本研究以 2017年5月至2024年12月，合計 1,836個交易日的日頻資料為樣本，並將 2020年3月23日逐筆撮合作為市場結構斷點進行微觀分析。首先透過 ADF 與 PP 單根檢定確認所有報酬與成交量序列在1％顯著水準內皆為一階定態，排除假迴歸疑慮。以GARCH(1,1)為基準，引入 ARJI模型將跳躍強度置入條件波松過程並以最大概似法估計。ARJI 量化出跳躍強度λ約為 0.0415、持續性ρ約為 0.753；負報酬對 λₜ 的邊際效應顯著為正，驗證「壞消息放大跳躍」的非對稱性。盤後成交量對翌日日盤報酬與開盤跳空具明顯正向預測力，在 FOMC、CPI 等重大事件夜，跳躍對總變異的貢獻超過三成，證明價格發現在盤後交易中領先性及有效性。
逐筆撮合上線後，盤後平均報酬由原先的微幅負值翻轉為小幅但具統計顯著性的正值；然而分布仍呈厚尾且偏向左側。日內振幅與波動持續性雙雙大幅抬升，呈現雜訊放大、跳躍占比較低但擴散變異更高的現象。監理層面上，建議將即時跳躍強度與隱含波動度納入動態風控管理，並在訊息密集時段搭配動態價格範圍、短暫冷靜期與造市獎勵，以降低極端跳空與滑價風險。專業法人可利用盤後交易報酬均值回歸進行隔日反向套利，並透過期權部位調整 vega、gamma 以因應跳躍群聚與波動記憶延長；一般投資者則宜避開重大事件前後的秒級搶單，以降低流動性枯竭與交易成本。綜上所述，ARJI 模型得以系統化描繪臺灣盤後市場的跳躍動態，揭示逐筆撮合帶來的即時性提升與尾部風險擴張雙刃效果，為日後的監理政策與投資策略提供定量基礎。

Using 1,836 daily observations (May 2017 – Dec 2024) and treating the 23 Mar 2020 shift to continuous auction as a micro structural breakpoint, this study first confirms—via ADF and Phillips Perron tests—that all return and volume series are I(0) at the 1 % level, eliminating spurious regression risk. A baseline GARCH(1,1) is then augmented with an ARJI specification, embedding jump intensity λₜ in a conditional Poisson process and estimating parameters by maximum likelihood. The model delivers λ ≈ 0.0415 and persistence ρ ≈ 0.753; negative shocks significantly raise λₜ, evidencing bad news jump amplification. After hours volume strongly predicts next day regular session returns and opening gaps, while on FOMC/CPI nights jumps generate >30 % of total variance—affirming the leading price discovery role of the overnight market. 
Post auction, mean after hours returns flip from –0.024 to 0.0857, yet distributions remain fat tailed and left skewed; intraday range and volatility persistence rise sharply, indicating amplified noise with fewer jumps but higher diffusive variance amid thinner depth. Regulators should integrate real time λ and implied volatility into dynamic risk controls, complemented by adaptive price limits, brief cooling off intervals, and market maker incentives to curb gap and slippage risk. Professionals can exploit mean reversion in overnight returns for contrarian trades and hedge vega/gamma against clustered jumps; retail traders should eschew millisecond orders around major announcements to avoid liquidity drain costs. The ARJI framework thus quantifies Taiwan’s after hours jump dynamics and highlights the continuous auction trade off: faster price discovery but greater tail risk—a crucial input for future policy and strategy.

目錄
第一章 緒論 ……………………………………………………………………………1
  1.1 研究背景  ………………………………………………………………………1
1.2 研究動機  ………………………………………………………………………3
1.3 研究目的  ………………………………………………………………………5
第二章  文獻回顧  ……………………………………………………………………9
2.1 期貨盤後交易是否存在價格發現功能  ………………………………………9
2.2 盤後交易市場衍生性商品與價格發現機制 …………………………………14
　2.3 盤後交易及到期效應   ………………………………………………………16
第三章 變數定義及研究模型  ………………………………………………………21
3.1 變數定義   ……………………………………………………………………21
    3.1.1指數報酬率  ……………………………………………………………21
    3.1.2指數波動度  ……………………………………………………………21
    3.1.3指數成交量  ……………………………………………………………22
3.2 實證模型定義及功能 …………………………………………………………22
      3.2.1 ADF單根檢定法 ………………………………………………………22
      3.2.2 PP單根檢定法…………………………………………………………24
3.2.3 ARJI模型………………………………………………………………24
第四章 資料處理與來源  ……………………………………………………………28
4.1 資料來源 ………………………………………………………………………28
4.2 資料處理 ………………………………………………………………………28
第五章  實證結果分析 ………………………………………………………………37
5.1 單根檢定   ……………………………………………………………………37
5.2 ARJI檢定  ……………………………………………………………………38
5.3 擴散與跳躍 ……………………………………………………………………41
第六章 結論……………………………………………………………………………43
參考文獻   ……………………………………………………………………………45









表目錄
表4-1 基本統計量 ……………………………………………………………………29
表4-2 非到期基本統計量 ……………………………………………………………30
表4-3 到期基本統計量 ………………………………………………………………31
表4-4 20200323 逐筆前基本統計量  ………………………………………………33
表4-5 20200323逐筆後基本統計量…………………………………………………34
表5-1各變數單根檢定結果 …………………………………………………………37
表5-2 GARCH模型 & ARJI模型實證結果 …………………………………………39
表5-3一般變異、跳躍變異與總變異之樣本平均值   ……………………………42

參考文獻
Andersen, T. G., & Todorov, V. (2010). [Jump risk–related article]. Journal of Finance, 65(2),  
Barclay, M. J., & Hendershott, T. (2003). [Price discovery in after hours trading]. Journal of Financial Markets, …  
Battalio, R., Jennings, R., & Saretto, A. (2014). [ETF expiry and liquidity traps]. Journal of Empirical Finance, …  
Black, F., & Scholes, M. (1973). The pricing of options and corporate liabilities. Journal of Political Economy, 81(3), 637 654.  
Black, F. (1975). Fact and fantasy in the use of options. Financial Analysts Journal, 31(4), 36 41.  
Boehmer, E., & Wu, J. (2008). [Information transmission in U.S. markets]. Review of Finance, …  
Chan, N., & Maheu, J. (2002). Conditional jump intensity: A new approach to modelling jumps in financial prices. Journal of Business & Economic Statistics, 20(3), 377 389.  
Chang, C. H., Hsieh, P. F., & Lai, H. W. (2009). [Option volume and future stock returns in Taiwan]. Pacific Basin Finance Journal, …  
Chen, H. (2023). Expiration day effects under day  and night session EGARCH: Evidence from TAIFEX. Asia Pacific Financial Markets, …  
Chen, L., & Yin, W. (2025). [Retail trading intensity around earnings announcements]. Journal of Behavioral Finance, …  
Chiu, C. L., Chang, T. H., Hsiao, I. F., & Chiou, D. S. (2024). Information spillover from after hours futures to next day cash markets. International Review of Economics & Finance, …  
Corwin, S. A., & Schultz, P. (2012). A simple way to estimate bid ask spreads from daily high and low prices. Journal of Finance, 67(2), 719 760.  
Cvetkovs, K., & Karcevs, A. E. (2019). ETF price deviations and arbitrage opportunities. Journal of Index Investing, …  
Dickey, D. A., & Fuller, W. A. (1979). Distribution of the estimators for autoregressive time series with a unit root. Journal of the American Statistical Association, 74(366), 427 431.  
Duong, H. (2025). Spoofing behaviour and order book dynamics in emerging markets. Journal of Financial Crime, …  
Griffin, J. M., Nardari, F., & Stulz, R. M. (2004). Are foreign investors price leaders or followers? Journal of Financial Economics, 74(3), 581 612.  
Gomber, P. (2024). Multi platform market design and price transparency. European Financial Management, …  
Kallinterakis, V., & Karaa, A. (2023). Night day price reversals and arbitrage dynamics. Review of Quantitative Finance and Accounting, …  
Kappou, K., Brooks, C., & Ward, C. W. (2010). Price discovery between S&P 500 constituents and ETFs overnight. Journal of Banking & Finance, 34(12), …  
Li, Y., Hsu, C., & Wang, T. (2025). Futures led price discovery: High frequency evidence from TAIFEX. Journal of Futures Markets, …  
Lin, E., & Chen, C. R. (2019). Information flow speed after micro structure reform: A five minute study. Journal of International Financial Markets, Institutions & Money, …  
Lin, Y. J., & Chang, J. R. (2024). Expiration day behaviour under dual sessions: Evidence from TAIFEX. Finance Research Letters, …  
Madhavan, A. (2012). Exchange traded funds: Market structure, pricing and trading mechanics. Annual Review of Financial Economics, 4, 189 208.  
Maharajan, R. (2025). AI based overnight signal extraction for price discovery. Journal of Financial Data Science, …  
Marshall, N. T., & Stephan, A. (2025). Speed of price adjustment after firm disclosures. Accounting & Finance, …  
Nguyen, V., & Phengpis, C. (2009). NAV deviations, ETF arbitrage and market efficiency. International Review of Financial Analysis, 18(4), 190 197.  
Pathak, T. (2025). Co location scandals and fairness in price formation: Evidence from NSE. Emerging Markets Review, …  
Phillips, P. C. B. (1986). Understanding spurious regressions in econometrics. Journal of Econometrics, 33(3), 311 340.  
Phillips, P. C. B., & Perron, P. (1988). Testing for a unit root in time series regression. Biometrika, 75(2), 335 346.  
Said, S. E., & Dickey, D. A. (1984). Testing for unit roots in autoregressive–moving average models. Journal of Business & Economic Statistics, 2(4), 369 379.  
Tosun, O. (2021). Contract design optimisation and after hours efficiency. Journal of Derivatives & Policy, …  
Vij, S., & Wan, C. (2020). High volatility expiry pressure and after hours price integration. Journal of Futures Markets, …