Not logged in.
Quick Search - Contribution
Contribution Details
| Type | Master's Thesis |
| Scope | Discipline-based scholarship |
| Title | Multivariate Methods for Stock Selection |
| Organization Unit | |
| Authors |
|
| Supervisors |
|
| Language |
|
| Institution | University of Zurich |
| Faculty | Faculty of Business, Economics and Informatics |
| Number of Pages | 49 |
| Date | 2018 |
| Zusammenfassung | A factor can be thought of as any firm characteristic that has predictive power in explaining its return and risk (Bender et al. (2013)). Evidence suggests that a variety of factors are related to cross-sectional return predictability. For example, Hou et al. (2017) document 447 factors that have been published in academic literature. Factors can be used for testing or implementing investment strategies by selecting stocks into portfolios based on the value of realized factors. In the academic literature, most studies present strategies that select stocks based on a single factor. Some studies jointly analyze factors and find that selecting stocks based on multiple factors provide some advantages. For example, using multiple factors instead of one allows to incorporate more of the information available about the expected cross-section of stock returns (Huang et al.(2017)). A combination of factors may jointly predict the cross-section, meaning that the aggregated measure provides more information than the sum of the single variables (Han et al. (2016)). Novy-Marx (2013) show that a strategy that selects stock based on valuation and quality measures jointly generates more exposure to both factors than a 50/50 combination of the returns based on a singlefactor sort. Stambaugh et al. (2015) find that aggregating the information contained in several anomalies factors provide a better signal as it eliminates some noise contained in each single factor. Some researchers point out that many of the supposedly significant factors reported in the academic literature are most likely spurious (Lo and MacKinlay (1990); Harvey et al. (2016); Linnainmaa and Roberts (2016); Hou et al.(2017)). Therefore, an investor faces the risk of not knowing exacte which factors really matter for predicting stock returns. If an investor applies an aggregation method that automatically ignores or at least underweights irrelevant variables, a composite strategy may also serve as a protection of not knowing ex-ante which factors really matter. Therefore, the selection of an appropriate aggregation method matters to investors who seek to aggregate multiple factors in a systematic way. The question of how to combine many factors when constructing an actual portfolio has received less attention in the empirical asset pricing literature. The purpose of this thesis is to detect available methods that allow aggregating the information from both price and fundamental variables when selecting stocks, identify potentially promising and appropriate methods, and investigate the performance of the strategy including robustness checks. |
| Export |
BibTeX
|