Work detail

Fama-French Model: Multiscale Portfolio Analysis

Author: Bc. Radek Spousta
Year: 2018 - summer
Leaders: Mgr. Lucie Kraicová
Consultants:
Work type: Bachelors
Language: English
Pages: 66
Awards and prizes:
Link: https://is.cuni.cz/webapps/zzp/detail/191028/
Abstract: This thesis studies the empirical relationship between excess asset returns and the
Fama−French risk factors at various scales using a combination of the Fama−French model
and wavelet-based methods. We re-examine previously published results obtained for six
portfolios formed on size and book-to-market ratio in the U.S. market, and focus on the
influence of different scales on the original results. We conclude that the most the total
variance of the risk factors and excess portfolio returns is concentrated at scale 1 and 2, which
corresponds to periodicities of 2-4 months and 4-8 months, respectively. Next, we observe
significant variation in estimated parameters across different scales. Furthermore, some of the
Fama−French risk factors are strongly correlated at scale 2, 3 and 4, which is unobservable in
standard correlation matrix. Overall, the multiscale approach seems beneficial for analysis of
the Fama−French three-factor model as it reveals information that remains hidden to
traditional methods.

Partners

Deloitte

Sponsors

CRIF
McKinsey
Patria Finance