[FIRST DRAFT CURRENTLY COMPLETING DRAFT WITH DATA TO 2002] DO FIRMS BENEFIT FROM STOCK REPURCHASES?

Transcription

1 [FIRST DRAFT CURRENTLY COMPLETING DRAFT WITH DATA TO 2002] DO FIRMS BENEFIT FROM STOCK REPURCHASES? Abstract Over the past few years, many firms have announced significant numbers of stock repurchases. The overwhelming reason given for stock repurchase announcements has been to reverse a trend of declining stock prices. Several studies have provided conclusive proof that in general stocks experience abnormal returns in the order of 3% immediately after repurchase announcements. This paper investigates whether there is a long-term difference in abnormal returns for firms that repurchase stocks versus those that do not repurchase stock, given a stock repurchase announcement. An examination of stock repurchase announcements during the period 1994 to 1998 [SECOND DRAFT INCLUDES DATA TO 2002] indicate that stock prices do not adjust to pre announcement levels if the repurchases are not carried through. Also, there is no statistically significant difference in abnormal returns in the postannouncement period for firms that repurchase and those that do not. These results imply that firms are not punished for failing to carry through on repurchase announcements, and that investors only focus on the repurchase announcement. I. INTRODUCTION Stock repurchase refers to the reversal of equity offerings. The issue of stock repurchase has been the focus of much theoretical and practical research and is often touted as one of the classic methods to raise a company's stock price. By reducing the number of shares outstanding, the interaction of demand and supply is expected to cause the stock price to float upward. However, many stock repurchases don t shrink the total equity pool of the firm. Instead, repurchases are often used to offset employee stock 1

2 options and prevent shareholder dilution. 1 Further, the surge in the number of stock repurchase announcements over the last decade has raised the question of whether repurchase transactions offer firms long-term benefits. II. RELATED RESEARCH Earlier empirical research concentrated on four issues: First, the reason for and effect of stock repurchase announcements and the abnormal returns associated with such announcements. Lakonishok, Josef and Vermaelen (1990) examine stock repurchase data for repurchase tender offers that occurred between 1962 and 1986 and presents evidence of abnormal returns of more than 9% over a period of less than one week. Vermaelen (1981) and Ikenberry, Lakonishok and Vermaelen (1995) find abnormal returns of approximately 3% over a two-day announcement period. In addition, Ikenberry, Lakonishok and Vermaelen (1995) hypothesize that stock repurchases primarily serve as a signaling mechanism and thereby provides new information. The signaling hypothesis led to an investigation of the long-run performance of companies announcing stock repurchases. Using a sample of 1,239 openmarket repurchases announced between 1980 and 1990, they show that investing in companies that announce stock repurchases results in 12.4% abnormal returns over 4 years. However their examination was done without regard to whether the programs were actually completed. Early work by Comment and Jarrell (1991) investigates the signaling hypothesis as it relates to stock repurchase announcements. They examine the three main methods of repurchasing stock tender offers, Dutch auction, and open market repurchases and 1 Farrell, Christopher. The Buyback Boom is Mostly a Boon, BusinessWeek April 13, 1998, page

3 conclude that fixed-price tender offers generally signal the most information to investors and open-market repurchase the least. The signaling hypothesis would also imply that firms time repurchases to gain maximum benefits from repurchasing when their stock price is most undervalued. Comment and Jarrell (1991) also find evidence that firms tend to announce open-market repurchase plans following a decline in their share price, when their stock is more likely to be undervalued. Stephens and Weisbach (1998) examine 450 repurchase programs between 1981 to 1990 and find that on average firms acquired 74 to 82 percent of the shares announced as repurchase targets within three years of the repurchase announcement. They find that share repurchases are negatively related to prior stock price performance, suggesting that firms increase their purchasing depending on the degree of perceived undervaluation. They also find evidence of consistency with liquidity theory as the level of repurchases within their sample was positively related to the levels of cash flow. Persons (1997) presents an asymmetric information model of share repurchases and finds that managers in the model repurchase shares at a premium above the postrepurchase share value transferring wealth from shareholders who do not tender to those who do in order to signal that the firm is undervalued. The model also demonstrates that share repurchases are more effective, i.e. less costly to the signaler, than many other possible signals (such as dividends) as repurchases dominate these as a signaling device in this model. Barth and Kasznik (1999) compares a sample of firms that made repurchase announcements between 1990 and 1994 with a sample of all firms that have 1992 Compustat data (the median year) and do not announce any share repurchase between 3

4 1985 and They find that firms with more intangible assets are more likely to repurchase shares and have more positive repurchase announcement returns. In addition, they also find that idle cash is positively related to repurchase likelihood and negatively related to announcement returns. The second body of research focuses on the effects of stock repurchases on rival firms. Research focusing on the impact on rival firms demonstrates that rival firms on average realize insignificant announcement period and long-term abnormal returns. Hertzel (1991) provides evidence showing that the information regarding repurchase announcements was primarily firm specific. Using a sample of 134 repurchase announcements between 1970 and 1984, Hertzel finds that rival firms on average realize insignificant announcement period abnormal returns; while negative but statistically insignificant rival stock performance is detected over longer intervals surrounding the announcement period for rival firms. Repurchase announcements therefore differ in impact on rival firms than that of other factors such as acquisitions that have been shown to result in abnormal returns. 2 Third, current research also focuses on firm earnings changes surrounding stock repurchases. Nohel and Tarhan (1998) examine tender share repurchases to differentiate between the information signaling and free cash flow hypothesis and conclude that the positive investor reaction to repurchases is best explained by the free cash flow hypothesis. Their sample consists of 282 announcements of both fixed-price and Dutch auction tender offers with the primary focus on operating performance changes surrounding repurchases. Nohel and Tarhan s research finds that operating performance 4

5 following repurchases improves only in low-growth firms, and that the gains are generated by more efficient utilization of assets, and asset sales, rather than improved growth opportunities. They argue that repurchases do not appear to be pure financial transactions meant to change the firm s capital structure but are part of a restructuring package meant to shrink the assets of the firm. Lie and McConnell (1998) examine announcements between September 1981 and December 1994 and test whether the earnings improvement following fixed-price self-tender offers is greater than that following Dutch auction self-tender offers. The analysis investigates the changes in earnings from before to after tender offers and changes in Value Line forecasts of earnings. Lie and McConnell find evidence that earnings improve following both types of self-tender offers, but find no statistically significant difference in earnings improvement between the two types of offers. Fourth, other research concentrates on whether or not firms that announce repurchases actually carried through on the announcements. Kracher and Johnson (1997) cites several examples where CEOs later admitted that they had no intention of purchasing stock as per the announced plans and showed that most firms do not repurchase anywhere near the amount of stocks they had earlier indicated. Kracher and Johnson (1997) examine the Securities and Exchange Commission (SEC) rule 10b-5 which implies that misleading repurchase announcements are fraudulent, but acknowledges that firms do not perceive that the SEC considers misleading announcements problematic. As a result, they advocate that since firms create distrust in the investment community by making false announcements, changes should be 2 For example, Song, Moon H. and Ralph A. Walkling (2000) find that rivals of initial acquisition targets earn abnormal returns because of the increased probability that they will be targets themselves. Akhigbe 5

6 forthcoming on reporting procedures to force the public dissemination of information regarding actual repurchases compared to announcements on a quarterly basis. This study differs from earlier research by examining the long-run impact on stock prices resulting from repurchasing or not repurchasing stock after announcing an intention to repurchase. The investigation focuses on whether or not there is any difference in abnormal returns related to the level of repurchases. The sample of firms that made a repurchase announcement during January 1994 to December 1998 [update to 2002] are arranged into quintiles based on the level of repurchases actually made. The theory is that if the stock market is efficient, then abnormal returns associated with repurchase announcements should reverse if firms do not repurchase stock as announced. Hence, the extent of repurchases across each group should result in statistically significant differences in post-announcement abnormal returns. Mean cumulative abnormal returns for the sample of firms in this study indicate that the level of repurchases does not influence the stock price over a subsequent twentysix week period. A derived regression equation that extends the analysis to include firm size, the size of the repurchase target as a percent of total stock outstanding, and industry classification also yield statistically insignificant coefficients. With several earlier works documenting the positive abnormal returns from stock repurchase announcements and the wealth implications for investors the importance of long-term benefits takes on significance. If stock repurchases can provide significant positive abnormal gains, then corporate governance would imply that stock repurchases should be actively pursued as a means of generating substantial wealth. and Whyte (Working paper) find that rivals gain upon termination of a merger. 6

7 Conversely, if markets are efficient, reacting to public information, then the movements of the level of outstanding stock after repurchase announcements, should provide information to influence the stock price for each firm. If mean abnormal returns are not statistically different between firms repurchasing and not repurchasing, then management has no financial incentive to repurchase stocks and the signaling effect from the announcement is sufficient action. III. HYPOTHESIS This paper uses data that is strikingly similar to Kracher (1997), in that the percentage of firms actually repurchasing stock subsequent to an announcement is small compared to the percentage of firms not repurchasing after announcing a repurchase plan. Fifty percent of firms purchase less than 20% of the indicated repurchase target within twenty-six weeks after an announcement to repurchase, and 22% of firms repurchase 80% or more of the indicated target in the same time period. Further, according to Stephens and Weisbach (1998) and Comment and Jarrell (1991), the signaling explanation of repurchases predicts that the event-day returns on announcement of a repurchase program should be related to the information contained in the announcement. If this hypothesis is true, then the subsequent level of repurchases should not be significant in determining the extent of abnormal returns surrounding the announcement. Given these issues, we calculate the actual level of repurchases in a twenty-six week period starting one week after the announcement date and compares this repurchase activity to the abnormal returns of the stock. Abnormal returns are examined 7

8 for two time periods; a one-week period starting on the announcement date, and a twenty-six week period starting one week after the announcement. The issue is whether stock repurchase announcements are only signaling events, with the market responding only in the short-term (immediately following the announcement) or whether subsequent repurchase activity provides additional incentives and hence further market response. Since the market is assumed to be efficient, the expectation is that stock prices should reflect the level of stock repurchases, given that an abnormal return is evident upon the announcement. This leads to the first hypothesis: H1: Stock prices in the post announcement period will reflect the level of stock repurchase. Alternatively, if the market reaction to the repurchase announcement is in response to a signal of under-pricing, then stock prices may not adjust to reflect the extent of actual stock repurchases. Hence, firms not repurchasing will not be penalized and those that follow through on the announcement to repurchase may not experience further stock-price increases. In such an event, firms would be able to reap permanent benefits from simply announcing stock repurchase plans. This leads to an alternate hypothesis: H2: Abnormal returns as a result of repurchase announcements are permanent. We next extend our investigation to ascertain whether or not the percent of shares actually repurchased, firm size, the size of the repurchase target as a percent of total stock outstanding, and industry classification are statistically significant indicators of abnormal returns for firms making repurchase announcements. 8

9 To test the hypothesis, market adjusted cumulative abnormal returns (the S&P 500 used as the market reference) are calculated for each stock for a one-week period starting on the announcement date and a twenty-six week period starting one week after the stock repurchase announcement. Next, the average cumulative abnormal returns are calculated and the means of each quintile compared for statistical difference. The findings contained in this paper have significant implications. Since it is generally believed that the announcement effect is in reaction to expected future wealth transfer to stockholders or a signaling effect of under-valuation, then lack of action on the announced repurchase should result in price reversion subsequent to abnormal returns surrounding the announcement. Several works and the sample used in this paper find that a large percentage of announced repurchases do not materialize and hence the sample should indicate significant numbers of stock price reversals sometime after the announcement. III. Sample The sample consists of stock repurchase announcements obtained from the Wall Street Journal Index between January 1994 and December 1998 (updated to 2002) by utilizing various keywords, including stock repurchases and stock announcements. (Although several announcements relating to repurchase of preference shares were recorded in the review period, the scope of this paper is limited to repurchases of common stock only.) A total of 215 repurchase announcements are found, with the final sample being reduced to 138. Several firms were taken over or merged with other firms and were thus eliminated to avoid survivorship bias, while detailed stock information was unavailable for 58 firms thereby preventing proper analysis. Of the final sample, 69 9

10 (50%) purchased less than 20%, 17 (12%) purchased between 21% and 40%, 14 (10%) purchased between 41% and 60%, 8 (6%) purchased between 61% and 80%, and 30 (22%) purchased more than 80% of the announced repurchase target. (See Table 1 & 2) The firms that repurchased typically repurchased close to 100% of the announced amount, while the firms that were classified as not repurchasing indicate repurchases of less than 20 percent, with 41 or 29.7 percent of the firms in the usable sample recording no purchases within the review period. The skew of the data is indicative of either a willingness to follow through on the announced repurchases or management basically ignoring the announcements. The Wall Street Journal and daily stock prices surrounding the announcements indicate positive short-term abnormal returns on days 0 and +1. In addition, the data is free of day-of-the-week skew as the announcements were fairly evenly spread across all five trading days. Table 1 gives the sample breakdown across years. Stock prices were obtained from CRSP and Standard and Poor s Compustat (some data also taken from Compact Disclosure, which uses SEC 10K and 10Q filings in its database) provided outstanding stock information and was supplemented with information regarding actual repurchases from the Wall Street Journal. Tables 1 and 2 partitions the sample by year with the actual announcements in Table 1 and the percentage distribution in Table 2. The mean announced size of the repurchase program is approximately 5.44 percent of the firm s total shares outstanding at the announcement date and the median is approximately 3.75 percent of the firm s shares outstanding. The number of repurchase programs announced per year between 1994 and 1997 is roughly consistent with a spike in 1998 when 49 announcements are 10

11 included in the final sample. Additionally, the programs are generally increasing in size during the sample period. (These sample characteristics are very similar to the sample used by Stephens and Weisbach (1998) although their time period was different.) The initial sample of 215 stock-repurchase announcements appearing in the Wall Street Journal was reduced to 138 due to several factors. In several cases, either no information was available regarding outstanding stock or weekly stock prices were unavailable. Several firms announced that they had suspended their repurchase plans for reasons unrelated to the stock price. For example, the Wall Street Journal (October 21, 1996) announced that several big U.S. companies abruptly rescinded repurchase programs. The programs were rescinded in response to a March 1996 Securities and Exchange Commission staff bulletin that effectively restricted stock repurchases by companies that make acquisitions and account for them using the pooling-of-interest accounting methodology. Of the total sample of 138 announcements, only a few were expansions of existing stock repurchase plans. Several companies did not have sufficient historical stock price information to facilitate the computation of the market-adjusted returns. For example, NationsBank Corp. merged with Bank of America after announcing a stock repurchase plan and hence subsequent long-term analysis is impossible. IV. Methodology Stephens and Weisbach (1998) points out that share repurchases could be neither observed at the time the transaction occurs or directly measured afterward. They use four methods as proxies for the actual number of shares repurchased by firms subsequent to 11

12 the announcement of an open market repurchase programs. The four methods are; 1) monthly decreases in the firm s shares outstanding from CRSP, 2) quarterly decreases in the firm s shares outstanding from Compustat, 3) dollars spent reacquiring firm stock using minimum and average quarterly purchase price, and 4) quarterly increases in the dollar value of treasury stock divided by minimum and average prices during the quarter. Empirically, the four methods produced results that were basically similar. This paper uses quarterly decreases in the firm s shares outstanding from Compustat as a proxy for the actual number of shares repurchased by firms subsequent to the announcement of an open-market repurchase program. To the extent that the firm both repurchases and distributes shares within the same month, this proxy for repurchased shares understates the true quantity of repurchases. However, this understatement of the actual repurchases is a mute point since the new issues effectively negates the benefits of the repurchases in respects to per share earnings. The number of shares to be repurchased is indicated in the WSJ announcements for most of the firms in the sample. In instances where only a dollar value was indicated for repurchases, the number of shares to be repurchased is determined by dividing the dollar value announced by the stock price on the last trading date prior to the announcement. The sample is divided into quintiles depending on the extent of share repurchases. The upper quintile (81% - >100% repurchases) and lower quintile (0% - 20% repurchases) accounts for the bulk of the categorizations with the lower quintile accounting for 50% and the upper quintile 22%. 12

13 To measure the extent of repurchases, the stock outstanding at the end of the review period was subtracted from the stock outstanding at the announcement date. In those cases where the stock outstanding was unavailable for some firms on the announcement date, the analysis uses data from the closest available date in determining the quintile to which each firm would be categorized. Long-term Abnormal Returns The first step in the analysis of the impact of actual stock repurchase on the level of abnormal returns requires computing the market adjusted cumulative abnormal returns (CAR) for the sample of 138 firms over the 26-week period starting one week after the repurchase announcement. The purpose of this analysis is to determine if the level of repurchases would influence the long-term cumulative abnormal returns. The exclusion of stock price data for the week of the announcement eliminates any bias from the short-term announcement effect and therefore indicates whether subsequent purchases have an impact on the average cumulative abnormal returns. Standard event-study procedures as used by Comment and Jarrell (1991) and Stephens and Weisbach (1998) were used to calculate the abnormal returns. The abnormal return in any given period is the market model residual, which is the difference between the stock s actual return and the predicted return based on the market return for that period. Hence the market adjusted abnormal returns were calculated as: ARij = RSij - RMi Where ARij is the abnormal return for firm j on day i. RSij is the actual return for firm j on day i. 13

14 And RMi is the return on the S&P 500 Index on day i. Next, the 26-week cumulative abnormal returns for each firm is calculated as: 26-Week CARij = ΣARij, for weeks i = 1, 2, 3, 26 where the announcement occurs in week 0. Cumulative abnormal returns were then averaged over the six-month period starting in week +1 to obtain the six-month cumulative average abnormal returns as: 26-Week ACAR = (ΣCARj)/n for all firms j = 1,2,..n The average cumulative abnormal returns are then compared for statistical difference between the means in each quintile. Statistical significance in the difference in means would indicate that cumulative abnormal returns are related to the level of repurchases undertaken during the 26-weeks. Barber and Lyon (1997) shows that using standard parametric tests for long run abnormal returns around firm specific events could draw biased inferences. Their study documents that test statistics based on abnormal returns using a reference portfolio, such as a market index, are misspecified with empirical rejection rates exceeding theoretical rejection rates. The recommended methodology is to correct for potential misspecification by using a matching sample of firms to compute abnormal returns. Despite Barber and Lyon s findings, this paper uses a market-adjusted methodology for computing abnormal returns based on two reasons. One, the findings referred to are based on abnormal stock returns over a one to five-year period while this study concentrates on a shorter reference period of 26 weeks. Two, the findings of both the 26 week and five-day tests are consistent with each other and hence do not indicate the existence of bias. 14

15 Short-term Abnormal Returns Next, the analysis is extended to the market adjusted returns for the sample over a fiveday trading period starting on the announcement date in the Wall Street Journal. By examining this shorter interval, the analysis investigates whether the abnormal returns just after the announcement ultimately impact the subsequent levels of repurchases. (The announcement date in the Wall Street Journal was included since the publication date would be normally a trading date and investors have the opportunity to respond to such announcements on the same date.) The market adjusted abnormal returns are calculated as in Equation (1) above. Next, the five-day cumulative abnormal returns for each firm is calculated as: 5-Day CARij = ΣARij, for days i = 0, 1, 2, 3, 4 where the announcement day is day 0. Cumulative abnormal returns were then averaged over the five-day period starting on the announcement date to obtain the five-day cumulative average abnormal returns as: 5-Day CAR = (ΣCARj)/n for all firms j = 1,2,..n The average cumulative abnormal returns are then compared for statistical difference between the means in each quintile. Statistical significance in the difference in the means would indicate that abnormal return is related to the level of repurchases undertaken during the five-day period. Total Abnormal Return Third, the total abnormal return, from announcement date to week +26, was calculated as: 15

16 5 Day ACARij + 26 Week ACARij Where the 5-day ACAR and 26-week ACAR were derived as in the respective sections above. A means test and one-way ANOVA to test for statistical significance between differences of the means of the quintiles was done on the ACARs. Cross-Sectional Analysis The third aspect of this research derived a cross-sectional regression model to predict the level of CAR as follows: CARj = β0j + β1jsrepo + β2jindus + β3jfsize + β3jasize Where; CARj is the abnormal return for firm j, SREPO is the percentage of repurchases undertaken by the firm compared to the amount announced. INDUS is a dummy variable for industry with a value of 1 if the firm is in the financial sector and 0 otherwise. FSIZE is a dummy variable for firm size based on the following: 1 if < 15 million shares outstanding 2 if million shares outstanding 3 if million shares outstanding 4 if million shares outstanding 5 if > 60 million shares outstanding ASIZE is a continuous variable of the actual percentage size of the repurchases relative to total shares outstanding. 16

17 V. Results The results from the calculation of the average cumulative abnormal returns confirms the signaling hypothesis. No evidence is found to support a theory of market efficiency in that stock prices are not shown to react to repurchase activity. The abnormal returns that were obtained immediately surrounding the announcement date was not replicated over the long run as returns in subsequent weeks closely mirrored market returns. (See Table 3) Long-term Abnormal Returns The results of the examination of the levels of repurchases over a twenty-six week period starting one week after a repurchase announcement indicates that a majority of firms do not carry out the stated levels of repurchases. Only 22% of firms repurchase more than 80% of the targeted repurchases and still fewer (20 or 14.5%) repurchase 100% or more of the targeted shares. Further, 41 or 29.7% do not record any repurchases. Arithmetically, the differences between the mean abnormal returns of each quintile appear to be large. However, the relatively large standard deviations results in overlap of mean abnormal returns. We also obtain an F-statistic of (Table 4) and a significance value of.339, indicating no statistical difference between the means at the 95% confidence interval. Additional tests of significance (not reported here) at the 90%, 80% and 50% levels had similar insignificant results. Next, a computation of the mean cumulative abnormal returns for a sub-sample of firms having positive short-term abnormal returns gives an ACAR value of 4.59% with an associated t-value of 1.672, and is significant at the 10% level. The ANOVA to 17

18 test for differences between the means of the quintiles for this sub-sample yields an F- value of.347 indicating statistically significant differences. A similar analysis of the sub-sample of firms with negative short-term abnormal returns indicates insignificant results. Short-term Abnormal Returns Table 3 indicates that the overall sample had an average five-day cumulative abnormal return of 0.04%, which was not statistically significant. This finding varies from earlier works that concentrates on the shorter time period of two days after announcement. For example, Ikenberry, Lakonishok, and Vermaelen (1995) find 2-day abnormal returns of approximately 3%. A possible explanation is that the market reaction is complete within a day or two. After grouping the firms in quintiles, an examination of the differences between the means (See Table 5) indicate no statistical significance thus implying that any shortterm abnormal returns does not influence subsequent levels of repurchases. Total Abnormal Returns The total 27-week period indicates average abnormal results of 0.60%, which is statistically insignificant. The F-statistic of was also statistically insignificant, implying that there was no difference between the mean returns of the quintiles. Since the long-term abnormal results were insignificant, the implication is that the gains from a stock repurchase announcement is concentrated around the time of the announcement and subsequent purchases of stock does not provide any additional influence on marketadjusted stock price returns. Cross Sectional Analysis 18

19 The cross sectional regression coefficients are presented in Table 7 and are not statistically significant at conventional levels. The coefficient with the largest impact is the firm size, and although not statistically significant, is consistent with the general trend in the marketplace where the larger firms generally tend to follow through on their repurchase announcement. For example, of the twenty firms that repurchased 100% of the announced share target, fifteen had total outstanding shares that resulted in the firm being classified within the largest of the five categories used. VI. Summary and Conclusion Earlier research indicates that stock repurchases primarily serve as a signaling mechanism of management s view that their firm s stock is undervalued. Stock repurchase announcements have also been shown to result in positive and statistically significant abnormal returns around the announcement date. This paper investigates whether or not firms experience long-term benefit in addition to the well-documented short-term positive abnormal returns from purchasing their stock after announcing plans to do so. The hypothesis that markets are efficient and hence stock prices will adjust to stock repurchase information is not supported. Instead, an alternative hypothesis that the market is not efficient and stock prices do not adjust to penalize firms for not repurchasing stocks as announced is shown to be valid. This paper reinforces the signaling hypothesis that stock repurchase announcements primarily serve to articulate management s views of under-valuation, rather than management s specific plans in relation to repurchasing stock. Data 19

20 presented in this paper and in other works point to a stock repurchase announcement as simply an announcement. Although such announcements permit firms to repurchase their shares, they are under no obligation to do so. Hence, if management perceives that the announcement has worked to their advantage in contributing to the rise in price of the shares, they would have accomplished their main objective of enhancing shareholder value. At the same time, the nature of open market repurchase programs allows firms to time their share repurchases to take advantage of changes in stock price and provides flexibility to firms that face uncertain cash flows during the repurchase period. The short-term effects of the announcements are inconsistent with earlier works with negative abnormal returns (although not significant). The long-term effect is also statistically insignificant in terms of the overall sample, but evidence of statistical significance is found for a sub-sample of firms with positive ACAR in the five-day window. More importantly, this study provides new insights into the variability of returns based on the extent of repurchases. There is no evidence of statistically significant differences between the mean abnormal returns for the sample of stocks over a 26-week period when divided into quintiles based on level of repurchases compared to an announced target. The range of data and the similarity of the sample with other research allow some generalization of the findings of paper. Since stock price performance is shown to be independent of the level of repurchase transactions, the implication is that firms will not reap additional stock price benefits from following through on repurchasing stock after announcing a plan to do so, and the repurchase announcement is sufficient to obtain abnormal returns. Investors therefore must view repurchase announcements as a means 20

21 of short-term gain and not a viable strategy for significant long-term abnormal gains. Future research could extend the analysis to investigate the influence on subsequent firm financial performance based on the extent of repurchases after an announcement to repurchase stock. 21

22 Table 1 Repurchase Announcements TOTALS Total Announcements Included in Final Sample Percentage Repurchased > Excluded from Final Sample Table 2 Percentage Distribution of Sample Repurchase Announcements TOTALS Percentage Repurchased % 8.70% 10.14% 10.87% 15.22% 50.00% % 2.17% 1.45% 2.17% 5.80% 12.32% % 1.45% 1.45% 0.72% 3.62% 10.14% % 0.72% 0.00% 2.17% 2.17% 5.80% 81 - > % 1.45% 5.07% 4.35% 8.70% 21.74% Totals 11.59% 14.49% 18.12% 20.29% 35.51% % 22

23 Table 3 Mean Abnormal Returns (in Percentage) TOTALS 5-Day Returns (-.185) (0.334) (-0.743) (-0.462) (-0.587) (-0.883) (0.403) (1.205) (-0.324) (1.018) (1.470) (1.675) (1.275) (0.923) (-1.276) (-0.339) (0.894) (-1.248) (-0.487) 81 - > (-.353) (2.195) (-0.545) (-0.223) (2.159) (1.066) Total (0.354) (1.126) (-0.200) (-0.747) (-.144) (-0.114) 26- Week Returns (1.419) (0.437) (1.049) (-0.233) (-0.360) (0.594) (-.376) (1.918) (-1.740) (-0.009) (-.289) (1.327) (-.273) (7.211)* (0.567) (1.380) N/A (-1.001) (-1.283) (-1.386) 81 - > (2.540) (-.038) (0.010) (0.267) (-.386) (0.046) Total (2.907)** (0.255) (1.505) (-1.144) (-0.482) (0.326) Total Returns (1.226) (0.464) (0.922) (-0.327) (-0.248) (0.402) (-0.304) (1.714) (-1.976) (0.193) (-0.028) (1.437) (-.233) (5.112) (0.328) (1.174) N/A (-1.007) (-1.277) (-1.384) 81 - > (2.315) (0.093) (-0.182) (0.192) (-0.180) (0.194) Total (2.971)** (0.391) (1.429) (-1.252) (-0.489) (0.294) t-values in parenthesis * 10% ** 5% 23

24 Table 4 ANOVA For 26 Week Cumulative Abnormal Returns Sum of Df Mean F Sig. Squares Square CAR * Between (Combined) Percent Repurchase Groups Within Groups Total Table 5 ANOVA For 5-day Cumulative Abnormal Return Sum of Squares Df Mean Square F Sig. WEEKCAR * Between (Combined) Percent Groups Repurchase Within Groups Total Table 6 ANOVA For Combined Cumulative Abnormal Returns Sum of Df Mean F Sig. TOTALCAR Between * Percent Groups Repurchase Within Squares Square (Combined) Groups Total Table 7 Regression Coefficients Unstandardized Coefficients Standardized Coefficients t Sig. 95% Confidence Interval for B Lower Bound Model B Std. Error Beta Upper Bound 1 (Constant) Percent Repurchase Industry Firm Size Percent of Shares Announced a Dependent Variable: CAR 24

25 References Chen, R. Carl, Nancy J. Mohan, Thomas L. Steiner Discount Rate Changes, stock market returns, volatility, and trading volume: Evidence from intraday data and implications for market efficiency, Journal of Banking and Finance 23, Akhigbe, Aigbe and Ann Marie Whyte The Source of Gains to Targets and Industry Rivals: Evidence Based on Terminated Merger Proposals. Working Paper. Barber, Brad M. and John D. Lyon Detecting Long-run Abnormal Stock Returns: The Empirical Power and Specification of Test Statistics, Journal of Financial Statistics Brown, Stephen and Jerold B. Warner Measuring Security Price Performance, Journal of Financial Economics Chen, R. Carl, Nancy J. Mohan, Thomas L. Steiner Discount Rate Changes, stock market returns, volatility, and trading volume: Evidence from intraday data and implications for market efficiency, Journal of Banking and Finance 23, Comment, Robert and Gregg A. Jarrell The Relative Signaling Power of Dutch Auction and Fixed-Price Self-Tender Offers and Open-Market Share Repurchases, The Journal of Finance Ikenberry, David, Joseph Lakonishok, Theo Vermaelen Market Underreaction to open market share Repurchases, Journal of Financial Economics 39, Kracher, Beverly and Robert R. Johnson Repurchase Announcements, Lies and False Signals, Journal of Business Ethics Lakonishok, Josef and Theo Vermaelen Anamalous Price Behavior Around Repurchase Tender Offers, The Journal of Finance Lie, Erik and John J. McConnell Earnings Signals in Fixed-price and Dutch Auction Self-tender Offers, Journal of Financial Economics Nohel, Tom and Vefa Tarhan Share Repurchases and Firm Performance: New Evidence on the Agency Costs of Free Cash Flow, Journal of Financial Economics Persons, John C Heterogeneous Shareholders and Signaling with Share Repurchases, Journal of Corporate Finance

26 Stephens, Clifford P. and Michael S. Weisbach Actual Share Reacquisitions in Open-Market Repurchase Programs, The Journal of Finance Vermaelen, Theo Common Stock Repurchases and Market Signaling: An Empirical Study, Journal of Financial Economics Vermaelen, Theo Repurchase Tender Offers, Signaling, and Managerial Incentives, Journal of Financial and Quantitative Analysis