| Summary: | This thesis applies and extends microeconometric methods in the analysis of economic questions related to education choice and fertility choice, and their interaction with the labour market. Chapter 2 extends the literature on non-parametric bounds on the returns from education by allowing for non-random selection into both education and labour market participation simultaneously. Allowing for both forms of non-random selection leads to very wide bounds on the estimated returns from education. This finding highlights both the role of parametric assumptions in pinpointing the magnitude of these effects, but also the importance of rigorously validating the assumptions leading to point estimates in such a wide identified set. Chapter 3 reviews the marginal treatment effect approach. This chapter also outlines how the MTE model can be used to estimate the selection effect, and to estimate whether an advantage exists for the treated group for either potential outcome (treated and nontreated). This chapter also rigorously discusses the comparison of ATE, OLS and IV estimates, and discusses what can be inferred from these comparisons. Chapter 4 estimates heterogeneity in the returns to higher education in the UK. Significant heterogeneity due to observable characteristics was found, in particular with high ability individuals receiving lower returns to higher education than lower ability individuals. Since graduates have higher mean levels of ability than non-graduates this leads to negative selection; individuals who do not attend higher education stand to gain more from participation than those who do attend. This counter-intuitive finding is discussed and possible economic explanations explored. Finally, chapter 5 found that extensions in the duration of paid maternity leave led to deteriorating female labour market conditions, with female employees receiving lower pay and experiencing higher levels of redundancy relative to males as a result of an expansion in the duration of paid maternity leave.
|
|---|
