混沌理論近年來大量應用於財務金融領域後挑戰了傳統資本資產訂價理論的隨機漫步假說,以往CAPM所無法解釋的市場異常現象,如過度反應、元月效應等;卻發現混沌行為的的碎形市場(fractal market hypothesis)假說可解釋這類異常,意味著混沌理論補充了傳統資本資產訂價理論於實務應用的不足。至於混沌行為的偵測,早期研究方法主要採用BDS檢定,但此方法難以區別真正的混沌行為且難適用於含有噪音的資料;故本研究採Gilmore(2001)發展出的複歸檢定(close returns test)偵測混沌,此方法特別適合資料數量不足且含有噪音(noise)的財金數據資料。在BDS檢定與複歸檢定的對照比較後,證實台灣股票市場報酬非隨機漫步而符合混沌理論的碎形市場假說。
Traditionally, market returns have been assumed to be consistent with the random walk hypothesis. Explaining anomalies in market returns, such as seasonal effects, the weekend effect, and the January effect, is difficult. This study examined time series data on Taiwan Weighted Index returns from the perspective of chaos theory. A topological method, the close returns test, was applied to test whether Taiwan stock market returns exhibited chaotic behavior. The main findings of this study are that close returns test outperformed the traditional BDS (Brock, Dechert, Scheinkman) test and that Taiwan stock market returns exhibit recursive behavior rather than random walk behavior.