Background. Due to prevailing technological development, telecommunication and computers have become very advanced. This has had a tremendous effect on the financial markets as well, various facilitating financial means have become much more common. One of such is passively managed index funds which does not only use index as a benchmark but also trade the stocks in the index. Thus, guaranteeing the fund a return equal to the market return and to a lower cost than an equally good actively managed fund. Index funds have in recent times increased in popularity, which has left its mark. The price of the stocks included to and excluded from an index has been observed to respectively increase and decrease in value.   Research on the price effects caused by index revision, or the effects that inclusion and exclusion have on the price of underlying shares, has been around since the 1980s. In the literature, it is generally accepted that inclusion to an index results in a positive price development, while exclusion results in a negative price development. However, the literature does not agree on whether the price effects are long-term or short-term. The disagreement began with the first studies in the field, where one author found that the price effects were long-lasting, even permanent. The others, on the other hand, found that the price effects were short-lived and returned to their original value when the trading ceased. Subsequent research is equally inconsistent. Some studies have found temporary changes, and some have found permanent ones. In this uncertainty, different theories of explanation have also been presented for the different outcomes, but these do not agree either. Objectives. To bring some clarity to the problems within the literature, the purpose of this study is to investigate the stock price effects from the reconstruction of a Swedish market index, with consideration of whether the effects are temporary or permanent. Methods. This study applied the event study methodology and the market model to examine the abnormal return found around the announcement day and the changing day. The study is based on 195 stocks that were included to and excluded from the OMX Stockholm Benchmark Index between the years 2009 and 2019.  Results. This study did not find any statistically significant price change in the period before the announcement date. However, there were indications that the announcement day did have a positive effect on the included stocks and a negative effect on the excluded stocks. But the time after the announcement day and prior to the changing day did not show any statistically significant price changes. The changing day and the period after were both found to be negatively significant for inclusions. Thus, indicating a negative price effect on the day of inclusion and the period that followed. These results are consistent with previous studies that have found a price drop on the changing day and the following period. A further test of the relationship between the abnormal return found on the announcement day and the changing day revealed that the price increase was concentrated to the announcement day. A possible explanation for this outcome may be that index funds that trade on the Swedish exchange have recognized the opportunity to trade closer to the announcement day without incurring any losses and acted accordingly. Regarding the exclusions, the changing days were not found to be statistically significant. Neither did the period following the changing day show any statistical significance. This result could be due to a delayed reaction to the changing day, given that this group showed a slow reaction to the announcement day as well. Both the announcement day and the day after the announcement day were statistically significant at the 1% level. Other possible causes for the deviating results are errors in execution or data. Conclusions. The result of this study is consistent with other studies that find a temporary price reaction to the index reconstruction.

本研究探討台灣50指數成分股變動造成的異常報酬與超額交易量現象，提供投資人作為選股參考，由於台灣50指數編製依據為總市值，僅使用公開資訊，可以用來檢驗效率市場假說。不同於過去研究的貢獻，本文考慮金融資產報酬的波動群聚現象及模型係數的不穩定性，分別使用卡爾曼濾嘴法及GARCH(1,1) 模型修正，以期獲得較穩健的實證結果。本研究結果發現，新增股及剔除股在宣告日的下一個交易日及生效日的前一個交易日有顯著異常報酬；新增股的累積異常報酬可持續大約兩個工作星期，且新增股有持續的超額交易量情形；而剔除股僅在生效日及生效日的前一個交易日有超額交易量。上述實證結果顯示投資人可利用此公開資訊，獲得異常報酬，不支持效率市場假說。 / We study the price and volume effects following the Taiwan 50 Index revisions. Since the decision to include (exclude) a stock in (from) the index is based merely on market capitalization, it provides a unique opportunity to test the efficient market hypothesis. Our contribution to the existing literature lies in the fact that we employ the Kalman filter approach and the GARCH(1,1) model to allow the parameters instability and time-varying residual variance, and hence to produce more robust empirical results. The results suggest that the newly added stocks and the currently dropped stocks experience significant price changes right after the announcement day and right before the change day; the abnormal returns of the newly added stocks last about two weeks, these stocks also exhibit persistently excessive trading volume while the currently dropped stocks only have excessive trading volume on and right before the change day. These results indicate that investors can earn abnormal returns through public information, and invalidate the efficient market hypothesis.

本研究旨在探討臺灣中型100 指數成分股調整事件之價量關係以及新聞報導效果。研究樣本分為純粹納入股、向下納入股、純粹剔除股以及向上剔除股，並分別以成份股調整之宣告日與執行日作為事件日，採用事件研究法，分析事件前後之價量變化，同時，進一步探討，宣告日至執行日期間，新聞報導對於成分股調整效果之影響。實證結果顯示，純粹納入(剔除)股於宣告日當天以及執行日前一天具有顯著的正(負)向異常報酬，且短期內皆有反轉的現象產生；向下納入(向上剔除)股，在宣告日與執行日前一交易日享有負(正)向異常報酬，且執行日後五日旋即反轉；此外，此四類個股於宣告日以及執行日附近，皆有異常週轉率生。在新聞效果部分，本研究將純粹納入(剔除)股分為有利多(空)新聞之組別以及無新聞之組別，探討其異常報酬現象。實證結果發現，有利多(空)新聞之純粹納入(剔除)股，相較於無新聞之純粹納入(剔除)股享有較高的正(負)向異常報酬率，顯見在成分股調整事件中，新聞報導亦會影響股價表現。 / This paper examines the effects of Taiwan mid-cap 100 index revision and news coverage on stock price and volume. Using event study method, the sample of this study is divided into four groups: pure additions, downward additions, pure deletions, upward deletions to analyze the changes of stock prices and volume on the announcement day and effective day. Furthermore, the effects of news coverage between announcement day and effective day are also investigated. Results of our analysis suggests that there are significantly positive (negative) abnormal returns (ARs) for pure additions (deletions) on announcement day and the day before effective day but transitory. For downward additions and upward deletions, there are significantly negative ARs for the former and significantly positive ARs for the latter. However, both ARs reverse within five days. Besides, there are abnormal volumes in the entire sample. As for the effects of news coverage, pure additions (deletions) are divided into two groups to examine the existence of ARs, one group with bullish (bearish) news and one group without. Results show that pure additions (deletions) with bullish (bearish) news have higher ARs than those without news coverage, supporting our hypothesis that in the event of index revision, news coverage do affect stock prices