摘要
在各個市場與金融商品中，透過曝險於不同的因子所形成不同的因子風格，能夠換取特定的因子溢酬，此一投資方式為：「因子投資」，或可稱 ”Smart Beta”。隨著這種投資方式逐漸受到金融市場投資人的青睞，甚至打入投資人心中成為主流的投資方法下，本研究利用SL Chung, CY Yeh (2011)中部分相同的研究方法：MS-VAR(1)模型(Markov switching vector autoregressive model)，亦即將不同市場景況(Regime)結合至VAR結構，以增進不同因子風格投資組合報酬率可預測性。 經實證後發現：(1) 預測變數在狀態2(景氣較差)時，預測效果較佳；(2) 對於大型股的預測能力較小型股來的顯著；(3) 淨值市價比所生成的平滑機率曲線與「景氣對策分數」對比後，發現較有顯著的預測效果；(4) Fama and French (1992,1993,2018) 的因子風格中，動能因子與PH Chou, YF Liu (2000) 所得的結論相符，其可能不存在於臺灣的金融市場當中。
並列摘要
This study examined the predictive patterns of stock return under different group’s characteristics and the state of the regime. The main purpose of this study was to examine the predict power between the relationship of predictor variables (Dividends index of listed company, Default premium, Secondary market interest rate (percent per annum) on 90 day commercial paper disclosed, and 10-Year Treasury Constant Maturity Minus 2-Year Treasury Constant Maturity) and Factor based portfolio (Size, Book-to-market ratios, Profitability, Investment and Momentum). We used Markov-switching first-order vector autoregressive model (MS-VAR(1)) incorporates the information from qualitative binary variables, such as market regimes, into the VAR structure to facilitate exploring predictive patterns in a conditional way. According to our research result, we found that (1) The appearance of factor style depends on the states of the regime；(2) The predict power of predict variables for large-cap is better than small-cap；(3)smooth probability product by book-to-market’s style portfolio more powerful to predict Business Indicator of Taiwan；(4) Momentum does not exist in Taiwan’s main financial market.