Title page for etd-0721113-014612


[Back to Results | New Search]

URN etd-0721113-014612
Author Pei-sang Lee
Author's Email Address No Public.
Statistics This thesis had been viewed 5349 times. Download 0 times.
Department Finance
Year 2012
Semester 2
Degree Master
Type of Document
Language English
Title A Multi-Factor Alpha Model Constructed Using Multi-lag-period Information— with Application in the Taiwan Market
Date of Defense 2013-07-22
Page Count 85
Keyword
  • Alpha model
  • Quantitative portfolio management
  • multi-factor
  • Abstract The main objective of this study is to generate values by combining the current and prior values of descriptors to improve the performance of a portfolio constructed based on the standard alpha model of Hsu et al. (2011). The Polynomial Distributed Lag Model, a time-series model, is adopted to detect the optimal lag length of each company in our research. After measuring an “adequate” lag length for each descriptor, we use the approach of exponential smoothing to combine the current and multi-lag-period descriptors. Instead of using the subjective method applied inHsu et al. (2011), our study calculates some statistics to filter the valid descriptors.
    The empirical results suggest that the new values of the monthly and weekly frequency descriptors should substitute the original ones, especially those within the Value factor. When compared with a portfolio constructed using the raw descriptor values, the IR of the portfolio with the new values of monthly and weekly descriptors is increased from 0.203 to 0.612. Although its tracking error rises slightly by 0.28%, this portfolio still achieves the requirement of an enhanced index fund, which is below 3%.
    Advisory Committee
  • Shyh-weir Tzang - chair
  • Yi-Hsi Lee - co-chair
  • Yih Jeng - advisor
  • Files
  • etd-0721113-014612.pdf
  • Indicate in-campus at 99 year and off-campus access at 99 year.
    Date of Submission 2013-08-21

    [Back to Results | New Search]


    Browse | Search All Available ETDs

    If you have more questions or technical problems, please contact eThesys