| 摘要: | 探討TSE、OTC、台指VIX關聯，並以回測驗證在台股投資之效益。本文獻先以時間序列的單根判定，判別資料是否定態，接著以Johansen共整合判定，評估有無長線平衡關係。若有，以向量差數匡正模型（Vector Error Correction Model, VECM）跑迴歸。針對迴歸之結果以Granger因果判定、變異數拆分、衝撞反饋，來衡量三者變數間的關係。再將此時間序列之結果，作為回測驗證（Backtesting）的定錨錨點，以華倫巴婓特選股法作為選股標準，投資策略為買進持有（Buy and Hold）及波段（Band Trade），TSE、OTC、台指VIX則為進、出場因子的36種投資組合，採回測方式驗證此三變數在台股投資之效益。
時間序列之研究結果，向量差數匡正模型（VECM）顯示，OTC、VIX對TSE都有長線關係，VIX對TSE長線負向，OTC對TSE長線正向。根據Granger因果判定，TSE與VIX有雙向回饋的因果關係；TSE對OTC有單向領先因果關係。根據衝撞反饋，TSE、VIX在面對來自本身衝撞時反饋最為激烈，OTC則是受到TSE的衝撞最劇烈；VIX與TSE二者之間確實存在長線負向影響關係；TSE對OTC全程影響高度顯著。
全部上市櫃個股及以時間序列之研究定錨，採XQ回測之結論，TSE對OTC，確實具有高強度的正向關係，且單向領先因果關係。在僅針對上櫃公司回測部分，TSE為因子之投資報酬率加總高居第一名，與時間序列Granger因果判定、衝撞反饋之分析完全相符。分別加總三個因子的「投報率加總」後，反映出了，因投資策略、入場時機的不同，OTC外生變數最弱，但依然可以在有效的投資組合運用下，獲得高於其他因子的報酬率。從時間加權報酬率解讀之數據可看出，全部因子TSE、VIX、OTC，都無可避免受到大趨勢影響，而一致性地有較好、較差的報酬率。因而可了解一點，TSE、VIX、OTC因子故然有影響投資績效，但大環境趨勢為何，影響卻是更為重大。
;Explore the relationship between TSE, OTC, and Taiwan VIX, and verify the benefits of investing in Taiwan stocks with backtesting. In this paper, Unit Root Test of the time series is used to determine whether the data is in a stationary state, and then the Johansen Cointegration Test is used to evaluate whether there is a long-term equilibrium relationship. If so, run the regression with the Vector Error Correction Model. For the regression results, Granger Causality Test, Variance Decomposition, and Impulse Response Function are used to measure the relationship between the three variables. The results of this time series are then used as the anchor point for the backtesting verification, and the Buffett′s Stock Selection Method is used as the stock selection standard. The investment strategy is Buy and Hold, and Band Trade. TSE, OTC, and Taiwan VIX are 36 investment portfolios of entry and exit factors. Backtesting is used to verify the benefits of investing in Taiwan stocks for these three variables.
The research results of time series, the VECM shows that OTC and VIX have a long-term relationship with TSE, VIX has a negative long-term relationship with TSE, and OTC has a long-term positive relationship with TSE. According to the Granger Causality Test, TSE and VIX have a two-way feedback causal relationship; TSE has a one-way leading causal relationship with OTC. According to IRF, the TSE and VIX have the most intense feedback when facing the collision from themselves, while the OTC is the most violently impacted by the TSE; there is indeed a long-term negative relationship between the VIX and the TSE; the TSE has a highly significant impact on the OTC throughout the process. .
All listed OTC stocks are anchored by time series research, and the conclusion of XQ backtesting is that TSE does have a high-strength positive relationship with OTC, and one-way leading causality. In the part of backtesting only for listed companies, TSE is the factor with the highest ROI, which is completely consistent with the analysis of time series Granger Causality Test and IRF. After summing up the "Return rate of Summation" of the three factors respectively, it reflects that due to different investment strategies and entry timings, OTC exogenous variables are the weakest, but they can still be used in an effective investment portfolio. From the data of time-weighted rate of return interpretation, it can be seen that all factors TSE, VIX, OTC are inevitably affected by the general trend, and consistently have better and worse rates of return. Therefore, it can be understood that TSE, VIX, and OTC factors have an impact on investment performance, but the impact of the general environmental trend is more significant. |
