Empirical Essays on Well-Being

This doctoral dissertation addresses a spectrum of research topics, each of significant importance, unified by the general objective of unfolding factors that shape well-being at the national and individual levels. The ambition is that findings from this study will improve policy decision-making and, therefore, boost societal and individual well-being. It is crucial to perceive each chapter as a standalone article, as they each explore different research questions requiring distinct data and methodological approaches. Chapter 1 focuses on predicting depression in old age using life course data and machine learning techniques. The study, drawing from the Survey of Health, Ageing, and Retirement in Europe (SHARE), evaluates the performance of six machine learning algorithms. The analysis demonstrates that models using semi-structured sequence data offer the highest predictive accuracy. Notably, the study identifies life course instability and low dental care utilization as novel predictors of depression risk in later life, alongside traditional factors like age, health in childhood, and education. Chapter 2 examines the inter-generational spillover effect of parental retirement on adult children's well-being using data from exogenous changes in the UK State Pension eligibility age. Fuzzy Regression Discontinuity Design reveals maternal retirement significantly increases adult children's life and income satisfaction. Fathers' postponed retirement, conversely, also improves their adult sons' well-being. These effects are particularly pronounced among lower-income families, underscoring the role of intergenerational time transfers in enhancing adult children's quality of life. Chapter 3 investigates ethnic disparities on the Airbnb platform and the impact of an anti-discrimination policy. The research finds that Black hosts experience lower occupancy rates than White hosts, even after controlling for differences in observable characteristics such as the location. Unexpectedly, the policy aimed at reducing discrimination by shrinking profile picture sizes exacerbated this disparity, increasing the Black-White gap. The findings suggest that reduced positive visual information in profile pictures, like a smile, may lead Airbnb guests to rely more heavily on racial cues like skin color, highlighting the complexity of addressing bias in online platforms. Chapter 4 applies spatial composite indicators to assess well-being across Italian provinces, utilizing a Bayesian latent factor model. This method captures spatial dependencies and quantifies uncertainty in well-being measurements. The results reveal stark regional disparities, particularly between northern and southern Italy, with the environmental dimension showing less geographic clustering. This approach provides a more nuanced understanding of well-being distribution, challenging conventional composite indicator rankings. Chapter 5 explores the relationship between institutional quality and environmental well-being in European regions. The study constructs composite indicators of environmental well-being in the same way as Chapter 4. It demonstrates that better regional institutions are associated with better environmental outcomes, particularly in air and soil quality. These findings emphasize the importance of institutional development in addressing environmental challenges, suggesting that policymakers should focus on strengthening governance to improve environmental well-being at the sub-national level. This dissertation contributes to understanding well-being by offering new empirical evidence and methodological innovations across multiple domains. Each chapter provides actionable insights for policymakers, aiming to enhance societal and individual well-being through informed, data-driven interventions.