In the leading chapter of this dissertation, "Unemployment, Vulnerability and Poverty: A New Class of Measures, Its Axiomatic Properties and Applications," we, Kaushik Basu and myself, derive a new class of measures for unemployment or poverty. Measures of unemployment and poverty have usually focused solely on those currently unemployed or below the poverty line, ignoring those who are vulnerable to becoming unemployed or falling into poverty. We fully characterize a class of measures that, unlike the standard measures of unemployment or poverty, account for the amount of vulnerability that exists in a society and apply those measures to the USA and South Africa.

In the second chapter of this dissertation, "Unemployment and Family-Values," I propose a new class of unemployment measures that incorporates the externalities one receives from living in a household with employed individuals. The standard measure of unemployment does not do this. A household with at least one employed person is more likely to have heating, water, and other household public goods than a household where everyone is unemployed. Given this, I axiomatize a class of unemployment measures that is sensitive to household unemployment levels. This is done by assuming that if unemployment is held constant, say at fifty percent, an economy where half of each household is employed is better than an economy where half of the households are fully unemployed and half are fully employed.

The third chapter of this dissertation, "Racial Identity, Performance and Self-Confidence: A South African Experiment," deals with the effects of racial identity in Post-Apartheid South Africa. Racial gaps across many measures of performance are well documented. Whether these gaps occur because of discrimination or are due to ex-ante differences between racial groups is still unanswered. Using experimental data from Cape Town, we, Erica Field and myself, examine how different environments and incentive schemes effect a student's performance and her level of self-confidence. We find that cuing a student to her racial identity has a significant affect on performance, self-confidence and where a student places herself within a distribution. These differences are statistically significant even if the groups are ex-ante similar.