Not logged in.

Quick Search - Contribution

Contribution Details

Type Bachelor's Thesis
Scope Discipline-based scholarship
Title Hierarchical Clustering of stocks: Return comovement and comparison to GICS
Organization Unit
Authors
  • Frederik Philipona
Supervisors
  • Patrick Eugster
Language
  • English
Institution University of Zurich
Faculty Faculty of Business, Economics and Informatics
Number of Pages 22
Date 2022
Zusammenfassung The purpose of this thesis is to investigate the quality of different stock groupings formed by hierarchical clustering based on past returns of stocks and to compare them to the GICS stock grouping. The quality of a stock grouping has in this thesis three aspects: 1. The ability of the grouping to put together stocks that exhibit similar behavior, 2. The consistency of the stock grouping over time, 3. Whether or not the resulting stock groups are of reasonable sizes. Finding a high-quality stock grouping can aid investment decisions. This thesis proposes and tests the following simple way of using the information gained from a stock grouping to build a stock portfolio: Instead of running an investment strategy on the whole stock population to build a portfolio, the same strategy is run on every individual stock grouping and then the resulting portfolios are added up equally weighted to a final portfolio, with the idea of lowering risk through better diversification. This approach is tested with the strategy of equal weights. Data of US stock prices from 2010 to 2020 is used. With a low amount of resulting stock groups, the right configurations of hierarchical clustering produce stock groupings of higher quality than the GICS grouping. The above outlined approach of building a portfolio is tested for the strategy of equal weights and the portfolio is compared to a normal portfolio of equal weights. For the right configuration of hierarchical clustering and for a small number of resulting stock groups the risk of a portfolio can be reduced by this simple approach.
Export BibTeX