我國法令原則上禁止庫藏股的交易，但實務上相當盛行母公司透過本身可完全控制的子公司來買回母公司股票，從經濟實質的觀點來看，此一交易屬於實質的庫藏股交易。本論文之目的在探討我國「子公司買回母公司股票」之實質庫藏股交易的資訊內涵，是否和資訊假說的含意相符；其次驗證對於發生諸如借殼上市之經營權移轉情形的公司而言，「子公司買回母公司股票」交易和公司大股東出脫持股俾追求自利的經濟誘因是否有關。研究主題一共發展3個研究假說，主要擬瞭解我國子公司買進母公司股票之交易行為的資訊內涵，俾了解此交易傳遞之資訊本質，並評估市場反應是否和此資訊有關。第二個研究主題共發展4個假說，係探討借殼上市公司內部人在借殼成功後，是否會利用子公司買進其借殼時增加的持股或融資交易的股票，俾舒解資金壓力。此外，亦驗證交易過程中的股價行為，是否和內部人為極大其獲利水準而可能進行股價操縱的股價行為相同。 　　本論文是以84年至87年為研究期間，兩個主題的樣本係相互獨立。實證結果顯示：關於主題一－子公司買進母公司股票傳遞之資訊本質，本研究發現，平均而言，以子公司買進母公司股票當季與後續四季為基準，母公司在此交易的後續期間並無表現出優於預期的盈餘績效（相對於配對公司），而且在交易後亦無法觀察到分析師對母公司盈餘預測有顯著的正向修正（相對於配對公司）。另外，在控制內部人淨買入、公司規模、子公司購買股數、資金需求與產業趨勢等變數後，本研究並未發現子公司買進母公司股票時的股價反應，和母公司當季與下一季的異常正向盈餘績效有顯著的正向關係。總而言之，國內實務界雖然宣稱子公司買進母公司股票係傳遞看好母公司未來前景的訊息，但本研究的證據顯示，這項交易並非傳遞母公司未來盈餘變動的資訊，同時，此交易引發的正向市場反應，似乎並非來自投資人對母公司未來績效的預期，故從股價操縱的角度來看，此一股價效應可能來自於母公司刻意創造，但其實與母公司經濟實質無關的泡沫股價。 　　關於主題二－借殼上市公司透過子公司買進母公司股票之經濟誘因－的實證結果顯示：在控制其他因素後，借殼上市公司在借殼成功的後續期間，子公司買回母公司股票之交易發生後，相較於無此交易的配對公司而言，其內部人持股比例和融資餘額有顯著降低的情形。基於這類公司大股東有套現的特殊動機，故可合理懷疑，這項證據和內部人籍子公司買進母公司股票來出脫持股俾套現獲利有關。此外，分析結果發現：子公司買回母公司股票交易前有正向的股價效應，此一股價行為和內部人操縱下的股價型態相符，但交易後的股價水準則並未顯著低於交易前的股價。吾人推測此一結果，可能肇因於投資人受到誤導，以致本研究未能觀察到後續反轉下跌的股價行為。 　　整體來看，子公司買進母公司股票之交易並非如實務界所宣稱，係傳遞公司未來前景較佳之資訊，反而可能和內線交易與股價操縱有關。因此，本研究支持目前財政部擬議中嚴格禁止或限制母子公司交叉持股的方案，此外，本研究應有助於投資人做投資決策時，對於母子公司交叉持股的本質有所了解。 / Companies are not allowed to buy back any of its own shares except in some special cases according to the Company Laws in Taiwan, but the practice that parent companies repurchase their own stocks through subsidiary companies (which is so called parent-subsidiary cross-holding transactions) is popular currently. From the economic substance view, the parent and subsidiary companies are treated as a single economic entity, and the parents' stocks held by their subsidiaries are treasury stocks in substance. Since the authorities and investors are all concerned with this issue, and previous work in this area is limited, this thesis investigates the information content and economic incentives of these transactions. Specifically, there are two research questions in this dissertation. The first is whether the transactions that subsidiary companies purchase parents' stocks convey any information about parents' future prospects? And whether the stock price reactions around announcement dates are positively related to parents' subsequent earnings changes. The next question is whether these transactions are motivated by insiders' intent to sell their own shares for self-interest? And is the stock price behavior around announcement dates consistent with the explanation that stock price is manipulated? 　　The findings of the first research question are as follows. First, I find no evidence that there are positive aggregate unexpected quarterly earnings relative to a size and industry-matched control sample over the announcement quarter and subsequent four ones. Second, the significant evidence that there are positive revisions of earnings forecasts by analysts around announcement dates also cannot be found. Finally, a regression analysis shows that there is no significant positive relation between positive abnormal stock returns at announcement and positive quarterly earnings changes over quarter 0 and quarter I. In summary, the results are not consistent with the implication of the information hypothesis. I find no systematic support for the firms' claims that the meanings behind parent-subsidiary cross holding transactions are to convey managements' private information about parents' favorable prospects. 　　With regard to the second research question, I focus primarily on those listed companies whose controls are obtained by unlisted companies through purchasing outstanding shares in the open market, these companies are not included in the sample of the first question. As predicted, I find systematic support for the hypotheses that shares owned by insiders and companies' margins decrease relative to control groups immediately following parent-subsidiary cross holding transactions announcements and that abnormal stock returns before and at announcements are positive. However, I show that the abnormal stock returns after the announcements do not fall bellow those before announcement dates significantly. In summary, the findings indicate that these transactions maybe a means for informed insiders to sell their own shares at manipulated high prices in order to alleviate the huge capital pressure by self-interest. 　　Overall, the transactions that subsidiary companies purchase parents' stocks are more likely to have relation to insider transactions and stock price manipulation, rather than convey favorable information about parents' prospects as what firms claim. Therefore, this thesis supports an act of prohibiting or limiting parent-subsidiary cross holdings proposed by the Ministry of Finance strictly. Besides, the results in this thesis may be helpful for investors to understand the nature of parent-subsidiary cross holdings when they make investment decisions.

庫藏股制度於國外已經行之有年，特別是美國，早在1960年代即有庫藏股交易。我國上市上櫃股票可合法買回自己公司股票制度始於2000年所通過的上市上櫃公司買回本公司股份辦法，除了希望藉以提振東南亞金融危機中不振的股市表現，也賦予企業多一種能向投資人傳遞訊息的管道。本研究基於股票買回的各項假說，希望探究公司股票買回宣告的動機還有事件宣告後對於公司股價的影響效果，最後並檢視公司的宣告決策是否會受到前次購回宣告的表現所影響。 在假說驗證方面，以2000至2004年上市公司832家、上櫃公司236家曾經宣告買回自家股票的公司為研究樣本，輔以選出與買回樣本同時間的相同數量「未買回樣本」，透過1-way ANOVA測試與鑑別分析探討影響公司宣告股票買回的因素符合哪些假說的推論結果。此外，依據各項假說，進ㄧ步利用1-way ANOVA測試和複迴歸分析找出與公司事件宣告後股價異常報酬率之間的關係，接著以1-way ANOVA檢視最後一部份研究主題。經由上述的實證分析，本研究獲得以下主要結論： 1.公司買回自家的股票平均而言對於股價有正面的影響效果。 2.以上市公司而言，影響公司宣告股票買回的假說包括了：自由現金流量假說、個人所得稅節稅假說與財務槓桿假說。以上櫃公司而言則包括：自由現金流量假說、個人所得稅節稅假說與資訊信號假說。 3.以上市公司而言，股票買回宣告後，股價的正向異常報酬與：自由現金流量假說、管理者的誘因假說呈現相關關係。上櫃公司則包括：自由現金流量假說。 4.本次是否再度宣告股票買回會受前次買回的正向累積異常報酬率所影響。 / Stock price reactions on stock repurchase announcements among publicly traded corporations have been widely studied over 40 years in United States. Several hypothesizes were established and provide logical reasons for why corporations buy back their own stocks. References show large proportion of positive effect on post-announcement stock price after stock repurchase announcements from empirical researches. It is since 2000 that corporations are allowed to repurchase their own stocks legally in Taiwan. Although many efforts were put in surveying the post-announcement stock price reactions, few focused on examining how corporations’ repurchase decision would be influenced by different hypothesizes. Moreover, whether or not previous repurchase outcome may affect future repurchase judgments met contradict conclusion from recent studies. This research aims on exploring the two main topics which form 7 hypothesize. Data collecting from Taiwan Economic Journal (TEJ) database, Commercial Times and Economic Daily News contains publicly traded corporations had ever made stock repurchase announcement except financial and government institutions. The period of the study was from Aug. 6, 2000 to Jul. 31, 2004 and total number of announced corporations included was 832 listed on the Taiwan Security Exchange with 236 listed on the OTC separately. Hypothesis 1 to 6 were tested both from individual model using 1-way ANOVA Analysis and integrated model using Discriminate Analysis on Hypothesis 1 to 5, Event Study Analysis on Hypothesis 6. Hypothesis 7 was also verified with 1-way ANOVA Analysis. The empirical results show supports in several hypothesize which reflect corporations did concern free cash flow amount, leverage level, stock underpriced and dividend payout level issues when making repurchase announcement. The stock price after announcement has certain degree of moving correlation toward the same direction as free cash flow level and managerial stockholding level. Finally, previous repurchase outcome was found holding opposite relation with future repurchase judgments.