本研究利用傳統的F test、OLS-CUSUM test、ME test 以及Zeileis, Leisch, Hornik, and Kleiber(2005)提出的監控的一般化波動性檢定，探討美國西德州中級原油、布蘭特原油與S&P500大盤指數之月報酬率是否存在結構性變化，希望找出最符合實際的模型建構方法。然後再引用McCraken(2004)和Clark and McCraken (2001)所提出的預測方法及預測能力比較找出結構性變化點是否顯著改善模型預測能力。 實證結果顯示，三者在使用傳統的檢定方法下多無法找出結構性變化，但使用監控的一般化波動性檢定法則可找到結構性變化。而在預測的顯著性檢定上，布蘭特原油月報酬率與美國西德州中級原油月報酬率之滾動預測皆為顯著，表示此兩項預測模型在考慮結構性轉變後能增加模型的準確度。 This research consider a wide array of fluctuation-type tests in a monitoring situation—given a history period for which a regression relationship is known to be stable, we test whether incoming data are consistent with the previously established relationship. We apply our methods to three data sets, returns of West Texas Intermediate oil, returns of Brent crude oil, and S&P 500 stock returns. Then, we generate simulated out-of-sample forecasts, forecast errors, and tests of mean square error (MSE) for a pair of nested models (the first model is a restricted version of the second) of a scalar prediction. The empirical results included that: (1) we can find structural changes from these three data sets by the generalized fluctuation test for monitoring. (2) the rolling forecast models of returns of WTI oil and returns of Brent crude oil both display that considering the factor of structure change will increase the power of forecast.
