Essays in applied economics

Abstract

This thesis considers two critical periods in life that can have long-term effects on health and prosperity. The first paper provides new evidence on the consequences of foetal exposure to high levels of pollution for the risk of stillbirth, and for the long-term health and labour market outcomes of those that survive. Variation in in utero exposure comes from a persistent weather system that affected London for five days in December 1952, preventing the dispersion of atmospheric pollution. This increased levels of total suspended particulate matter by around 300%. Unaffected counties in England are used in a difference-in-differences design to identify the short and long-term effects. Historical registrar data for the nine months following the smog show a 2% increase in reported stillbirths in London relative to national trends. As foetal deaths often go unreported, the exercise is then repeated for registered births. The data show around 400 fewer live births than expected in London, or a reduction of 3% against national trends. Survivors are then identified by district and quarter of birth, and their health and labour market outcomes observed at fifty and sixty years old. Differences-in-differences estimates show that survivors are in general less healthy, less likely to have a formal qualification, and less likely to be employed than those unaffected by the smog. The second chapter considers the decision over which skills to acquire at university - taken at seventeen, this decision has significant impacts on both unemployment on graduation, and long-term incomes over the life cycle. Under the hypothesis that more expensive tuition might lead students to acquire skills in high demand, this paper examines the effects of the 2006 increase in fees from 1,200 to 3,000 a year on the probability that a given student would study a stem subject. A propensity matching methodology is used to control for sample selection caused by reduced university participation following the increase in fees. Results indicate that the fee increase caused a five percent reduction in the probability that a given student would study a stem subject. Course level data from the Higher Education Statistics Agency suggests that that subjects most affected were nursing, pharmacology, and medical technology.