Although filled with uncertainties and risks, the stock market attracts many traders and investors who try to exploit it using different approaches. While many see potential for profit in forecasting if a certain stock goes up or down, a team at Morgan Stanley pioneered a way to profit from the financial markets without betting on the direction of the stock in the mid-1980s. Belonging to the category of statistical arbitrage and mostly used by hedge funds, this strategy is called pairs trading. The returns of pairs trading have been long documented in the United States, with data going back to the 1960s. However, little information is available on returns after 2015 and even less on returns in Switzerland. The question therefore arises:
- Is it possible to profit from pairs trading on the Swiss Stock Exchange?
The approach taken in this thesis, called the distance method, is based on two periods, the formation and trading period. Lasting one year, the formation period is when the pairs of stocks that meet a certain criterion are selected and eventually traded in the half-year long trading period given a pre-defined rule to enter and exit a trade. Drawing on 20 years of daily data from all Swiss Performance Index stocks, the above-stated question is answered by back-testing the strategy. In the first iteration, a naïve approach to the method was taken without any restrictions. With this approach unable to achieve a positive return, a second version was developed that introduced a volume filter. All stocks are now subject to a minimum standard of volume during the formation period in order to be selected for trading. However, despite its innovations, this version was also unable to generate a positive return. This underperformance was due to a weak relationship between the chosen pairs which could be traced back to a lack of a fundamental relationship between the stocks themselves. As a solution, a new approach was developed. This version implemented a rule that only stocks operating in the same sector could be matched. Although this solution managed to achieve statistically significant returns of over zero, this profit disappeared when trading costs were subtracted, leaving the strategy mostly negative again. In response, a final version was developed that expanded on the idea of matching related companies – perhaps to an extreme – by pairing stocks representing the same company. It did so by pre-selecting all pairs that will be traded over the whole back-testing period. Although very profitable over the 20 years tested, this strategy was unable to sustain its high returns, eventually losing the edge it seemed to have in 2013. In sum, the implementation of the distance method using swiss stocks exclusively was able to generate a statistically significant return before trading costs but was unable to do so after.