This thesis contains three chapters, each of which employs both reduced-form and structural tools of applied microeconomics to answer questions collectively in the fields of education, labour and health economics. The first chapter considers the effect of a recent expansion of private tertiary education in a set of six developing countries on the equity and efficiency of labour-market outcomes. In particular, this paper uses a structural modelling approach and data from Armenia, Bolivia, Colombia, Georgia, Ghana and Kenya to explain variation in the growth of private universities and, consequently, to consider the range of impact these institutions have on the efficiency of human capital allocation, aggregate output, and the dynamic evolution of income inequality. I show that differences in underlying parameter values across countries have important implications for the composition of the graduate workforce, the growth of private universities, output per worker and income inequality. The second chapter employs a structural model of endogenous education and occupational choice to demonstrate that youth unemployment in Ghana increases in parental wealth, and to consider the consequences of such a relationship for wage inequality, educational attainment, and aggregate productivity. I argue that, in the absence of unemployment insurance, only workers with a sufficiently high stock of parental wealth can afford to remain unemployed, and do so in order to search for scarce, high-productivity jobs. This leads to high income inequality and low match efficiency among workers of heterogeneous ability. I use the estimation results to compare the effectiveness of two alternative policy interventions: an education subsidy and unemployment insurance. I find that the education subsidy is most effective at increasing aggregate productivity, but comes at the cost of increasing income inequality, while unemployment insurance has a smaller effect on aggregate productivity but also decreases income inequality. The third chapter examines the consequences for efficiency of herding behaviour among UK hospitals in organ transplant queues. My co-authors and I employ unique administrative data from the NHS Blood and Transplant, and a combination of reduced-form and structural modelling techniques, to show that herding behaviour plays an important role in observed organ wastage: once an organ is rejected by one or more centres, subsequent centres emulate their behaviour, ignoring their own assessment of the organ's quality. We use counter-factual analyses to demonstrate that, while herding behaviour is common among UK transplant centres, the resulting increase in discard rates is not substantially higher than that of the full-information benchmark. Equally, it prevents centres accepting organs of poor quality, such that, overall, the benefits derived from observing predecessors' decisions outweigh the costs of herding traditionally emphasised in the theoretical literature. In contrast with this literature, therefore, we find that, in this context, herding enhances efficiency.