Asset characteristics based portfolio optimization on country indices
1University of Oulu, Oulu Business School, Department of Finance, Finance
|PDF Full Text (PDF, 1.3 MB)
Mean-variance model of Markowitz is important milestone in the history of the quantitative finance but the model is problematic in real portfolio optimization implementations. The estimation error remains an insuperable problem to overcome despite of many improvements that enhance the performance of the mean-variance model. We derive an asset characteristic based portfolio solution based on the work of Brandt et al. (2009) and Hjalmarsson and Manchev (2012). The data include stock markets in 21 countries in the period of January 1986 to December 2011. Our objective is to show the performance of this kind of simple portfolio optimization method with a set of asset characteristics. We do not seek the best set of characteristics but choose five characteristics that are earnings-to-price ratio, dividend yield, price-to-book ratio, market value and momentum. In addition to asset characteristic portfolios we show performances of equally weighted portfolio and two simple risk parity strategies which we then combine with the asset characteristic portfolio. We also show the importance of the selected asset characteristics for portfolio performance. For portfolio performance metrics we compute Sharpe ratio, Jensen’s alpha ant turnover. Our results support the claim that equally weighted and simple risk parity portfolios are great alternatives to mean-variance model. By out-of-sample performance measures they beat the sample efficient mean-variance tangency portfolio easily. They also show better performance values than the benchmark index, MSCI WI, when measured by Sharpe ratio or Jensen’s alpha although we do not find these measures to be statistically significant. Moreover, our results suggest that performances of these alternative methods can be improved even further by combining them with an asset characteristic based portfolio. The presented portfolio selection methods provide stable and financially sensible results which are in-line with the previous literature.
© Niilo Kärkkäinen, 2014. This publication is copyrighted. You may download, display and print it for your own personal use. Commercial use is prohibited.