This dissertation looks into whether quantile regression (QR), a distribution-sensitive estimation technique, provides additional information on the behavior of Fama–French factor exposures relative to the traditional Ordinary Least Squares (OLS) approach. Motivated by the well-known limitations of OLS whereas there are fat tails, asymmetries, and heteroscedasticity (features that characterize financial return distributions) the study applies both OLS and QR to the 25 size–book-to-market portfolios using monthly data from July 1926 to August 2025. However, the results suggest that, while QR offers interesting and methodologically appealing distributional insights, it does not introduce genuinely innovative elements in the analysis of systematic risk exposures within the Fama–French framework. The traditional OLS approach remains sufficient to capture the fundamental risk–return relationships in this empirical setting. The dissertation concludes by discussing implications for asset pricing and proposing future research directions in contexts where nonlinearities and tail dependencies may play a more important role.
Questa tesi esamina se la regressione quantile (QR), una tecnica di stima sensibile alla distribuzione, fornisca ulteriori informazioni sul comportamento delle esposizioni ai fattori Fama-French rispetto al tradizionale approccio dei minimi quadrati ordinari (OLS). Motivato dai ben noti limiti dell'OLS in presenza di code spesse, asimmetrie ed eteroscedasticità (caratteristiche che contraddistinguono le distribuzioni dei rendimenti finanziari), lo studio applica sia l'OLS che la QR ai 25 portafogli size-book-to-market utilizzando dati mensili dal luglio 1926 all'agosto 2025. Tuttavia, i risultati suggeriscono che, sebbene il QR offra approfondimenti distributivi interessanti e metodologicamente accattivanti, non introduce elementi realmente innovativi nell'analisi delle esposizioni al rischio sistematico nell'ambito del quadro di Fama-French. L'approccio tradizionale OLS rimane sufficiente per cogliere le relazioni fondamentali tra rischio e rendimento in questo contesto empirico. La tesi si conclude discutendo le implicazioni per la valutazione degli asset e proponendo future direzioni di ricerca in contesti in cui le non linearità e le dipendenze di coda possono svolgere un ruolo più importante.
Assessing the Value of Quantile Regression in Fama–French Factor Analysis: Evidence from Size and Value Portfolios
SU, TONG
2024/2025
Abstract
This dissertation looks into whether quantile regression (QR), a distribution-sensitive estimation technique, provides additional information on the behavior of Fama–French factor exposures relative to the traditional Ordinary Least Squares (OLS) approach. Motivated by the well-known limitations of OLS whereas there are fat tails, asymmetries, and heteroscedasticity (features that characterize financial return distributions) the study applies both OLS and QR to the 25 size–book-to-market portfolios using monthly data from July 1926 to August 2025. However, the results suggest that, while QR offers interesting and methodologically appealing distributional insights, it does not introduce genuinely innovative elements in the analysis of systematic risk exposures within the Fama–French framework. The traditional OLS approach remains sufficient to capture the fundamental risk–return relationships in this empirical setting. The dissertation concludes by discussing implications for asset pricing and proposing future research directions in contexts where nonlinearities and tail dependencies may play a more important role.| File | Dimensione | Formato | |
|---|---|---|---|
|
SU_TONG.pdf
Accesso riservato
Dimensione
6.02 MB
Formato
Adobe PDF
|
6.02 MB | Adobe PDF |
The text of this website © Università degli studi di Padova. Full Text are published under a non-exclusive license. Metadata are under a CC0 License
https://hdl.handle.net/20.500.12608/101985