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.
Questa tesi di dottorato affronta uno spettro di argomenti irrisolti nell'ambito della ricerca sul benessere (Well-Being). L'ambizione è che i risultati di questa tesi migliorino le decisioni politiche con conseguenze positive sul benessere sociale e individuale. È fondamentale percepire ogni capitolo come un articolo a sé stante, poiché ciascuno di essi esplora domande di ricerca che richiedono dati e approcci metodologici diversi. Il capitolo 1 si concentra sulla previsione della depressione in età avanzata utilizzando dati sul corso della vita e tecniche di Machine Learning (ML). Lo studio utilizza dati dalla Survey of Health, Ageing, and Retirement in Europe (SHARE) e valuta le prestazioni di sei algoritmi di apprendimento automatico. L'analisi dimostra che i modelli che utilizzano dati semi-strutturati in sequenza offrono la massima accuratezza predittiva. In aggiunta, lo studio identifica l'instabilità del corso di vita e il basso utilizzo delle cure dentistiche come nuovi predittori del rischio di depressione in età avanzata, insieme a fattori tradizionali come l'età, la salute nell'infanzia e l'istruzione. Il capitolo 2 esamina l'effetto di ricaduta intergenerazionale del pensionamento dei genitori sul benessere dei figli adulti, utilizzando i dati relativi alle variazioni esogene dell'età di accesso alla pensione pubblica nel Regno Unito. Troviamo che il pensionamento materno aumenta significativamente la soddisfazione di vita e del reddito dei figli adulti. Il pensionamento posticipato dei padri, al contrario, migliora anche il benessere dei figli adulti. Questi effetti sono particolarmente pronunciati tra le famiglie a basso reddito. Il capitolo 3 analizza le disparità etniche sulla piattaforma Airbnb e l'impatto di una politica anti-discriminazione. La ricerca rileva che gli host neri registrano tassi di occupazione inferiori a quelli degli host bianchi, anche avendo considerato le differenze nelle caratteristiche osservabili, come la posizione. Inaspettatamente, la politica volta a ridurre la discriminazione attraverso la riduzione delle dimensioni delle immagini del profilo ha esacerbato questa disparità, aumentando il divario tra bianchi e neri. I risultati suggeriscono che la riduzione delle informazioni visive positive nelle foto del profilo, come il sorriso, può indurre gli ospiti di Airbnb a dare maggiore peso ad indizi razziali come il colore della pelle, evidenziando la complessità di affrontare i pregiudizi nelle piattaforme online. Il capitolo 4 costruisce indicatori spaziali compositi per valutare il benessere economico, sociale e ambientale nelle province italiane, utilizzando un modello Bayesiano a fattori latenti. Questo metodo incorpora le dipendenze spaziali e quantifica l'incertezza nelle misurazioni del benessere. I risultati rivelano forti disparità regionali, in particolare tra il Nord e il Sud Italia, mentre la dimensione ambientale mostra una minore concentrazione geografica. Questo approccio fornisce una comprensione più accurata della distribuzione del benessere, e colma alcuna lacune delle metodologie convenzionali per costruire indicatori compositi di benessere. Il capitolo 5 esplora la relazione tra qualità del governo e benessere ambientale nelle regioni europee. Lo studio costruisce indicatori compositi di benessere ambientale nello stesso modo del capitolo 4. Dimostra che migliori istituzioni regionali sono associate a migliori risultati ambientali, in particolare per quanto riguarda la qualità dell'aria e del suolo. In sintesi, questa tesi contribuisce alla comprensione del benessere offrendo nuove prove empiriche e innovazioni metodologiche in diversi ambiti. Ogni capitolo fornisce spunti di riflessione con l'obiettivo di migliorare il benessere sociale e individuale attraverso interventi informati e guidati dai dati.
EMPIRICAL ESSAYS ON WELL-BEING / Carlotta Montorsi , 2024 Nov 25. 36. ciclo, Anno Accademico 2022/2023.
EMPIRICAL ESSAYS ON WELL-BEING
MONTORSI, CARLOTTA
2024-11-25
Abstract
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.File | Dimensione | Formato | |
---|---|---|---|
Carlotta_Montorsi_PhD_thesis.pdf
accesso aperto
Descrizione: Empirical Essays on Well-Being
Tipologia:
Tesi di dottorato
Dimensione
26.69 MB
Formato
Adobe PDF
|
26.69 MB | Adobe PDF | Visualizza/Apri |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.