Co-integration is an econometric property of time series variables. If two or more series are themselves non-stationary (unit root process), but a linear combination of them is stationary, then the series are said to be co-integrated. If there is a co-integration among some time series, we can say there is a long-run equilibrium. That is the non-stationary time series may diverge from each other in short-run, however they would arrive at equilibrium in long-run. Therefore, we can use this methodology to test the existence of commonality of some non-stationary time series. Here we apply a semi-parametric cointegration test introduced by Cheng and Phillips (2008) to three issues: commonality of hedge funds with different strategies, the co-movement of different industries and financial markets of different countries. The test shows that there is a co-integration among nine different hedge funds strategies and this result provides a support for the factor-seeking methodology used in Agarwal and Naik (2004), Fung and Hsieh (2001), and Fung and Hsieh (2004) which find factors for hedge funds from specific strategy and use these factors to the whole industry. For industry, there is also a full rank cointegration among five industries: consumer, manufactory, high-tech, health and other and therefore different industries co-move with each other in long-run. The test of financial markets of different countries shows that there is no long- run equilibrium among financial markets of USA, UK, Germany, France, Hong Kong, Japan and Singapore.
MSc in Finance
Portfolio and Security Analysis
Three Sections of Applications of Co-Integration: Hedge Funds, Industry and Main Global Equity Markets. (2009). Dissertations and Theses Collection (Open Access).
Available at: http://ink.library.smu.edu.sg/etd_coll/47