Teachers are among the most important school inputs for pupil attainment (Hanushek, 2011). Despite this, economically-advanced countries experience recurring shortages of teachers, resulting in sub-optimal hiring and deployment of teachers and reduced pupil attainment. This thesis investigates the determinants of entry to and exit from the teaching profession in order to understand how these shortages can be reduced. Very little is known about the correlates of entry to the teaching profession. Non-cognitive skills and personality-type have been shown to be important predictors of occupational choice in general (Cobb-Clark & Tan, 2011; Nieken & Stormer, 2010). However, these have not been used to model entry to the teaching profession. Chapter 2 of this thesis uses rich data from a household panel survey to model entry to the profession. The model identifies groups of people who are up to four times more likely to enter teaching than the typical graduate. This information can be used to target recruitment efforts. Retaining teachers is also important for ensuring sufficient supply. Research using administrative data generally finds the proportion of disadvantaged pupils in a school to be the strongest correlate of turnover. However, recent literature suggests that working conditions are important omitted variables in such analysis. Chapter 3 uses data on teachers from thirty-five countries to develop a rich set of working conditions measures and uses these to model teacher job satisfaction and intention to quit. The results highlight the importance of school leadership and assigning teachers to subjects in which they have been trained. Chapter 4 builds on this analysis by evaluating the impact of a subject-specific professional development intervention for science teacher retention. Double- and triple-difference models suggest that participation in the programme improves retention in the profession, though not in the participant’s original school.