Return to search

Portfolio Construction Methodology with the Equally-Weighted Risk Contribution Strategy¢wEvidence from Taiwan Weighted Index

Even though the framework of mean-variance analysis is convincing, in practice, investors encounter serious drawbacks. Understandably, a more stable and rather simple method to make investment decisions without depending on the expected returns would obviously be preferred by some investors. In this study, we adopt a newly proposed equally-weighted risk contribution portfolio (ERC), without the assumption of expected returns, in order to observe its risk and return, as well as the timing of use compared to different benchmarks, the Taiwan 50 index (TWN50) and Taiwan weighted index (Y9997). For comparison, we adopt the other two commonly used methods, the minimum variance portfolio (MVP) and the equal weight portfolio (EW).
It is interesting to observe the ERC¡¦s risk and return profile because, similar to the EW, it invests in every asset in a portfolio while adjusting weights to consider the marginal risk between each component. Therefore, no individual asset¡¦s risk contribution is dominated by an other and they all have the same risk contribution.
In addition, to strike a balance between risk control and pursuing excess return, we incorporate arbitrage portfolios into the standard ERC portfolio; to satisfy the standard of the passive funds, we combine the minimum tracking error portfolio with the standard ERC portfolio by a specific ratio.
From the TWN50 and the Y9997 cases, we presume that the problem relating to whether or not the benchmark is a full sample or a partial sample and whether it¡¦s replicable, affects the performance of the ERC. If our benchmark is a partial sample like the TWN50, the ERC strategy can outperform the benchmark and even provides some degree of defensive ability when the market trend is down.

Identiferoai:union.ndltd.org:NSYSU/oai:NSYSU:etd-0720112-153532
Date20 July 2012
CreatorsTseng, Yi-Chiang
ContributorsShyh-Weir Tzang, Yih Jeng, Yi-Hsi Lee
PublisherNSYSU
Source SetsNSYSU Electronic Thesis and Dissertation Archive
LanguageEnglish
Detected LanguageEnglish
Typetext
Formatapplication/pdf
Sourcehttp://etd.lib.nsysu.edu.tw/ETD-db/ETD-search/view_etd?URN=etd-0720112-153532
Rightsuser_define, Copyright information available at source archive

Page generated in 0.0026 seconds