This thesis in applied microeconomics consists of three chapters that are devoted to specific questions in the fields of subjective well-being (life and job satisfaction) and human capital formation. All three chapters aim at deepening our understanding of the determinants of human well-being by investigating causal medium and long-term effects of unexpected and exogenous political, institutional or environmental changes. The first chapter analyses the question whether different levels of job satisfaction between public and private sector workers represent public sector rents or merely a spurious correlation due to workers' self-selection across sectors. The novel identification strategy exploits the massive post-Soviet privatization process in Ukraine as a quasi-experiment based on which each individual can be assigned an ex ante, exogenous privatization probability. The results point to a causal public sector satisfaction premium which even slightly increases after correcting for self-selection. Chapter 2 is devoted to the assessment of the long-run effects of the Chernobyl catastrophe in 1986 on subjective well-being in post-Soviet Ukraine. The analysis is based on a nationally representative Ukrainian data set and reveals that even 20 years after the accident subjective well-being is negatively associated with self-reported assessments of being affected by the catastrophe. The causal long-term effect of the disaster on life satisfaction is being tested and supported by exploiting official radiation measures which are linked to survey respondents through information on their place of living in 1986. The final chapter turns to the question whether long-term cognitive child outcomes are affected by the duration of parental leave, i.e. by the time mothers spend at home after giving birth and before returning to work. Employing a difference-indifference approach, this chapter evaluates an unexpected reform in Austria which extended the maximum duration of paid and job protected parental leave from 12 to 24 months for children born on July 1, 1990 or after. While the results based on the pooled sample suggest no significant overall impact of extended maternal leave taking on standardized PISA test scores of children aged 15, the subgroup of boys of highly educated mothers seem to have benefited.