Pairs Trading at the Prague Stock Exchange
| Název práce v češtině: | Párové obchodování na Burze cenných papírů Praha |
|---|---|
| Název v anglickém jazyce: | Pairs Trading at the Prague Stock Exchange |
| Klíčová slova: | kointegrace, párové obchodování, výběr párů, návrat ke střední hodnotě, statistická arbitráž, Burza cenných papírů Praha, součet čtverců odchylek, dlouhá a krátká pozice |
| Klíčová slova anglicky: | cointegration, pairs trading, pairs selection, mean reversion, statistical arbitrage, Prague Stock Exchange, sum of squared deviations, long/short position |
| Akademický rok vypsání: | 2011/2012 |
| Typ práce: | bakalářská práce |
| Jazyk práce: | angličtina |
| Ústav: | Institut ekonomických studií (23-IES) |
| Vedoucí / školitel: | prof. PhDr. Ladislav Krištoufek, Ph.D. |
| Řešitel: | skrytý - zadáno vedoucím/školitelem |
| Datum přihlášení: | 04.06.2012 |
| Datum zadání: | 04.06.2012 |
| Datum a čas obhajoby: | 10.09.2014 08:00 |
| Místo konání obhajoby: | IES |
| Datum odevzdání elektronické podoby: | 07.07.2014 |
| Datum proběhlé obhajoby: | 10.09.2014 |
| Oponenti: | Mgr. Tomáš Křehlík, Ph.D. |
| Kontrola URKUND: | |
| Zásady pro vypracování |
| 1. Introduction
2. Literature review and theoretical background a. Minimum-distance method for matching stocks into pairs, justification of its suitability b. Johansen cointegration analysis c. Engle-Granger cointegration test 3. Formation of pairs from stocks traded on the United States’ stock exchange market a. Execution of pairs using minimum-distance method b. Verification of cointegration of pairs in MATLAB using cointegration framework (Engle and Granger (1987), Johansen (1988)) 4. Empirical model a. Determining entering and unwinding position for each pair, separately for various trading horizons b. Backtesting the strategies and verifying their functionality 5. Conclusion a. Relevance of the strategies b. Comparison of different trading horizons c. Assessing the possible profitability of both methods |
| Seznam odborné literatury |
| Engle, R., & Granger, C. (1987). Co-Integration and Error Correction: Representation, Estimation and Testing (Sv. Vol. 55, No. 2). Econometrica.
Gatev, E., Goetzmann, W., & Rouwenhorst, K. (2006). Pairs Trading: Performanceof a Relative-Value Arbitrage Rule. Oxford University Press. Johansen, S. (1988). Statistical Analysis of Cointegration Vectors. Copenhagen: Journal of Economic Dynamics and Control 12. Tsay, R. (2005). Analysis of Financial Time Series. John Wiley & Sons, Inc. Vidyamurthy, G. (2004). Pairs Trading: Quantitative Methods and Analysis. John Wiley and Sons, Inc. |
| Předběžná náplň práce |
| The objective of this thesis is to propose strategies for pairs trading using cointegration approach. The idea of pairs trading is based on fluctuations around the long-run equilibrium of the two stocks that form a pair. The quantity representing the difference in normalized prices between the two stocks is called the spread. When the value of the spread substantially deviates from its mean value, a long-short position is taken with the assumption that the spread will revert back to its equilibrium. If this is the case, the position is unwound and, consequently, profit is made. In the thesis we will focus on implementing trading strategies for various trading horizons, ranging from intraday to weekly and monthly holding periods. We will compare our results and discuss profitability of each strategy.
As a first step of the process, after providing theoretical background supporting the whole thesis, we will need to identify suitable stocks for trading pairs. There are two possibilities how to form pairs. One of them is stock fundamentals analysis, which involves looking at company’s data, e.g. revenue, debt-to-equity ratio, etc. The second approach focuses on technical analysis and takes into account historical prices of stocks. The latter approach is the one we will follow along in our thesis. We will form suitable pairs by ordering the stocks with minimum-distance method of their normalized historical prices. After identification we need to test whether the stocks in a pair are cointegrated. We will use MATLAB computing software to attest the right choice of pairs using cointegration framework (Engle and Granger (1987), Johansen (1988)). Once this is achieved, we will set the trade signals for each pair according to historical comovement of both stocks. As a last step, the two strategies will be backtested and their functionality will then be verified on out-of-sample data. The purpose of this work, however, is not to practically prove the profitability of the strategies; it should serve as an inspiration on a possible method of trading pairs in various trading horizons, which, no matter how promising results it may yield, does not take into account the real-world trading obstacles. |
| Předběžná náplň práce v anglickém jazyce |
| The objective of this thesis is to propose strategies for pairs trading using cointegration approach. The idea of pairs trading is based on fluctuations around the long-run equilibrium of the two stocks that form a pair. The quantity representing the difference in normalized prices between the two stocks is called the spread. When the value of the spread substantially deviates from its mean value, a long-short position is taken with the assumption that the spread will revert back to its equilibrium. If this is the case, the position is unwound and, consequently, profit is made. In the thesis we will focus on implementing trading strategies for various trading horizons, ranging from intraday to weekly and monthly holding periods. We will compare our results and discuss profitability of each strategy.
As a first step of the process, after providing theoretical background supporting the whole thesis, we will need to identify suitable stocks for trading pairs. There are two possibilities how to form pairs. One of them is stock fundamentals analysis, which involves looking at company’s data, e.g. revenue, debt-to-equity ratio, etc. The second approach focuses on technical analysis and takes into account historical prices of stocks. The latter approach is the one we will follow along in our thesis. We will form suitable pairs by ordering the stocks with minimum-distance method of their normalized historical prices. After identification we need to test whether the stocks in a pair are cointegrated. We will use MATLAB computing software to attest the right choice of pairs using cointegration framework (Engle and Granger (1987), Johansen (1988)). Once this is achieved, we will set the trade signals for each pair according to historical comovement of both stocks. As a last step, the two strategies will be backtested and their functionality will then be verified on out-of-sample data. The purpose of this work, however, is not to practically prove the profitability of the strategies; it should serve as an inspiration on a possible method of trading pairs in various trading horizons, which, no matter how promising results it may yield, does not take into account the real-world trading obstacles. |