In Chapter Five, I use fixed effects regression with well-being measurements as the dependent variables. The choice of fixed effects regressions is supported by Hausman tests (p- value <0.000 for all indicators). The different measures yield no evidence for unit roots, as augmented Dickey-Fuller tests with one lag rejected the null hypothesis that all panels contain unit roots for all well-being indicators (p-value <0.001 for all indicators). Models are thus estimated in levels. Following Lagrange multiplier tests, annual dummies are included (p <0.000 for all well-being indicators) in all models. This holds potentially unobserved time-specific variables, such as crisis years that could affect well-being outcomes, constant.