none

This dissertation studies the reaction of trading behavior of investors, especially institutional investors, to the public information in Taiwan. Two kinds of public information are chosen in this dissertation. One is open-ending closed-end funds under the regulation set up by Taiwan authority. The other is the change of stocks in MSCI Taiwan Index that is decided by Morgan Stanley Capital International (MSCI), a well-known foreign institution in constructing various indices. Consistent with earlier studies using U.S. data, our results show that open-ending is a wealth-enhancing event for shareholders. We also provide evidence of the existence of noise traders in the closed-end fund market. The evidence is derived from the trading behavior of domestic institutional investors and small individual investors, who ignore price discounts when open-ending is imminent. The trading by noise traders impedes price adjustments to the discounts, and provides profit opportunities to arbitragers. Furthermore, we show that foreign investors gain considerable wealth, largely at the expense of domestic institutional investors and small individual investors, in the open-ending process. On average, their gains account for 30% of the total gains associated with open-ending, or NT$562 millions per case. On the issue of the change of stocks in MSCI Taiwan Index, we find that MSCI prefers to select the securities with good performance, high liquidity, and large firm size into MSCI Taiwan Index while tends to drop the securities with poor performance, lower liquidity, and small firm size from MSCI Taiwan Index. Besides, consistent with the previous studies, prices increase (decrease) significantly for stocks added to (deleted from) the MSCI Taiwan Index after the announcement date. As well as the deletions, the price decreases for unchanging stocks after the announcement date. However, there is no evidence to find that foreign investors have information advantage in MSCI news over domestic investors. Foreign investors increase (decrease) their holdings on stocks included in (excluded from) the MSCI Taiwan Index after the announcement date. Moreover, price pressure hypothesis is not supported. Visibility hypothesis, information content hypothesis, downward sloping demand curves hypothesis are supported. Finally, for additions and deletions, the market-adjusted returns are driven by the contemporaneous excess buy of foreign investors and the contemporaneous excess sells of domestic corporations and individuals.

本文以摩根台指與台灣50指數成分股調整的股價效應做比較，兩者對於成分股調整的宣告及生效是否存在異常報酬，而異常報酬的不同是否與其指數編制方式有關。實證結果發現摩根台指新增股具有 顯著正報酬、剔除股具有顯著負報酬，且在宣告日二十天後價格呈現反轉，符合價格壓力假說。而台灣50指數新增股異常報酬則不顯著異於零，兩種指數的新增股在宣告成份股調整後皆有超額成交量、流動性持續增加。再者摩根台指的宣告效果比台灣50指數強，且透過加權的方式較能表現出指數成份股調整所帶來的現貨價格影響。 / The study examines both the price and volume effect of stock additions or deletions on both the MSCI Taiwan index and Taiwan50 index. We document significant abnormal returns of stock additions and deletions for the MSCI Taiwan Index both on the announced period or on the effective period. In addition, we also find a significant abnormal return of stock deletions for Taiwan 50 Index either announced period or the effective period. While we do not find any significant abnormal return of stock additions. Further more, both the announced date effect and the effective date effect for MSCI Taiwan Index are stronger than those for Taiwan 50 Index. Our results support the price pressure hypothesis.

本研究針對臺灣摩根指數的成分股進行分析，研究樣本期間從2008年至2010年，合計三個年度，正好歷經景氣的一個多空循環週期。本研究利用技術指標作為判讀多空的工具，技術指標包含價與量的技術分析工具，價格的技術指標有趨勢指標MA、擺盪指標KD與MACD，量的技術指標則是OBV。並利用優化的方式挑選出合適的參數值。本研究的風險控管則是控管個股的偏離程度，當允許的偏離程度愈大，模型便愈能區別出強勢股與弱勢股，風險的衡量指標則是採用年化追蹤誤差值來衡量，本研究設定的限制條件為最大累積年化追蹤誤差值不得超越6%。 實證結果發現，當模組的模型年化追蹤誤差值設定愈大，個股的偏離程度就愈大，模組的報酬表現就愈佳，但同樣的風險也愈大，即年化追蹤誤差值愈大。當模型年化追蹤誤差值設定在24%，並搭配MA、MACD與OBV三個技術指標得到的績效最佳，同時亦能夠將風險控制在設定的6%水準之下。 / This study analyzed the component stocks in MSCI Taiwan Index. The analyzed data from 2008 to 2010 was exactly an economic cycle. The study was based on technical analysis, including price and volume to judge that the price was bullish or bearish. The price technical analysis included Moving Average (MA), Stochastic Line (KD) and Moving Average Convergence and Divergence (MACD). The volume technical analysis was On Balance Volume (OBV). The study used the method of optimization to choose the best parameter of each technical analysis. The risk control was to limit the bias of each stock. When the bias of each stock was larger, the model could easily distinguish the stock was bullish or bearish. The risk indicator was annual tracking error limited to 6% in the study. The empirical results showed that the larger the model annual tracking error set, the large bias the stock show, and the outperformance of the return. But with the performance of the return larger, the risk of tracking error was also getting larger. When the model annual tracking error set to 24%, and utilized MA, MACD and OBV would get the best performance and the risk of annual tracking error was under 6%.