策略資產配置是影響長期投資績效的重要因素,很多研究都曾將不同理論模型運用在策略資產配置上,如:多變量模型(Campbell et al.(2003))、動能交易策略(Bange and Thomas(2004))、(Brouwer and Philippe(2009))等,而本研究採用共整合分析法在策略資產配置上。共整合自1987年被提出後,很多研究採用共整合分析法去尋找非定態時間序列的長期關係,而Lucas(1997)認為共整合會影響戰術和策略上的財務決策,故本研究欲探究共整合分析法在策略資產配置上的風險分散效果。 本研究以12檔避險基金指數及代表傳統資產的標準普爾500指數(S&P 500)、那斯達克指數(NASDAQ)與J.P.摩根債券指數(J.P. Morgan bond)為研究對象,採用共整合分析法,探究經共整合測試後的策略資產配置之效率前緣。實證結果顯示在平均數-變異數及平均數-條件風險值的報酬-風險架構下,以不具共整合的資產為選取標的之策略資產配置能有較高報酬,且風險較低。
The strategic asset allocation is an important factor that influenced long-term investment performance; many researchers have already made use of different theories in the strategic asset allocation, such as multivariate model (Campbell et.al. (2003)), momentum trading strategy (Bange and Thomas (2004)), behavioral investing (Brouwer and Philippe (2009)), while this study applied cointegration analysis in the strategic asset allocation. Cointegration had been defined and developed since 1987, a lot of studies used cointegration to look for the long-turn relationship among the nonstationary time series. Lucas(1997) believed that cointegration affects both tactical and strategic financial decision making. This study wants to investigate cointegration analysis in strategic asset allocation effect on diversification potential. This study applies cointegration analysis to a sample containing 12 hedge fund indices and indices representative of traditional assets, such as S&P500、NASDAQ and J.P. Morgan bond to explore the efficient frontier of strategic asset allocation.Empirical results show that under mean-risk framework, such as mean-variance model and mean-conditional value at risk model, the strategic asset allocation based on the portfolio of non-cointegrated assets has higher return and lower risk.