment year. The impact on food security is mixed across indicators but similar in both treatment years and cannot be attributed to whether or not households adopted new technologies. The third essay examines the measurement of social welfare, poverty and inequality taking into account features that have been found to be important welfare determinants for people. Most notably, we incorporate reference-dependence, loss aversion and diminishing sensitivity – aspects emphasized in Prospect Theory – to social welfare measurement. We suggest a new notion of equivalent income, the income level with which the individual would be as well off, evaluated using a standard concave utility function, as he actually is, evaluated with a reference-dependent utility function. We examine the differences between standard poverty and inequality measures based on observed income and measures that are calculated based on equivalent income. These differences are illustrated using household-level panel data from Russia and Vietnam