Since its’ invention at Morgan Stanley in 1987 pairs trading has grown to be one of the most common and most researched strategies for market neutral returns. The strategy identifies stocks, or other financial securities, that historically has co-moved and forms a trading pair. If the price relation is broken a short position is entered in the overperforming stock and a long in the underperforming. The positions are closed when the spread returns to the long-term relation. A pairs trading portfolio is formed by combining a number of pairs. To detect adequate pairs different types of data analysis has been used. The most common way has been to study historical price data with different statistical models such as the distance method. Gatev et al (2006) used this method and provided the most extensive research on the subject and this study will follow the standards set by that article and add new interesting factors. This is done through an investigation on how the analysis can be improved by using the stocks fundamental data, e.g. P/E, P/B, leverage, industry classification. This data is used to set up restrictions and Lasso models (type of regression) to optimize the trading portfolio and achieve higher returns. All models have been back-tested using S&P 500 stocks between 2001-04-01 and 2015-04-01 with portfolios changed every six months. The most important finding of the study is that restricting stocks to have close P/E-ratios combined with traditional price series analysis increases returns. The most conservative measure gives annual returns of 3.99% to 4.98% depending on the trading rules for this portfolio. The returns are significantly (5%-level) higher than those obtained by the traditional distance method. Considerable variations in return levels is shown to be created when capital commitments are changed and trading rules, transaction costs and restrictions on unique portfolio stocks are implemented. Further research regarding how analysis of P/E-ratios can improve pairs trading is suggested. The thesis has been written independently without an external client and studied an area that the author found interesting.
| Identifer | oai:union.ndltd.org:UPSALLA1/oai:DiVA.org:umu-104314 |
| Date | January 2015 |
| Creators | Jakobsson, Erik |
| Publisher | Umeå universitet, Institutionen för matematik och matematisk statistik |
| Source Sets | DiVA Archive at Upsalla University |
| Language | English |
| Detected Language | English |
| Type | Student thesis, info:eu-repo/semantics/bachelorThesis, text |
| Format | application/pdf |
| Rights | info:eu-repo/semantics/openAccess |
Page generated in 0.0017 seconds