Share Repurchase Reasons and the Market Reaction to Actual Share Repurchases: Evidence from Australia

Share Repurchase Reasons and theMarket Reaction to Actual Share

Repurchases: Evidencefrom Australia

ALI C. AKYOL AND CHI CHONG FOODepartment of Finance, Faculty of Business and Economics, University of

Melbourne, Parkville, Melbourne, Victoria, Australia

ABSTRACT

Using repurchase reasons provided by Australian companies for their stockrepurchase programs, we ask if the market’s response is different acrossrepurchase motivations by examining actual daily share repurchases. We findthat firms with the undervaluation motive experience a more positive stockprice reaction when they report their repurchases to the market. We showthat undervaluation motive firms repurchase fewer shares and have a lowerprogram completion rate than other motives firms. During the 1-year periodfollowing the repurchases, undervaluation motive firms do better than theircontrol sample firms whereas other motive firms do not perform better orworse than their control sample firms. Overall, our results suggest that theundervaluation motive is a stronger signal than other repurchase motives,and contrary to the predictions of the standard signaling theories, manage-ment statements carry some value for the market. We also present someevidence suggesting that a costly action may not be needed for a signal to becredible.

I. INTRODUCTION

This paper examines management statements provided for share repurchaseprograms by examining actual repurchases to test the argument of the tradi-tional signaling theories that management signals have no value for the market.We study open-market share repurchases because share repurchases are one ofthe most popular methods for companies to distribute cash to shareholders.1

Banyi et al. (2008) report that from 1985 to 2004, the total announced value ofrepurchases in the United States was USD 1.96 trillion and 89% of them were toopen-market repurchases.

1 In Australia, open-market repurchases are referred to as on-market buybacks and tender offersare referred to as off-market buybacks (Brown 2007). In this paper, the terms ‘repurchase’ and‘buyback’ are used interchangeably.

bs_bs_banner

International Review of Finance, 2012DOI: 10.1111/j.1468-2443.2012.01159.x

© 2012 The Authors. International Review of Finance © International Review of Finance Ltd. 2012. Publishedby Blackwell Publishing Ltd., 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main Street, Malden, MA02148, USA.

There are numerous studies on the motivations of open-market repurchases.2

The signaling of undervaluation has been found to be one of the most popularmotivations that emerge from these studies. The empirical evidence is sup-ported by management surveys, where managers consider stock price under-valuation as one of the most important considerations when deciding whetherto repurchase shares (see, for example, Brav et al. 2005 in the United States andMitchell et al. 2001 in Australia).3 Most studies also report negative abnormalreturns prior to and positive abnormal returns around stock repurchase programannouncements, suggesting that managers time the market to exploit potentialundervaluation when making the announcements (Dann 1981; Vermaelen1981; Asquith and Mullins 1986; Comment and Jarrell 1991; Stephens andWeisbach 1998 in the United States; and Lamba and Ramsay 2005 in Australia).It has also been shown that investors initially view repurchase announcementswith some skepticism, which leads to positive long-term drifts in stock returnsfollowing the announcements (Ikenberry et al. 1995; Konan et al. 2004; Chanet al. 2007; Peyer and Vermaelen 2009).

Although signaling of undervaluation is found to be important in manystudies, the traditional theories of signaling (Bhattacharya 1979; Vermaelen1984; Miller and Rock 1985) do not consider management statements – that thefirm is undervalued – as credible signals and judge them as ‘cheap talk’ aspointed out by Peyer and Vermaelen (2009). According to these theories, for thesignal to be credible it should be costly.

One of the key features of open-market repurchases is that they do notcommit the firm to repurchase shares in the open market. A repurchaseannouncement effectively grants the firm an option to repurchase shares in theopen market at management’s discretion (Ikenberry and Vermaelen 1996).Therefore, for the undervaluation signal to be convincing (as argued by thetraditional signaling theories), firms actually have to repurchase shares in theopen market so that those firms that are truly undervalued can be distinguishedfrom mimicking firms. Consequently, data on actual share repurchases areimportant for testing the undervaluation hypothesis.

In the United States, firms are not required to report their daily buybackactivities.4 Thus, in the absence of actual share repurchases data, studies in theUnited States must rely on quarterly estimates of shares repurchased (Stephensand Weisbach 1998) or questionnaires completed by repurchasing firms (Cooket al. 2004). Both studies find evidence in support of managerial timing abilities.

2 See, for example, Dittmar (2000), Konan et al. (2004), Oded (2005), Billett and Xue (2007) inthe United States and Mitchell and Dharmawan (2007) in Australia.

3 Graham (1999), in a survey study, reports that as the Dow Jones Industrial Average wasapproaching a new height of 10,000, only 3% of the financial executives thought that theirstock was overvalued and more than 60% of them considered their stock undervalued.

4 However, after making changes to Rule 10b-18 in 2004, the Securities and Exchange Com-mission now requires companies to disclose more about their repurchase activities in theirquarterly statements. See Ben-Rephael et al. (2011) for an example of a study that uses suchdata from financial reports.

International Review of Finance

© 2012 The Authors2 International Review of Finance © International Review of Finance Ltd. 2012

Stephens and Weisbach (1998) provide further evidence that managers not onlytime the repurchase announcements, but also the daily execution of repurchaseprograms.

Unlike the United States, some other markets (such as Australia and HongKong) have transparent share buyback regimes in which companies are requiredto disclose their buyback activities on a daily basis. Studies that use actual sharebuyback data generally find that firms repurchase shares following price drops,managers time their repurchases, and firms are able to buy their shares at alower price than a naïve accumulation strategy (see, for example, Brockman andChung 2001; Zhang, 2005; and McNally et al. 2006).

In this paper, we study share repurchases in Australia between 1998 and2008. Australian legislation requires companies to disclose their buyback activi-ties daily. Different from other international markets, the Australian buybackenvironment also requires repurchasing firms to provide at least one reason fortheir repurchases when the buyback program is initiated. Using daily announce-ments of actual repurchases with the reasons provided for those repurchases, weexamine the market’s reaction to daily stock repurchases in Australia and if themarket’s response differs across repurchase reasons, also testing the claim of thetheories that management statements have no value. Specifically, we investigateif one specific repurchase reason – the undervaluation of the company’s stockprice – elicits a different market reaction compared with other repurchasereasons. We also exploit the special nature of the Australian data and use thesequential nature of the daily repurchase activities to investigate if undervalu-ation motive firms behave differently than other repurchasing firms; forexample, if undervaluation motive firms are more aggressive in repurchasingshares in the first months of the announcement, if undervaluation motive firmsbuy more aggressively when the price is lower, and if they stop their buybackactivity when the price has increased. The examination of daily repurchaseactivity is important as it may answer the question of if cheap-talk is moreimportant than action. A recent paper by Douthit et al. (2011) employs anexperimental setting and argues that cheap-talk can be an effective tool inreducing agency problems. By studying daily-repurchase activity for undervalu-ation and other motives firms, we can gain insight into this question.

We find that both initial program announcements and actual repurchases arepreceded by a period of negative abnormal returns. The cumulative abnormalreturn (CAR) in the preannouncement window is -0.91% (p value = 0.09) forinitial program announcements and -1.15% for announcements of actual sharerepurchases (p value = 0.00). The market responds significantly positively toboth the initial repurchase program announcements and daily repurchaseannouncements. Moreover, we find that firms that announce a repurchaseprogram and firms that actually repurchase shares with the undervaluationmotive receive a more favorable market response than firms with other repur-chase motives. Our univariate results, overall, suggest that repurchasing firms(both undervaluation motive and other motives firms) are underperforming themarket prior to the actual repurchases; managers are taking advantage of this

Actual Share Repurchases and Management Statements

© 2012 The Authors3International Review of Finance © International Review of Finance Ltd. 2012

opportunity to benefit nonselling shareholders; and repurchases with theundervaluation motive have more positive abnormal returns.

The examination of daily repurchase announcements reveals that undervalu-ation motive firms, on average, buy back fewer shares than other motives firmsthroughout the program’s life and subsequently they have a lower programcompletion rate than other motives firms. Also, undervaluation firms do notseem to buy back shares in response to their share price movement.

In the cross-section, we find that the market reaction to repurchases made byundervaluation firms is more positive after controlling for other factors. Exam-ining monthly repurchase activities in a cross-sectional setting, we again showthat undervaluation firms tend to repurchase fewer shares during the month,and past and current stock price performances are negatively related to thepercentage of shares repurchased in the market. As a final test for actual repur-chases, we examine program completion rates in a multivariate setting and findthat undervaluation firms have lower completion rates than other motives firmsafter controlling for firm and repurchase program characteristics.

Overall, the results from the cross-sectional analysis support our univariatefindings and indicate that the market is taking into account the repurchasereason when reacting to actual repurchase announcements as well as to initialrepurchase program announcements, and undervaluation firms buy fewershares under the repurchase program than other motives firms. Our findingsalso suggest that the signal sent by the firm when the firm cites undervaluationas the repurchase reason is stronger and seen as a credible one by the market.

When we examine the long-term stock price performance of repurchasingfirms over a horizon of up to 3 years against a control sample of firms that donot have a repurchase program, we find that repurchasing firms that use othermotives do not do better or worse than their matched control sample firms.However, undervaluation motive firms statistically perform better than theircontrol sample firms at least in the year following the repurchase, suggestingthat undervaluation firm managers are able to detect valuation errors when theybuy back shares. However, when we examine firms that do not repurchase anyshares after their initial program announcements, we find that nonrepurchasingundervaluation firms do also better than their control sample firms during the1-year period following repurchase program announcements. Although we donot emphasize this result due to the very small sample size, it suggests that acostly action (such as actual share repurchases), as argued by the traditionalsignaling theories, may not be needed for a signal to be credible.

We contribute to the share repurchases literature as well as to the growingliterature of repurchase studies that use actual repurchase data in couple ofways. First, different from other studies, we use the time-series pattern of dailyrepurchasing activities to gain more insight into firms’ buyback patterns andshow that undervaluation motive firms repurchase fewer shares than othermotives firms throughout the program, which leads to a lower program comple-tion rate. Second, our results indicate that not all repurchases are equal as themarket’s response to the ones with the undervaluation motive is more positive,



confirming Peyer and Vermaelen (2009). Our results, thus, suggest that man-agement’s signal is more valuable to the market when the firm sends the marketa stronger signal besides actually buying back shares (i.e., when the repurchasereason is the undervaluation of the firms’ shares). Third, by using Australiandata, we show that firms buy back their shares after a fall in their share pricesand provide further evidence that managers time the market to take advantageof the stock price undervaluation.

We deviate from other papers that use repurchase motives, such as Peyer andVermaelen (2009), in that we focus on actual repurchases rather than only onthe announcements of repurchase programs to better capture the signalingeffect of management statements and their effect on firm value. Overall, as inPeyer and Vermaelen (2009), we report that contrary to the presumption of thetraditional signaling models, management statements are valuable for themarket and find evidence that a costly action may not be needed for the signalto be credible.

The remainder of this paper is organized as follows. We describe the Austra-lian buyback environment in Section II and discuss prior studies in Section III.Section IV discusses our hypotheses and Section V explains our data andpresents descriptive statistics. We report the empirical results in Section VI andSection VII concludes the paper.

II. THE AUSTRALIAN BUYBACK ENVIRONMENT

In Australia, share repurchases became possible in 1989 following the report ofthe Companies and Securities Law Review Committee. However, as a result ofthe restrictive requirements imposed on repurchasing firms, there were notmany buybacks until the end of 1995.5 The introduction of First CorporationsLaw Simplification Bill, effective December 1995, simplified the process of sharebuybacks. Most importantly, the mandatory requirements on share repurchaseswere reduced substantially (Lamba and Ramsay 2005). These changes haveresulted in an increase in the popularity of share repurchases (Mitchell et al.2006).

To undertake a share repurchase program, a company in Australia is requiredto make an initial formal announcement of the intention to buy back shareswith detailed information. The information includes the duration of the repur-chase program, the number of shares intended to be repurchased, and thereason for the buyback.6 In addition, according to Corporation Act 2001, com-panies in Australia can only buy back a maximum of 10% of the number ofshares issued over a 12-month period without shareholders’ approval (Brailsfordet al. 2008). Furthermore, the buyback process throughout the buyback periodis transparent, and firms are required to make timely disclosures relating to the

5 Lamba and Ramsay (2005) report that there were only 32 buybacks between 1989 and 1995.6 Prior to September 30, 2001, a firm also needed to disclose whether the directors would

participate in the share repurchase program.



buyback activity. Moreover, any changes relating to the buyback program haveto be announced to the market.

According to the Australian Securities Exchange (ASX) Listing Rules 3.8A,firms are required to disclose the repurchase transactions at any given day by9:30 AM on the next business day. The information disclosed is contained inAppendix 3E – Daily Share Buyback Notice. The information includes therepurchase date, the number of shares bought back, total consideration paid,and the highest and lowest prices paid. In addition, Listing Rules 7.33 restrictsthe repurchase price to be no more than 5% of the 5-day average price prior tothe repurchase day. Moreover, the shares repurchased must be cancelled andcannot be treated as treasury shares (Mitchell and Dharmawan 2007). Theavailability of actual share buybacks data as well as of repurchase reasons makethis study possible.7

A share buyback program finishes with a final buyback notice issued to themarket. It signifies the completion of the buyback program. The informationdisclosed includes total number of shares bought back, total consideration paid,and highest and lowest prices paid throughout the buyback duration.

III. PRIOR STUDIES

The literature on the motivations of open-market share repurchases is extensive:signaling, free cash flow, capital structure adjustment, dividend substitution,and earnings management. Of all these motivations, signaling of undervalua-tion has emerged as one of the few motivations that are well accepted both inthe academia (Dann 1981; Vermaelen 1981; Ikenberry et al. 1995; Lie 2005;Peyer and Vermaelen 2009) and the corporate world (Mitchell et al. 2001; Bravet al. 2005).

The signaling of undervaluation hypothesis suggests that in an incompletemarket where information asymmetry exists, managers are better informedthan shareholders, and managers can use repurchases to signal better futureprospects or undervaluation to shareholders (Grullon and Ikenberry 2000;Grullon and Michaely 2004). Peyer and Vermaelen (2009), on the other hand,report findings which are consistent with the overreaction hypothesis thatrepurchases are conducted because managers disagree with the market as tothe correct value of the firm rather than managers’ signaling better futureprospects to investors – rejecting the inside information hypothesis. Peyer andVermaelen (2009), however, note that their results are not inconsistentwith the signaling of undervaluation hypothesis. They argue that managersinitiate repurchase programs not because they think the future prospects are

7 One may argue that we should be skeptical of the reasons provided by management, becausethey may not be the true motivations for a share repurchase. However, as discussed byMitchell and Robinson (1999) and Peyer and Vermaelen (2009), the career concerns ofmanagers and other market forces, such as the probability of a lawsuit, encourage managersto disclose the true reasons for share repurchases.



improving, but because they disagree with the market’s assessment that thefirm’s performance is falling.

The signaling theories suggest that an effective signal must be costly, so thatthose firms that are truly undervalued can be distinguished from mimickingfirms and a share repurchase is a costly signal because it distributes cash back toshareholders (Oded 2005). Managers prefer share repurchases as they are moreflexible than dividends and they can use this flexibility to time the market byincreasing (or initiating) repurchases when their company’s stock price is low(Brav et al. 2005).

Consistent with the signaling/undervaluation hypothesis, numerous studiesdocument positive and significant abnormal returns around buyback programannouncement dates. For example, Peyer and Vermaelen (2009) report that theaverage 3-day CAR surrounding the initiation of repurchase programs is 2.39%from 1991 to 2001.

In recent years, some studies, including studies on repurchases, have foundlong-term drifts in stock returns after different types of corporate events (Konanet al. 2004).8 The results from those studies suggest that the short-run marketresponse is incomplete and that long-run stock performance may be informa-tive; therefore, studies that draw conclusions only based on initial marketresponse do not tell the full story (Konan et al. 2004). Ikenberry et al. (1995)argue that the market underreacts to repurchase announcements because inves-tors perceive the announcements with skepticism, which leads to a positivelong-term drift in stock returns. Their results indicate that investors underreactto repurchase announcements by value stocks and stock prices react slowly tothis underreaction over time.9

In addition, various studies observe significant negative abnormal returnsprior to repurchase announcements, suggesting that managers time the marketto exploit potential undervaluation when making repurchase announcements(Ikenberry et al. 1995; Stephens and Weisbach 1998; Chan et al. 2007).Stephens and Weisbach (1998) posit that the flexibility to repurchase shares inthe open market enables firms to benefit from fluctuations in stock price bytiming their share repurchases. Consistent with their argument, they find thatquarterly estimates of actual share repurchases are negatively related to a firm’sprior quarter’s stock return. This provides further evidence that managers notonly time the repurchase announcements, but also the daily execution of therepurchase programs.

In the United States, a direct measurement of the number of shares repur-chased is not possible as firms are not required to disclose shares repurchasedwhen making actual share repurchases. Therefore, the accuracy of the estimates

8 Two of these corporate events are seasoned equity offerings (Loughran and Ritter 1995) andinitial public offerings (Purnanandam and Swaminathan 2004).

9 However, it is important to note that the results on long-run stock performance are sensitiveto the procedures used (Konan et al. 2004). Therefore, the methodologies used to studylong-run performance are critical for the accuracy of the results. See Barber and Lyon (1997)for the problems of event studies used to test long-run abnormal returns.



used is important for studies on share repurchases (Banyi et al. 2008). Banyiet al. (2008) claim that the estimation error can be very high and argue thatwith the availability of actual share repurchases data, many previous studiesneed to be re-examined.

Brockman and Chung (2001) argue that the issue of managerial timing is bestexamined in a buyback environment where the actual shares repurchased areknown to the market. They examine this issue with daily shares repurchasesdata in Hong Kong and report that managers possess superior timing ability andable to acquire shares at lower costs than a naïve accumulation strategy.McNally et al. (2006) obtain similar results with monthly repurchases data inCanada. However, McNally and Smith (2007) find no evidence of managerialtiming ability in France. Zhang (2005) examines the market reaction followingactual share repurchases in Hong Kong. He finds that market responds posi-tively to the actual repurchases, where the 3-day CAR is significant at 0.43% andthat the market reacts more positively to the actual repurchases by small andhigh book-to-market firms. Arguing that event clustering problems may affectZhang’s results, Wang et al. (2009) examine actual share repurchase announce-ments in the UK with a modified multivariate regression model to solve theevent clustering problems. Consistent with Zhang (2005), they find that marketreacts positively to actual repurchase announcements.

Share buybacks have become a popular payout method for many Australianfirms to distribute cash to shareholders.10 Five types of buyback are recognizedin Australia: on-market, equal access, selective buyback, employee sharescheme, and minimum holding.11 Further, Mitchell and Robinson (1999) findthat signaling incentives are the most important motivation for on-marketbuybacks in Australia.

Consistent with the US studies, many Australian studies find a significantlypositive stock price reaction to the announcement of a share buyback program.For example, the 3-day CAR surrounding an on-market buyback announcementis 3.3% from 1989 to 1998 (Lamba and Ramsay 2005). Further, buybackannouncing firms are undervalued prior to the announcement (Mitchell et al.2006) and industry rivals of buyback announcing firm are positively affected(Otchere and Ross 2002). Taken together, these results suggest that managementuses buybacks when its stock is undervalued and has the ability to time themarket to reduce the undervaluation, and the results are similar to the findingsin US repurchase studies.

Mitchell and Dharmawan (2007) argue that the formalized and transparentbuyback regime in Australia provides a better environment in which to capturethe signaling incentives for on-market buybacks. Therefore, in contrast to thelow information content of the US open-market repurchase environment,the buyback programs in Australia are more of a definite commitment given the

10 Brown (2007) documents that from 1996 to 2003, there were about 350 (45) on-market(off-market) buybacks in Australia to buy back $10.5 billion ($12.1 billion) worth of shares.

11 See Mitchell et al. (2001) for a discussion of the various buyback types in Australia.



rich and timely information content throughout the entire buyback period andthe expectations of the market.12

The results of Otchere and Ross (2002), Mitchell et al. (2006), and Mitchelland Dharmawan (2007) support the argument that the reasons provided bymanagement has information content and that share buybacks in Australia aredifferent from other countries.

Our study is different from other studies on repurchases in Australia as wefocus on actual share repurchases – rather than announcements related to theinitiation of buyback programs – by using repurchase reasons provided by firms.Our analysis of actual repurchases by their stated motives may help us under-stand which types of repurchases have information content and which ones donot change the market’s view of the firm. Additionally, as firms are required todisclose their actual repurchases to the market, we are able to identify the firmsthat do not repurchase any shares (zero repurchasers). Finally, we also test theclaim of the traditional signaling theories that management statements are notvaluable for the market.

IV. HYPOTHESES

A share repurchase announcement is not a firm commitment to repurchaseshares in the open market. While most companies repurchase some shares inthe open market, other companies make zero repurchases after the announce-ment, and, most often, the actual shares repurchased are much lower than theintended number of shares to repurchase (Bonaimé 2010). Therefore, actualshare repurchases are important for investors to verify a firm’s commitment toa share repurchase program.

As discussed in the previous section, repurchase announcements are nor-mally preceded by a period of negative abnormal returns and the announce-ments are greeted positively by the market. This suggests that managers timethe market to exploit potential undervaluation when making repurchaseannouncements.

Hypothesis 1: If managers also time the daily execution of the repurchaseprogram, then we should observe a negative abnormal return prior to actualshare repurchases.

Otchere and Ross (2002), Mitchell et al. (2006), and Peyer and Vermaelen (2009)report that buyback announcing firms with the undervaluation motive earnhigher abnormal returns than firms with other motives around the initiation oftheir share buyback programs. This suggests that the reason provided by man-agement is of value to the market.

12 See Dharmawan and Mitchell (2001) for a comparison of the buyback regulations andprocedures in Australia, United Kingdom, Hong Kong, New Zealand, United States, andCanada.



Hypothesis 2: We should observe a stronger market response to actual sharerepurchases by undervaluation firms as those firms show their commitment totheir repurchase programs.

The share-repurchase literature finds that the market initially underreacts torepurchase announcements and the underreaction is corrected in the long-term(see, for example, Peyer and Vermaelen 2009).

Hypothesis 3: Firms with the undervaluation motive should perform betterthan other repurchasing firms in the long-run if management’s signal is of valueto the market.

As noted previously, we also know the firms that do not make any repur-chases. We call those firms as zero repurchasers. The signaling theories suggestthat firms that repurchase shares are sending a credible signal to the market. Wecall the firms that buy shares back as positive repurchasers. By examining zerorepurchasers and comparing them with positive repurchasers, we can determineif short- and long-term performances of zero and positive repurchasers aresimilar and if actual share buybacks are needed for the signal (sent through theinitial buyback announcement) to be credible.

Hypothesis 4: A costly action is needed for a signal to be credible.

V. DATA AND DESCRIPTIVE STATISTICS

We construct a database of repurchase announcements and daily buybacktransactions with manually collected data from Aspect Huntley’s DatAnalysisdatabase. To be included in the sample, each announcement must have infor-mation on issued capital and the number of shares intended to be repurchased.The sample includes all companies except for the financials and utilities listedon the ASX that conducted on-market buybacks from September 1, 1998 toDecember 31, 2008.

Using data from IRESS and Datastream, we create the following variables:daily return on the Australian All Ordinaries Index, market-to-book ratio (MB),and market value of equity.13 As MB is required for subsequent analysis, firmswithout MB in Datastream are excluded from the analysis. For each actualrepurchase day, we compute the size and MB quartiles using all firms listed onthe ASX and then allocate each repurchasing firm to the appropriate size andMB quartiles on the actual repurchase day.

Peyer and Vermaelen (2009) read share repurchase-related press releases fortheir sample of 3725 initial announcements and find buyback reasons for only2093 of them (56%). In the United States, firms are not required to disclose anyreason as to why the repurchase program was initiated. However as explained in

13 IRESS is a supplier of stock market data in Australia.



Section II, firms in Australia are required to provide at least one reason for theirbuyback program. We provide an example of an initial buyback announcementfor one of our sample companies in the Appendix (Autron Corp.). As seen inItem 7 of the buyback notice, the firm provides a reason. Different from Peyerand Vermaelen (2009), we do have, at least, one reason for every firm in oursample, thus mitigating any sample selection issues that might arise becausesome firms may choose not to give any reason for their repurchase program.

We read initial program announcements and classify each announcement inone of the following seven categories: undervaluation, capital management,efficient use of capital, improve EPS, provide liquidity, reduce share price vola-tility, and others.14 We only include those announcements that state either‘undervaluation’ or ‘increase shareholders’ return’ as a motive in the undervalu-ation subsample.15 We exclude repurchase announcements that state ‘offsetdilution’ as the reason for a buyback from the sample as target shares tend to besmall and it is unlikely that the buyback reason is undervaluation.16 In addition,announcements that made zero actual repurchases are grouped separately. Aszero repurchasers do not have any actual repurchases, we only examine themarket’s initial reaction to program announcements and the long-term perfor-mance of zero-repurchasers following initial program announcements. The finalsample of positive repurchasers is then grouped into two subsamples based onthe reasons provided by management at the time of initial announcement. Thefirst subsample contains announcements of undervaluation motive and thesecond subsample contains announcements of other motives. There are 192buyback program announcements by positive repurchasers (132 firms) and 20buyback program announcements by zero repurchasers (16 firms). We providethe frequency distribution of repurchase announcements for each repurchasemotive in Table 1 [similar to table 5 in Peyer and Vermaelen (2009)]. The tableshows the number of times each motive is cited in a buyback programannouncement and also the number of times the motive is cited together withany other six motives. The table shows that 31 program announcements men-tioned undervaluation as the only motive and 11 announcements mentionedanother motive besides undervaluation. In total, 47 announcements mentionedundervaluation as a motive. Of the 47 announcements, 41 are made by positiverepurchasers and 6 by zero repurchasers.17

14 See Table 1 for how the announcements are classified into seven groups.15 Our classification of repurchase reasons is similar to that of Peyer and Vermaelen (2009) who

classify ‘undervaluation’ and ‘best use of money’ in their undervaluation category.16 There were 37 such announcements with 0.77% for shares sought. The mean market reaction

in the (-1, 1) window to the initiation of dilution motivated repurchase programs is 0.22%(p value = 0.90). Including those announcements in the analysis do not change the mainresults.

17 In our analysis, we do not distinguish between single and multiple motivation announce-ments. However, our main results stay quantitatively similar if we restrict our analysisto only single motivation announcements (e.g., to only 31 single undervaluation motiveannouncements).



Tab

le1

Freq

uen

cyd

istr

ibu

tion

ofsh

are

rep

urc

has

escl

assi

fied

bym

otiv

atio

n

Mot

ivat

ion

s

Un

der

valu

atio

nC

apit

alm

anag

emen

tEf

fici

ent

use

ofca

pit

alIm

pro

veEP

SPr

ovid

eli

qu

idit

yR

edu

cesh

are

pri

cevo

lati

lity

Oth

ers

Nu

mbe

rof

mot

ivat

ion

sp

eran

nou

nce

men

t1

3112

215

47

14

211

41

50

00

35

02

00

00

Tota

l–

pos

itiv

ere

pu

rch

aser

s41

116

187

71

2

Tota

l–

zero

rep

urc

has

ers

610

02

00

2

Tota

l–

full

sam

ple

4712

618

97

14

Th

ista

ble

pro

vid

esth

en

um

ber

ofob

serv

atio

ns

per

mot

ivat

ion

for

firm

sth

atan

nou

nce

da

rep

urc

has

ebe

twee

nSe

pte

mbe

r19

98an

dD

ecem

ber

2008

.T

he

mot

ivat

ion

sar

eob

tain

edfr

omea

chfi

rm’s

rep

urc

has

ean

nou

nce

men

tp

oste

don

the

ASX

’sw

ebsi

te.

Und

erva

luat

ion

refe

rsto

thos

ean

nou

nce

men

tsw

her

eth

efi

rmcl

earl

yst

ates

un

der

valu

atio

nor

alo

wsh

are

pri

ceas

the

rep

urc

has

ere

ason

.R

epu

rch

ases

that

men

tion

cap

ital

man

agem

ent

asa

reas

onar

ecl

assi

fied

asC

apit

alm

anag

emen

t.If

anan

nou

nce

men

tm

enti

ons

low

pro

babi

lity

ofa

futu

reac

qu

isit

ion

orm

ater

ial

cap

ital

inve

stm

ent,

then

we

clas

sify

the

ann

oun

cem

ent

asEf

fici

ent

use

ofca

pita

l.A

nn

oun

cem

ents

ofim

pro

vin

gea

rnin

gsp

ersh

are

are

clas

sifi

edas

Impr

ove

EPS.

An

nou

nce

men

tsof

rep

urc

has

esth

atm

enti

onim

pro

vin

gli

qu

idit

yin

the

mar

ket

soin

vest

ors

can

sell

thei

rsh

ares

wit

hou

tre

du

cin

gth

ep

rice

they

may

rece

ive

beca

use

ofth

eli

mit

edm

arke

tfo

rth

efi

rm’s

shar

esar

ecl

assi

fied

un

der

Prov

ide

liqui

dity

.R

epu

rch

ases

that

men

tion

min

imiz

atio

nof

shar

ep

rice

vola

tili

tyar

ecl

assi

fied

asR

educ

esh

are

pric

evo

lati

lity.

All

oth

eran

nou

nce

men

tre

ason

sar

eca

tego

rize

das

Oth

ers.



We provide descriptive statistics for the full sample and the subsamplesin Table 2; Panel A is for all repurchasers (both zero and positive repur-chasers) and Panel B is for positive repurchasers. There are 148 firms in thefull sample with 212 buyback program announcements. Of the 212announcements, we classify 47 programs as undervaluation and the rest asother motives repurchases. We observe in Panel B that the undervaluationsample firms seek to repurchase fewer shares than other motives firms. Theundervaluation firms repurchase fewer shares than other motives firms:2.31% versus 2.94% as a percentage of the outstanding shares leading to acompletion rate of 40.15% and 47.22%, respectively. The undervaluationgroup firms have fewer repurchase days (34 days versus 38 days) throughoutthe program.18

18 It is possible that the undervaluation firms are coming from specific industries, which maybias our subsequent results and conclusion. In an unreported table, we examine the numberof repurchase program announcing firms in each industry by repurchase motive. We observethat undervaluation firms come from the industries in which we also observe other repur-chase motives firms.

Table 2 Descriptive statistics

Variables Fullsample

Undervaluation Othermotives

Panel A: Full sampleNumber of firms 148 34 114

Positive repurchasers 132 31 101Zero repurchasers 16 3 13

Number of announcements 212 47 165Positive repurchasers 192 41 151Zero repurchasers 20 6 14

Panel B: Positive repurchasersNumber of daily repurchases 6.986 1.378 5.608Shares sought (percent of

outstanding)5.95 5.75 6.22

Shares repurchased (percent ofoutstanding)

2.78 2.31 2.94

Shares repurchased/sharessought (%)

46.65 40.15 47.22

Average repurchase days perfirm

37 34 38

We present descriptive statistics in this table. Panel A reports descriptive statistics for the fullsample of 148 companies that announced an open-market repurchase program between Septem-ber 1, 1998 and December 31, 2008. Panel B reports statistics only for positive repurchasers (i.e.,firms that repurchased shares following a repurchase program announcement). Shares sought is asof the percentage of outstanding shares.



VI. EMPIRICAL RESULTS

We examine repurchase program announcements in Section VI.A, actual repur-chase announcements in Section VI.B, and the long-term performances ofrepurchasing firms in Section VI.C.

A. Initial share repurchase program announcements

We examine abnormal returns surrounding program announcements in SectionVI.A.i, how firm and program characteristics are related to abnormal returns inSection VI.A.ii, and the relationship between firm and program characteristicsand stated repurchase motivation in Section VI.A.iii.

i. Abnormal returnsWe use the standard event study methodology to examine share price perfor-mance around both initial program and actual share repurchase announce-ments. We estimate abnormal returns using the market model with returns fromday -300 to day -46 before a program or an actual repurchase announcement.19

Day ‘0’ is the event day, which is either the initial program announcement dayor the actual repurchase announcement day. Our market index is the ASX AllOrdinaries Index.

We present the abnormal returns around initial program announcements inTable 3 for the full sample as well as different subsamples using five differentevent windows. We note in the first row for the full sample that repurchaseprogram announcements follow poor stock price performance. The market’sreaction to the announcements is significantly positive with an abnormalreturn of 3.06% in the (-1, +1) window and is similar to 3.3% reported in Lambaand Ramsay (2005) for Australia, but higher than 2.39% reported in Peyer andVermaelen (2009) for the United States between 1991 and 2001. Note thatunlike in the United States, in Australia firms must cancel the shares and cannottreat them as treasury shares. This implies a stronger initial response to repur-chase program announcements and weaker responses to actual daily repurchaseannouncements.20 The next two rows are for positive and zero repurchasers. Themean preannouncement abnormal return for the positive repurchasers group issignificantly negative, but not for the zero repurchasers group. However, themarket’s reaction to the announcements is significantly positive for both groups(3.01% and 3.49%).

We next examine the market’s reaction to announcements by undervalua-tion and other motives firms. The mean CAR for positive repurchasers is insig-nificantly negative prior to the announcements (-0.49%), significantly positive

19 The literature shows that the abnormal returns are not sensitive to the model used toestimate normal returns when daily stock returns are used for event study (Brown andWarner 1985; Campbell et al. 1997). We repeat our analysis with market-adjusted returns andobtain similar results.

20 We thank the Referee for this point.



(4.29%) around the announcements, and insignificantly positive after theannouncements (2.99%). Zero repurchasers on average do not have an under-performance problem prior to the announcements. The market’s reaction to theannouncements is insignificantly positive around the announcements (1.96%).It seems that the market’s overall reaction to repurchase announcements byundervaluation group firms indicates which firms are more likely to buy backshares subsequently.

In the rest of the table, we examine other motives repurchasers. Positiverepurchasers underperform the market prior to announcements as evidenced bya statistically significant CAR of -1.24%. The overall CAR for the (-1, +1)window is 2.67% (p value = 0.00). We note that market reacts more positively tothe announcements by undervaluation firms (4.29% versus 2.67%), whichindicates that management statements may contain new information for themarket.21 The CAR for zero repurchasers is insignificantly negative (-0.22%)prior to and significantly positive around program announcements (4.15% witha p value of 0.00). Compared with 4.15% for zero repurchasers, 2.67% is lower,but the difference is not statistically significant.

21 The difference, however, is not statistically significant.

Table 3 Abnormal returns around program announcements

Sample Event window

(-11, -2) (-1, -1) (0, +1) (-1, +1) (+2, +11)

All announcements (212) (%) -0.91c 0.21 2.85a 3.06a 1.33(0.09) (0.17) (0.00) (0.00) (0.11)

Positive Repurchasers (192) (%) -1.08c 0.14 2.87a 3.01a 1.41(0.06) (0.32) (0.00) (0.00) (0.16)

Zero Repurchasers (20) (%) 0.72 0.87 2.62a 3.49a 0.55(0.63) (0.16) (0.00) (0.00) (0.45)

Undervaluation motiveannouncements (47) (%)

-0.06 1.04b 2.95a 3.99a 2.48(0.69) (0.02) (0.00) (0.00) (0.18)

Positive repurchases (41) (%) -0.49 1.09b 3.20a 4.29a 2.99(0.47) (0.03) (0.00) (0.00) (0.11)

Zero repurchases (6) (%) 2.90 0.70 1.26 1.96 -0.99(0.44) (0.56) (0.33) (0.26) (0.71)

Other motives announcements(165) (%)

-1.15c -0.03 2.82a 2.79a 1.00(0.09) (0.74) (0.00) (0.00) (0.28)

Positive repurchases (151) (%) -1.24b -0.12 2.79a 2.67a 0.98(0.02) (0.90) (0.00) (0.00) (0.73)

Zero repurchases (14) (%) -0.22 1.00 3.15a 4.15a 1.21(0.94) (0.20) (0.00) (0.00) (0.25)

We report cumulative abnormal returns around share repurchase program announcements in thistable. We use the standard event study methodology based on the market model. The marketindex is the ASX All Ordinaries Index. The estimation period is from 300 to 46 days prior to theprogram announcement and day 0 is the announcement day. The Patell Z-test is used for thesignificance of means. p Values are provided in parentheses. a, b, and c denote statisticalsignificance at the 1%, 5%, and 10% levels.



Overall, we observe in Table 3 that announcements are welcomed by themarket as the market reaction to announcements is significantly positive.However, we note that the market reacts more positively to the announce-ments by undervaluation firms that subsequently repurchase shares, whichsuggests that management statements may create additional value forshareholders.

ii. Determinants of market reaction to program announcementsWe now turn our attention to how abnormal returns are related to firm andrepurchase characteristics in Table 4. We regress abnormal returns from the (-1,+1) window on several firm and repurchase characteristics: Undervaluation is anindicator variable equal to 1 if the repurchase reason is undervaluation. Zerorepurchaser is an indicator variable equal to 1 if the firm does not repurchase anyshares after the initial program announcement. Program duration is the length ofthe repurchase program in years. Repeat is an indicator variable equal to 1 if thefirm announced another repurchase program within the prior 15 months. Sizequartile is the relative size of the firm compared with all the ASX firms on theprogram announcement date (Quartile 1 is for small firms and 4 is for largefirms). Firm size is calculated as the market value of equity. Market-to-book is therelative market-to-book of the firm compared with all ASX firms on the programannouncement date (Quartile 1 is for small market-to-book firms and 4 is forlarge market-to-book firms). Market-to-book is calculated as the market value ofequity divided by book-value of equity. Program size is the ratio of sharesrepurchased to the number of shares outstanding. CAR (-11, -2) is the cumu-lative abnormal return during the 10-day period preceding the programannouncement day. We control for year and industry effects in all models andcluster standard errors at the firm level.

There are three models with slightly different control variables in Table 4.However, no variable except size quartile is significant in any of the models.Thus, the market reacts more positively to program announcements made bysmall firms. Although insignificant, we observe that undervaluation, repeat,program size have positive coefficients in all three models and recent share priceperformance [CAR (-11, -2)] is negatively related to abnormal returns.

iii. Repurchase motivation and program announcementsWe now ask the question of what makes firms choose undervaluation as arepurchase motivation over other motivations and whether firms that do notsubsequently repurchase shares have different characteristics than other firms.To answer the first question, we regress the undervaluation dummy on firm andrepurchase characteristics in the first three columns of Table 5. Similar to theresults in Table 4, the size of a firm is the only significant variable related to thechoice of repurchase motivation. Thus, small firms are more likely to chooseundervaluation over other repurchase motivations. This is consistent with thenotion that small firms have greater information asymmetry than large firms, asthey receive less analyst and media coverage.



In the last three columns of the table, we try to answer the second questionand regress an indicator variable for zero repurchasers on firm and repurchasecharacteristics. We find that firms that have announced another repurchase inthe prior 15 months are less likely to repurchase any shares under the currentprogram. Also, large firms or firms with large repurchase programs are less likelyto buy any shares. Finally, the program duration variable has a significantlynegative sign, indicating that firms with shorter durations are less likely to buyany shares.

Table 4 Determinants of market reaction

Variables (1) (2) (3)

Undervaluation 0.004 0.004 0.004(0.72) (0.78) (0.77)

Zero repurchaser 0.017 0.016(0.42) (0.45)

Program duration -0.005(0.57)

Repeat 0.008 0.006 0.005(0.55) (0.63) (0.67)

Size quartile -0.020a -0.020a -0.020a

(0.01) (0.01) (0.02)Market-to-book quartile 0.001 0.000 0.000

(0.90) (0.97) (0.99)Program size 0.107 0.099 0.116

(0.49) (0.54) (0.49)CAR (-11, -2) -0.098 -0.097 -0.098

(0.33) (0.33) (0.33)Intercept 0.050 0.052 0.050

(0.18) (0.17) (0.19)Industry dummies Yes Yes YesYear dummies Yes Yes YesR-squared 0.12 0.12 0.12N 212 212 212

We examine how abnormal returns around share repurchase announcements are related to firmand repurchase program characteristics in this table. The dependent variable in all models is theabnormal return from the (-1, +1) window. Undervaluation is an indicator variable equal to 1 if therepurchase reason is undervaluation. Zero repurchaser is an indicator variable equal to 1 if the firmdoes not repurchase any shares after the initial program announcement. Program duration is thelength of the repurchase program in years. Repeat is an indicator variable equal to 1 if the firmannounced another repurchase program within the prior 15 months. Size quartile is the relativesize of the firm compared with the all ASX firms on the program announcement date (Quartile1 is for small firms and 4 is for large firms). Firm size is calculated as the market value of equity.Market-to-book is the relative market-to-book of the firm compared with all ASX firms on theprogram announcement date (Quartile 1 is for small market-to-book firms and 4 is for largemarket-to-book firms). Market-to-book is calculated as the market value of equity divided by bookvalue of equity. Program size is ratio shares repurchased to the number of shares outstanding. CAR(-11, -2) is the cumulative abnormal return during the 10-day period preceding the programannouncement day. N is the number of observations. All models include year and industrydummies. p Values are reported in parentheses. Standard errors are clustered at the firm level. adenotes statistical significance at the 5% level.CAR, cumulative abnormal return.



Overall, we find that firm size is a good indicator of whether a firmannounces a buyback program claiming its share price is undervalued, and theresults also indicate which firm and program characteristics are associated withfirms that do not subsequently buy back any shares.

Table 5 Repurchase motivation and nonrepurchasers

Variables Undervaluation Zero repurchaser

(1) (2) (3) (1) (2) (3)

Undervaluation 0.108 0.100(0.81) (0.82)

Zero repurchaser 0.431 0.462(0.37) (0.33)

Program duration 0.123 -0.984(0.56) (0.02)b

Repeat -0.268 -0.304 -0.292 0.842 0.845 0.846(0.23) (0.16) (0.18) (0.04)b (0.04)b (0.04)b

Size quartile -0.310 -0.332 -0.354 0.674 0.681 0.767(0.04)b (0.04)b (0.03)b (0.02)b (0.02)b (0.01)b

Market-to-book quartile 0.011 -0.017 -0.016 0.654 0.641 0.686(0.95) (0.91) (0.92) (0.00)a (0.00)a (0.00)a

Program size 0.793 0.739 0.304 8.179 8.216 10.858(0.76) (0.78) (0.91) (0.06)c (0.06)c (0.02)b

CAR (-11, -2) 0.532 0.533 0.556 -2.271 -2.266 -2.371(0.51) (0.52) (0.50) (0.17) (0.18) (0.14)

Intercept 0.145 0.172 0.184 -4.598 -4.612 -4.643(0.86) (0.83) (0.82) (0.00)a (0.00)a (0.01)b

Industry dummies Yes Yes Yes Yes Yes YesYear dummies Yes Yes Yes Yes Yes YesR-squared 0.10 0.12 0.16 0.32 0.32 0.36N 205 205 205 153 153 153

In this table, we examine, using Probit models, how the choice of undervaluation as a repurchasemotive depends on firm and program characteristics and also if nonrepurchasing firms havecharacteristics that separate them from other firms. The dependent variable in the first threemodels is an indicator variable equal to 1 if the repurchase reason is undervaluation. Thedependent variable in the last three models is an indicator variable equal to 1 if the firm does notrepurchase any shares after the initial program announcement. Program duration is the length ofthe repurchase program in years. Repeat is an indicator variable equal to 1 if the firm announcedanother repurchase program within the prior 15 months. Size quartile is the relative size of thefirm compared with all ASX firms on the program announcement date (Quartile 1 is for smallfirms and 4 is for large firms). Firm size is calculated as the market value of equity. Market-to-bookquartile is the relative market-to-book of the firm compared with all ASX firms on the programannouncement date (Quartile 1 is for small market-to-book firms and 4 is for large market-to-book firms). Market-to-book is calculated as the market value of equity divided by book value ofequity. Program size is ratio shares repurchased to the number of shares outstanding. CAR (-11, -2)is the cumulative abnormal return during the 10-day period preceding the program announce-ment day. N is the number of observations. All models include year and industry dummies. pValues are reported in parentheses. Standard errors are clustered at the firm level. a, b, and cdenote statistical significance at the 1%, 5%, and 10% levels.CAR, cumulative abnormal return.



B. Actual repurchases

In this section, we examine actual daily share repurchases. We provide somesummary statistics for actual repurchases in Section VI.B.i and report abnormalreturns in Section VI.B.ii. We then examine the relationship between firm andrepurchase characteristics and the market’s reaction to actual repurchases inSection VI.B.iii. Section VI.B.iv and Section VI.B.v examine monthly repurchaseactivity and program completion rates, respectively.

i. Summary statisticsWe provide summary statistics for actual daily repurchases in Table 6. There are6986 reported daily repurchases in our sample. Panel A reports average cumu-lative buyback activity following program announcements up to 18 months.22

We report cumulative buyback activity both as a percentage of program size andthe number of shares outstanding.

Firms in our sample buy back about 12.89% of the program size (0.66% ofoutstanding shares) in the first month. At the end of the 12-month periodfollowing program announcements, this percentage increases to 43.31%(2.66%). We are, however, interested in if undervaluation firms have differentrepurchase patterns than other motive firms. Undervaluation motive firms buyfewer shares in the first month compared with other motive firms (10.70%versus 13.49%). At the end of the 18th month, other motives firms will havebought 46.94% of the program size compared with 39.07% for undervaluationfirms.

It looks like undervaluation firms are repurchasing fewer shares than othermotives firms. The literature shows that managers time the market when buyingback shares in the market and share price performance may explain if under-valuation firms are responding to their share price underperformance duringthe repurchase program. We report average excess cumulative stock pricereturns following initial program announcements in Panel B. At the end of thefirst month after the program announcement, the average share price perfor-mance of an undervaluation firm is 4.73% and that of another motives firm is3.76%. We do not observe any underperformance issue in the first month of theprogram. However, note that these firms are buying back more than 10% of theprogram size in the first month; thus, we are seeing the positive effect ofrepurchase activities in the first month. By the end of 18th month, undervalu-ation firms perform much better than other motives firms (5.38% versus0.64%). Compared with Panel A, it seems other motives firms are buying moreshares as their stock price deteriorates, but not undervaluation firms. We illus-trate this point in Figure 1 by plotting cumulative monthly buyback activityand stock price returns for undervaluation and other motives firms. Graph Adoes not indicate a strong correlation between stock returns and monthlybuyback activity. Graph B, however, indicates a strong negative correlation forother motives firms. In Panel C of Table 6, we report correlation coefficients

22 Most firms in the sample have a program duration of 1 year.



Table 6 Summary statistics for actual daily repurchases

Panel A: Average cumulative repurchase activity (%)

Month(s) after program announcement

1 3 6 9 12 15 18

All programs (192)Program size 12.89 28.62 37.58 42.05 43.31 44.42 45.13Shares outstanding 0.66 1.36 2.05 2.54 2.66 2.71 2.77

Undervaluation programs (41)Program size 10.70 26.26 30.57 33.86 35.92 38.47 39.07Shares outstanding 0.39 1.05 1.59 1.87 2.04 2.25 2.29

Other motive programs (151)Program size 13.49 29.26 39.52 44.29 45.41 46.17 46.94Shares outstanding 0.74 1.45 2.18 2.72 2.83 2.84 2.91

Panel B: Average excess cumulative returns (%)


1 3 6 9 12 15 18

All programs (192) 3.94 1.61 1.46 -3.21 -1.72 -1.34 1.73Undervaluation programs (41) 4.73 1.50 7.89 1.73 2.78 1.63 5.38Other motive programs (151) 3.72 1.64 -1.99 -4.56 -3.00 -2.19 0.64

Panel C: Correlations

Program size Shares outstanding

All programs (192) -0.73 -0.74Undervaluation programs (41) -0.17 -0.12Other motive programs (151) -0.80 -0.79

Panel D: Average duration (days)


1 3 6 9 12 15 18

All programs (192) 22 45 73 90 100 106 108Undervaluation programs (41) 21 41 65 77 88 99 103Other motive programs (151) 22 45 75 93 104 108 110

We provide summary statistics on the repurchasing patterns in this table. There are 6986 dailyrepurchases. Panel A presents the monthly average number of shares bought after the initialrepurchase program announcement as a percentage of the program size and shares outstanding.Average excess cumulative stock returns are provided in Panel B. The correlation coefficientsbetween the actual monthly repurchase activities in Panel A and cumulative stock returns inPanel B are presented in Panel C. Finally, the average duration of actual repurchases in days,defined as the weighted average of repurchase time scaled by the number of shares repurchased,is presented in Panel C.



between monthly cumulative repurchase activity and stock returns. There is aweak negative relationship between monthly repurchase activity and excessstock returns for undervaluation firms as the correlation coefficient ranges from-0.12 to -0.17. The correlation coefficient for other motives firms, however, is-0.80, suggesting that these firms buy back more shares as their stock price falls.

0

5

10

15

20

25

30

35

40

45

1 3 6 9 12 15 18

Rep

urch

ase

(%)

/ Sto

ck R

etur

n (%

)

Month

Cumulative Repurchase (%) Cumulative Excess Stock Return (%)

-10

0

10

20

30

40

50

1 3 6 9 12 15 18Rep

urch

ase

(%)

/ Sto

ck R

etur

n (%

)

Month

Cumulative Repurchases (%) Cumulative Excess Stock Return (%)

(a)

(b)

Figure 1 Cumulative Repurchase Activity and Stock Return.This figure shows overall monthly cumulative repurchase activity and cumulativeexcess stock returns for undervaluation (Graph A) and other motives programs (GraphB) up to 18 months after the initial share repurchase program announcement. Allvalues are in percentages.(a) Undervaluation programs.(b) Other motives programs.



Finally, we examine the duration of actual repurchases in days, defined as theweighted average of repurchase time scaled by the number of shares repur-chased, in Panel D. We note that in the first month, the duration is almost thesame for undervaluation and other motives firms. The duration, however, ishigher for other motives firms by the end of 18th month (110 days versus103 days).

The table, overall, shows that undervaluation firms buy back fewer sharesunder the repurchase program. We also note that there is a weak relationshipbetween monthly repurchase activity and stock returns for undervaluationfirms. Other motives firms, however, seem to be buying in response to fallingshare prices throughout the repurchase program.

ii. Market reaction to actual repurchase announcementsIn this part, we focus on positive repurchasers and examine their actual repur-chase announcements. We mostly follow Zhang (2005) to identify event daysrelated to actual repurchases. If there is only a single repurchase in a month, thefollowing day is defined as an event day. However when there are multiplerepurchases within a month by a sample firm, we identify the first repurchaseday of each month and define the following day (the announcement day) as anevent day. This is because if we consider every repurchase announcement day asan event day, then firms with frequent repurchases will have too much weightin the sample portfolio returns. Conversely, if we only consider the first repur-chase announcement day of the repurchase program as an event day, then theinformation contained in the subsequent repurchases will be overlooked. Inaddition, if managers are signaling information to the market via share repur-chases, then subsequent repurchases are important for investors to confirm thereliability of the signal. In total, we have 927 daily actual repurchase announce-ments and 120 of them are by undervaluation firms.

We begin our analysis by first examining CARs for the undervaluation andother motives samples in Figure 2. We measure CARs from day -11 to day +11around repurchase days. Figure 2 shows that other motives firms underperformthe market starting 10 days before the event date, indicating that stock priceunderperformance is a potential reason for actual repurchases. After the repur-chase on day -1, the CAR starts to improve until day 3 and then levels off.However, the CAR line for undervaluation firms paints a different picture. Westill observe that firms that cite undervaluation as the reason for their repur-chases also underperform the market during the 10-day period before the eventday, albeit at a lower level than the other motives firms. However, it seems theundervaluation sample firms repurchase in response to a sudden fall in the stockprice on day -2. The CAR after day -1 starts to improve and by day 3, it becomespositive and finishes above zero on day +11. Compared with the CAR line forother motive firms, the line for undervaluation firms suggests that the market istaking into account the reason for the buyback and responds accordingly. Weobserve that underperformance is corrected for undervaluation firms by day+11, but not for other motives firms.



We formally report the results of the event study in Table 7 for the full,undervaluation, and other motives samples. Similar to Table, 3, there are fiveevent windows with different lengths. The window (-11, -2) is used to examinewhether firms tend to repurchase shares when their stock prices underperformthe market, thus time the market. We use the event window (-1, -1) to inves-tigate whether the abnormal return on the actual repurchase day is negative. Ifan actual share repurchase has information content, the market reaction could

-2.00%

-1.50%

-1.00%

-0.50%

0.00%

0.50%

1.00%

1.50%

-11 -10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10 11

Cum

ulat

ive

Abn

orm

al R

etur

n

Days

Undervaluation Other Motives

Figure 2 Cumulative Abnormal Returns.This figure presents cumulative abnormal returns from day -11 to day 11 around sharerepurchase announcement days (day 0). All values are in percentages.

Table 7 Cumulative abnormal returns around actual repurchases

Sample Event windows

(-11, -2) (-1, -1) (0, +1) (-1, +1) (+2, +11)

Full sample (927) (%) -1.15a -0.36a 0.43a 0.07 0.10(0.00) (0.00) (0.00) (0.36) (0.47)

Undervaluation (120) (%) -1.08 -0.03 1.02b 0.99 1.30b

(0.20) (0.20) (0.01) (0.08)c (0.02)Other motives (807) (%) -1.16a -0.41a 0.34a -0.07 -0.07

(0.00) (0.00) (0.00) (0.18) (0.23)

This table reports cumulative abnormal returns for various event windows. The event day (day‘0’) is the announcement day following the first repurchase day of each month. There are 927,120, and 807 event days for the full sample, undervaluation, and other motives subsamples.Abnormal returns are calculated using the market model. The market index is the ASX AllOrdinaries Index. The estimation period is from 300 to 46 days prior to the event day. The PatellZ test is used for the significance of means. p Values are provided in parentheses. a, b, and c denotestatistical significance at the 1%, 5%, and 10% levels.



happen on the actual repurchase day or on the announcement day (Rees 1996).The former is possible if: (1) market participants detect the presence of the firm;(2) abnormal trading volume is regarded as news in itself; or (3) increaseddemand for the stock puts an upward pressure on the share price (Rees 1996). Asthe announcement can be made after-hours on the repurchase day or by9:30 AM on the following business day, the market will respond to the repur-chase announcement on day 0. The window (+0, +1) is used to capture theinitial market reaction to the actual share repurchases. We also include the (-1,+1) window that includes both the actual repurchase and announcement days.Finally, CAR (+2, +11) is used to examine the short-term market responsefollowing the initial market reaction.

The results for the full sample show that repurchasing firms underperformthe market prior to the repurchase day. The abnormal return is significantlynegative (-1.15%) for the (-11, 2) window and the market’s reaction to actualrepurchase announcements is significantly positive (0.43%) at the 1% level.However, there is no significant market reaction in the postannouncementperiod.

The undervaluation subsample CAR for the (-11, -2) window is an insignifi-cant -1.08%. However, note that most of the fall in the stock price is observedon day -2. In an unreported table, we find that the abnormal return on day -2is a statistically significant -1.07% (p value = 0.00). Thus, undervaluationmotive firms are also buying back when their stock prices are falling, but theymostly react to a sudden fall in the stock price. On the repurchase day, theabnormal return is an insignificant -0.03%. The 2-day return following therepurchase day is 1.02% and significant at the 5% level. During the following10 days, we observe a significantly positive CAR (1.30%) for the undervaluationsample compared with the same window for the other motives subsample.

Other motives firms significantly underperform the market during the 10-dayperiod before the repurchase. The CAR is -1.16% and significant at the 1% level.The abnormal return on the repurchase day is negative (-0.41%) and significantat the 1% level. When the market learns about the repurchase on the next2 days, there is a significantly positive reaction of 0.34%. The differencebetween 1.02% for the undervaluation sample and 0.34% is statistically signifi-cant with a p value of 0.03. However, during the following 10 days, in the (+2,+11) window, these firms continue to underperform the market. The CAR forthis period is an insignificant -0.07%.

Overall, similar to the results in Table 3, the results show that firms arebuying back in response to a fall in their share price and the market is respond-ing more positively to actual repurchases when the stated reason of a repurchaseis undervaluation. These results are consistent with our first and second hypoth-eses. Although the traditional theories of signaling (Bhattacharya 1979; Ver-maelen 1984; Miller and Rock 1985) do not attach any value to managementstatements due to the fact that the signal is not credible, the results of Peyer andVermaelen (2009) suggest otherwise. Their result indicates that managementstatements of repurchase reasons are valuable, at least those that state



undervaluation. Our results so far are in line with those of Peyer and Vermaelenin that we find that the market reaction to actual share repurchases is morepositive if the repurchase reason is undervaluation. Our results, thus, constitutean out-of-sample confirmation for those of Peyer and Vermaelen. Our resultsfurther suggest that management is repurchasing shares because it does notagree with the market’s value for the firm, which is further supported by therepurchases following a sudden decline in the stock prices of undervaluationmotive firms on day -2. The market seems to be viewing management’s signalas credible and is reacting accordingly. When the stated reason for the repur-chase is anything but undervaluation, the market does not seem to be respond-ing more positively. The underreaction to other repurchase motives mayindicate that there is no new information disclosed to the market. As therepurchase may not alter the market’s assessment of the firms, the market maynot be responding to the repurchases of other motives firms more positively.

iii. Analysis of actual repurchasesWe will now examine in Table 8 how firm and repurchase characteristics affectthe market’s reaction to actual repurchases. We regress CARs from threewindows – (-1, -1), (0, +1) and (-1, +1) – on firm and repurchase characteristics.We add the (-1, +1) as the market’s response may start on day -1 as argued byRees (1996).

Specifically, we regress cumulative abnormal returns on an indicator variablewhether the repurchase reason is undervaluation, size and market-to-bookquartiles of the firm (both determined on the actual repurchase day), repurchasesize (calculated as the ratio of shares purchased to the number of shares out-standing), repurchase lag (calculated as the number of days since the lastrepurchase), and the number of previous repurchases in the last 3 months underthe same repurchase program. The last two control variables are added follow-ing Zhang (2005), who argues that if a repurchase is a surprise to the market,then the market should react more favorably. We also include CAR (-11, -2) inthe regressions to control for potential mean reversion problem. Standard errorsare clustered at the firm level and we control for year and industry effects in allregressions. Repurchase lag is zero when the firm announces its buybackprogram. Therefore, we do not include the first repurchase day of the firstmonth in our analysis which reduces the sample by 191 observations and weend up with 736 observations.

There are three models in Table 8 for the full, undervaluation, and othermotives samples. We include the undervaluation dummy in the models for thefull sample. The coefficient for CAR (-11, -2) is negative and statisticallysignificant at the 1% level in the first and third models, indicating that moreunderperformance prior to the repurchase leads to more positive returns on therepurchase day. Additionally, repurchase size and the number of prior actualrepurchases are positively and significantly related to abnormal returns onthe repurchase day in the first model, which indicates that the structure ofa repurchase program also affects the market’s reaction to the actual share



Tab

le8

Dai

lysh

are

rep

urc

has

es

Var

iabl

esFu

llsa

mp

leU

nd

erva

luat

ion

Oth

erm

otiv

es

(-1,

-1)

(0,

+1)

(-1,

+1)

(-1,

-1)

(0,

+1)

(-1,

+1)

(-1,

-1)

(0,

+1)

(-1,

+1)

Un

der

valu

atio

n0.

007

0.00

50.

012b

(0.2

4)(0

.29)

(0.0

3)Si

zeq

uar

tile

0.00

1-0

.001

0.00

00.

007

-0.0

020.

005

0.00

2-0

.002

0.00

0(0

.63)

(0.5

1)(0

.88)

(0.5

1)(0

.84)

(0.6

9)(0

.23)

(0.4

3)(0

.94)

Mar

ket-

to-b

ook

qu

arti

le0.

000

-0.0

01-0

.002

0.01

0-0

.001

0.01

00.

000

-0.0

03c

-0.0

02(0

.87)

(0.3

9)(0

.40)

(0.1

0)(0

.94)

(0.3

1)(0

.75)

(0.0

9)(0

.28)

Rep

urc

has

esi

ze0.

435c

-0.3

940.

041

1.95

8-0

.759

1.19

90.

589b

-0.2

340.

355

(0.0

9)(0

.46)

(0.9

5)(0

.17)

(0.2

7)(0

.37)

(0.0

2)(0

.71)

(0.6

2)R

epu

rch

ase

lag

-0.0

120.

017

0.00

60.

059

-0.0

010.

058

-0.0

400.

001

-0.0

39(0

.70)

(0.2

4)(0

.88)

(0.2

9)(0

.98)

(0.3

5)(0

.15)

(0.9

1)(0

.15)

Nu

mbe

rof

rep

urc

has

es0.

028a

-0.0

130.

015

0.08

5c0.

036

0.12

10.

018b

-0.0

15b

0.00

3(0

.00)

(0.1

1)(0

.31)

(0.0

8)(0

.46)

(0.1

4)(0

.02)

(0.0

4)(0

.78)

CA

R(-

11,

-2)

-0.1

54a

-0.0

12-0

.166

b-0

.344

b-0

.212

a-0

.556

a-0

.109

a0.

024

-0.0

85b

(0.0

0)(0

.68)

(0.0

1)(0

.02)

(0.0

0)(0

.00)

(0.0

0)(0

.33)

(0.0

1)In

terc

ept

-0.0

22b

0.01

8c-0

.004

-0.0

79b

0.01

5-0

.064

c-0

.027

a0.

026b

-0.0

01(0

.02)

(0.0

6)(0

.80)

(0.0

3)(0

.48)

(0.0

5)(0

.00)

(0.0

3)(0

.95)

Ind

ust

ryd

um

mie

sYe

sYe

sYe

sYe

sYe

sYe

sYe

sYe

sYe

sYe

ard

um

mie

sYe

sYe

sYe

sYe

sYe

sYe

sYe

sYe

sYe

sR

-sq

uar

ed0.

140.

040.

090.

480.

270.

550.

110.

050.

05N

736

736

736

8888

8864

864

864

8

Th

ista

ble

rep

orts

the

resu

lts

ofth

ere

gres

sion

sin

wh

ich

cum

ula

tive

abn

orm

alre

turn

sar

oun

dac

tual

shar

ere

pu

rch

ases

are

regr

esse

don

firm

and

rep

urc

has

ech

arac

teri

stic

s.T

he

dep

end

ent

vari

able

sar

eC

AR

sfr

omth

e(-

1,-1

),(0

,+1)

,an

d(-

1,+1

)w

ind

ows.

Th

em

odel

sar

ees

tim

ated

for

the

full

sam

ple

asw

ella

sth

eu

nd

erva

luat

ion

and

oth

erm

otiv

essu

bsam

ple

s.U

nder

valu

atio

nis

anin

dic

ator

vari

able

equ

alto

1if

the

rep

urc

has

ere

ason

isu

nd

erva

luat

ion

.Siz

equ

arti

leis

the

rela

tive

size

ofth

efi

rmco

mp

ared

wit

hal

lASX

firm

son

the

ann

oun

cem

ent

dat

e(Q

uar

tile

1is

for

smal

lfirm

san

d4

isfo

rla

rge

firm

s).

Firm

size

isca

lcu

late

das

the

mar

ket

valu

eof

equ

ity.

Mar

ket-

to-b

ook

quar

tile

isth

ere

lati

vem

arke

t-to

-boo

kof

the

firm

com

par

edw

ith

all

ASX

firm

son

the

ann

oun

cem

ent

dat

e(Q

uar

tile

1is

for

smal

lm

arke

t-to

-boo

kfi

rms

and

4is

for

larg

em

arke

t-to

-boo

kfi

rms)

.M

arke

t-to

-boo

kis

calc

ula

ted

asth

em

arke

tva

lue

ofeq

uit

yd

ivid

edby

book

valu

eof

equ

ity.

Rep

urch

ase

size

isra

tio

shar

esre

pu

rch

ased

toth

en

um

ber

ofsh

ares

outs

tan

din

g.R

epur

chas

ela

gis

the

nu

mbe

rof

day

ssi

nce

the

last

rep

urc

has

eu

nd

erth

esa

me

pro

gram

div

ided

by36

5.N

umbe

rof

repu

rcha

ses

isth

en

um

ber

ofre

pu

rch

ases

inth

ela

st3-

mon

thp

erio

du

nd

erth

esa

me

pro

gram

div

ided

by90

.C

AR

(-11

,-2

)is

the

cum

ula

tive

abn

orm

alre

turn

du

rin

gth

ep

ast

10-d

ayp

erio

dp

rece

din

gth

eac

tual

rep

urc

has

ed

ay.

Nis

the

nu

mbe

rof

obse

rvat

ion

s.A

llm

odel

sin

clu

de

year

and

ind

ust

ryd

um

mie

s.p

Val

ues

are

rep

orte

din

par

enth

eses

.St

and

ard

erro

rsar

ecl

ust

ered

atth

efi

rmle

vel.

a,b,

and

cd

enot

est

atis

tica

lsi

gnifi

can

ceat

the

1%,

5%,

and

10%

leve

ls.

CA

R,

cum

ula

tive

abn

orm

alre

turn

.



repurchases. The undervaluation dummy is significantly positive in the lastmodel only, which provides some support for the argument that undervalua-tion motive sends a credible and a strong signal to the market and those firmsthat use the undervaluation motive experience more positive returns.

In the last two sets of regressions, we investigate how abnormal returns arerelated to firm and repurchase characteristics separately for the undervaluationand other motives subsamples. We find that CAR (-11, -2) and the number ofrepurchases have statistically significant coefficients in the undervaluation sub-sample. However, we do not find that market-to-book quartile is significantlyrelated to abnormal returns. This finding, which is in line with our earlierfindings, indicates that the market is paying attention to the stated reason ofthe repurchase rather than the level of potential undervaluation.

For the other motives subsample, CARs are significantly affected by priorstock price performance, repurchase size, and the number of repurchases as inthe full sample. However, this time market-to-book is significantly negative inthe second model, indicating that the market takes into account the firm’spotential undervaluation when reacting to the actual repurchase announce-ment by other motives firms.

Our results in Table 8 suggest that repurchase reasons send different signals tothe market. When the stated reason of the repurchase is undervaluation, themarket reaction is overall more positive than when the repurchase reason isanother reason.

iv. Monthly repurchase activitySo far, we have not examined what affects how much firms buy back in a givenmonth. This is an important question as firms may be buying back shares inresponse to recent share price performance.

In Table 9, we regress the monthly fraction of shares bought back under theprogram (relative to shares outstanding) on various firm and program charac-teristics in the first two models. In the last two models, we use a dummy variableas the dependent variable that equals one if the firm makes a repurchase in thatmonth. We follow the firms from the time of the initial program announcementto the end of the year following the announcement. Following Ikenberry et al.(2000), we add excess returns from previous and following months as well as theexcess returns in the month of interest to the model.

We find that excess returns from the previous and current month are goodpredictors of how much firms are likely to purchase during the month. Under-performing firms are more likely to repurchase shares in the market. In thesecond model, we add the undervaluation dummy and find that undervalua-tion firms are significantly less likely to repurchase shares in the market follow-ing the initial program announcement, which is in line with our earlierfindings.

The next two models use the monthly repurchase dummy. We find that largefirms or firms with large programs are more likely to repurchase shares. We addthe undervaluation dummy in the last model. Although insignificant, the



coefficient on the undervaluation dummy is negative and consistent with theone from the model in the second column. Excess return variables, althoughmostly negative, are not significant.

Consistent with our earlier findings, we find that undervaluation firmsare less likely to repurchase shares in the market following the program

Table 9 Determinants of monthly repurchase activity

Variables Monthly percentage Monthly dummy

(1) (2) (1) (2)

Undervaluation -0.014a -0.101(0.00) (0.39)

Repeat -0.001 -0.003 0.014 0.005(0.82) (0.66) (0.91) (0.97)

Size quartile 0.003 0.002 0.194a 0.185a

(0.20) (0.41) (0.00) (0.00)Market-to-book quartile 0.000 0.000 -0.035 -0.038

(0.98) (0.86) (0.56) (0.53)Program size -0.006 0.001 4.171a 4.223a

(0.91) (0.98) (0.00) (0.00)Excess return (previous month) -0.046b -0.045b -0.241 -0.236

(0.01) (0.01) (0.29) (0.30)Excess return (current month) -0.046b -0.046b -0.293 -0.290

(0.01) (0.01) (0.20) (0.20)Excess return (following month) -0.011 -0.010 0.163 0.170

(0.54) (0.56) (0.58) (0.56)Intercept 0.093a 0.103a -0.067 0.000

(0.00) (0.00) (0.93) (1.00)Industry dummies Yes Yes Yes YesYear dummies Yes Yes Yes YesR-squared 0.02 0.03 0.04 0.04N 2168 2168 2168 2168

In this table, we analyze monthly repurchase activity during the first 12 months after the initialprogram announcement. Firms that buy back more than 90% of their program size in givenmonth are excluded. The dependent variable in the first two models is the monthly percentageof shares bought back and in the last two models is an indicator variable equal to 1 if the firmbuys back in a given month. Undervaluation is an indicator variable equal to 1 if the repurchasereason is undervaluation. Repeat is an indicator variable equal to 1 if the firm announced anotherrepurchase program within the prior 15 months. Size quartile is the relative size of the firmcompared with all ASX firms on the program announcement date (Quartile 1 is for small firmsand 4 is for large firms). Firm size is calculated as the market value of equity. Market-to-book is therelative market-to-book of the firm compared with all ASX firms on the program announcementdate (Quartile 1 is for small market-to-book firms and 4 is for large market-to-book firms).Market-to-book is calculated as the market value of equity divided by book value of equity.Program size is ratio shares repurchased to the number of shares outstanding. Excess returns arecumulative stock returns net of the market return. Excess returns are calculated for the previous,current, and following month. N is the number of observations. All models include year andindustry dummies. p Values are reported in parentheses. Standard errors are clustered at the firmlevel. a and b denote statistical significance at the 1% and 5% levels.



announcement and firms time the market as they are buying back when theirshare price is underperforming.

v. Program completion ratesThe literature shows that firms do not buy back the full number of shares theydisclose in the buyback announcement (Bonaimé 2010). In this section, weexamine what factors affect program completion rates.

In Table 10, we regress the percentage of shares bought back (relative to thenumber of shares outstanding) during the repurchase program on firm andrepurchase characteristics. There are five models in the table with slightly

Table 10 Determinants of program completion rates

Variables (1) (2) (3) (4) (5)

Undervaluation -0.164b -0.167b -0.148b

(0.02) (0.02) (0.03)Zero repurchaser -0.442a

(0.00)Repeat -0.031 -0.047 -0.059 -0.069 -0.030

(0.66) (0.50) (0.35) (0.26) (0.62)Size quartile 0.060 0.046 0.039 0.025 0.048

(0.13) (0.25) (0.33) (0.54) (0.22)Market-to-book quartile -0.040 -0.044 -0.064c -0.063c -0.039

(0.28) (0.23) (0.08) (0.07) (0.26)Program size -0.674 -0.610 -0.829 -0.797 -0.576

(0.46) (0.49) (0.37) (0.36) (0.48)Excess return (previous year) -0.020 -0.011 -0.039 -0.027 -0.009

(0.81) (0.90) (0.62) (0.73) (0.91)Intercept 0.534b 0.639a 0.530b 0.614b 0.564a

(0.03) (0.00) (0.03) (0.01) (0.00)Industry dummies Yes Yes Yes Yes YesYear dummies Yes Yes Yes Yes YesR-squared 0.10 0.13 0.12 0.15 0.25N 192 192 212 212 212

In this table, we analyze program completion rates. The dependent variable is the percentage ofshares repurchased during the repurchase program. Undervaluation is an indicator variable equalto 1 if the repurchase reason is undervaluation. Zero repurchaser is an indicator variable equal to1 if the firm does not repurchase any shares after the initial program announcement. Repeat is anindicator variable equal to 1 if the firm announced another repurchase program within the prior15 months. Size quartile is the relative size of the firm compared with all ASX firms on theprogram announcement date (Quartile 1 is for small firms and 4 is for large firms). Firm size iscalculated as the market value of equity. Market-to-book is the relative market-to-book of the firmcompared with all ASX firms on the program announcement date (Quartile 1 is for smallmarket-to-book firms and 4 is for large market-to-book firms). Market-to-book is calculated as themarket value of equity divided by book value of equity. Program size is ratio shares repurchased tothe number of shares outstanding. Excess return is the cumulative stock return net of the marketreturn during the year before the program is finalized. N is the number of observations. Allmodels include year and industry dummies. p Values are reported in parentheses. Standard errorsare clustered at the firm level. a, b, and c denote statistical significance at the 1%, 5%, and 10%levels.



different control variables and sample sizes. In the first two models, we excludezero repurchasers. None of the control variables in the first model is signifi-cantly related to completion rates. We add the undervaluation dummy in thesecond model and find that it is significantly negatively related to completionrates. Thus, after controlling for other factors, undervaluation firms not onlyrepurchase fewer shares in a given month (Table 9), but they also buy fewer totalshares during the program. In the next three models, we add zero repurchasersto the sample. In Model (3), market-to-book is significantly negative, implyingthat firms that are not associated with potential undervaluation have a lowercompletion rate. This is not surprising as we know from earlier tables that zerorepurchasers have higher MBs. We next add the undervaluation dummy to themodel and find that it is still significantly negative and predicts a lower comple-tion rate. In the last model, we add the zero repurchaser dummy and asexpected the market-to-book variable loses its significance. The zero repurchaserdummy is significantly negative as predicted and the undervaluation dummy isstill significantly negative. Overall, the results indicate that undervaluationfirms tend to buy back a smaller fraction of their intended number of shares atthe end of their repurchase program.

C. Long-term performance of positive and zero repurchasers

If managers are actually taking advantage of mispricing, then examining long-term returns may help us to find out if undervaluation has been corrected.Following Barber and Lyon (1997), we calculate 1-, 2- and 3-year buy-and-holdreturns (BHRs) to examine the stock price performances of repurchasing firmsagainst a control sample of firms. To form the control sample, we find amatching firm based on firm size and market-to-book. We first require that thematching firm has a market value between 70% and 130% of the repurchasingfirm and then choose the firm with the closest MB to that of the repurchasingfirm as the control firm. Firms with missing data needed to calculate BHRs andsample firms that are delisted are eliminated from the analysis, leaving us 761,629, and 495 announcements for the full sample to calculate 1-year, 2-year, and3-year BHRs, respectively.23

We present the long-term returns for positive repurchasers in Table 11. PanelA reports the returns for the full sample. Although repurchasing firms outper-form their control sample firms in all holding period horizons, the differencesare not significant. For example, in the 3-year window, the BHR of repurchasingfirms on average is 55.18%, which is more than 48.91% for the control samplefirms. However, the difference is not statistically significant.

Panels B and C report the average BHRs for the undervaluation and othermotives subsamples. Undervaluation firms do better than their matching firmsin the first year following the repurchases. The average BHR is 38.90% compared

23 We also lose observations as those firms with repurchases in 2007 and 2008 do not have datato compute 2- and 3-year long-term returns.



with only 9.54% for the control sample. The difference is 29.36% and statisti-cally significant at the 5% level. The performance of the undervaluation samplefirms, however, in the second year dies down and the control sample firms doslightly better in the 2-year period. However, the difference between the averagereturns is only -2.45%, which is not statistically significant. There is no statis-tical difference between the returns in the 3-year horizon between undervalu-ation firms and their control firms.

Other motives firms in Panel C do not do better or worse than their controlsample firms. For example, in the 1-year column, the average buy-and-holdreturn is 16.71% compared with 17.47% for the control sample firms. Althoughother motives sample firms do better in the 2-year period, the differencebetween the returns is not statistically significant. Other motives firms doslightly better in the 3-year horizon; however, the difference of 7.47% is notstatistically significant.

Our results in Table 11 suggest that repurchasing firms on average do betterthan their matching control sample firms. The better performance mainly inthe first year comes from the firms that cite the undervaluation motive as theirrepurchase reason (consistent with our third hypothesis). Thus, undervaluationmotive firm managers are on average able to detect valuation errors and take

Table 11 Long-term buy-and-hold returns for positive repurchasers

Sample 1-Year 2-Year 3-Year

N Return N Return N Return

Panel A: Full sampleRepurchase firms 761 19.13% 629 41.74% 485 55.18%Control firms 16.61% 34.28% 48.91%Difference 2.53% 7.46% 6.27%

(0.44) (0.15) (0.38)Panel B: Undervaluation sampleRepurchase firms 103 38.90% 72 25.86% 49 -0.41%Control firms 9.54% 28.30% 3.98%Difference 29.36%a -2.45% -4.39%

(0.04) (0.86) (0.73)Panel C: Other motives sampleRepurchase firms 658 16.71% 557 43.79% 436 61.42%Control firms 17.47% 35.05% 53.95%Difference -0.76% 8.74% 7.47%

(0.81) (0.12) (0.34)

This table reports buy-and-hold returns (BHRs) following actual share repurchases up to 3 years.We BHRs relative to a control sample of matched firms for the full sample and the two sub-samples. Firms are matched on first firm size then on market-to-book and the methodology issimilar to Zhang (2005). N is the number of actual repurchases with available data to calculateabnormal returns. p Values are reported in parentheses. a denotes statistical significance at the5% level.



advantage of such errors as evidenced by short- (Table 7) and long-term stockprice performances (Table 11).

A natural question that comes to mind is if zero repurchasers do better orworse than positive repurchasers in the long run. Our earlier results indicatedthat the market reaction to program announcements by zero repurchasers waspositive. Examining zero repurchasers can also help to determine if the initialsignal from the company (repurchase program announcement) is sufficient forunderperformance to be corrected or, as suggested by the traditional signalingtheories, if a costly action (actual repurchases) is needed for the underperfor-mance to be corrected.

We present the results for zero repurchasers in Table 12. There are only 20announcements by zero repurchasers. We measure buy-and-hold returns (BHRs)relative to the repurchase program announcements. The full sample results inPanel A are mixed. In the 1-year horizon, the BHR for zero repurchasers is notmuch different from that for control sample firms. In the 2-year period, thedifference is an insignificant -18.82%. In the 3-year period, the pattern reversesand the difference of BHRs is 3.38%, again not significant. We next divide zerorepurchasers as undervaluation and other motives and report the BHRs in thelast two panels. Undervaluation firms do better than their control sample firmsin the first year as the difference is 21.54%. Other motives firms on the other

Table 12 Long-term buy-and-hold returns for nonrepurchasers

Sample 1-Year 2-Year 3-Year

N Return N Return N Return

Panel A: Full sampleRepurchase firms 20 19.08% 18 1.90% 17 11.94%Control firms 21.02% 20.72% 8.56%Difference -1.94% -18.82% 3.38%

(0.88) (0.28) (0.85)Panel B: Undervaluation sample

Repurchase firms 6 37.66% 4 14.21% 4 -20.70%Control firms 16.11% -10.95% -12.83%Difference 21.54% 25.16% -7.88%

(0.20) (0.16) (0.90)Panel C: Other motives sample

Repurchase firms 14 11.12% 14 -1.62% 13 21.98%Control firms 23.12% 34.29% 18.43%Difference -12.00% -35.91% 3.55%

(0.48) (0.10) (0.89)

This table reports buy-and-hold returns (BHRs) for repurchase program announcements that arenot followed by actual repurchases. We calculate BHRs relative to a control sample of matchedfirms. Firms are matched on first firm size then on market-to-book on the buyback programannouncement date. N is the number of repurchase program announcements with available datato calculate abnormal returns. p Values are reported in parentheses.



hand do worse (the difference is -12.00%). However, because of the smallsample size, none of the differences is statistically significant.

Compared with Table 11, the results in Table 12 show that there is no differ-ence between positive and zero repurchasers that mention undervaluation intheir announcements in the 1-year horizon. This suggests that, unlike the claimof the traditional signaling theories and contrary to our last hypothesis, a costlyaction (i.e., actual repurchase) may not be needed by the market for the under-performance to be corrected. However, due to a very small sample size for zerorepurchasers (six announcements for the first BHR return for the undervalua-tion subsample), we do not emphasize this finding.

VII. CONCLUSION

Before conducting share repurchases, Australian companies are required toprovide at least one reason as to why the repurchase program is being initi-ated. Among the reasons provided, the undervaluation of the firm’s stockprice particularly stands out. Using actual share repurchases between 1998and 2008 in Australia, we examine whether the market’s reaction to actualrepurchases is different for the undervaluation motive compared with otherrepurchase motives. We show that firms tend to repurchase shares after aperiod of decline in share price and that firms buy back on average on dayswhen their stock prices are declining, suggesting that managers are timing themarket. In addition, both initial repurchase program and actual repurchaseannouncements are greeted positively by the market. We, however, find evi-dence that the market reaction is more positive for the announcements ofrepurchases (both initial program and subsequent actual repurchaseannouncements) made by firms where the stated motive for the buyback isundervaluation of the company’s stock price. We further report that under-valuation motives firms repurchase fewer shares at the end and have a lowerprogram completion rate.

We examine long-term stock price performances of repurchasing and non-repurchasing firms separately for the undervaluation and other motives sub-samples and find that firms that repurchase shares because of undervaluationperform better than a control sample of firms matched on size and market-to-book during the 1-year period following the repurchases. On the other hand,other motives firms do not perform better or worse than their control samplefirms. Our results for nonrepurchasers suggest that nonrepurchasing undervalu-ation firms also do better in the long-run than their control sample firms, whichmay suggest that a costly action may not be necessary as suggested by thetraditional signaling theories. However, due to lack of statistical significancebecause of the small sample size for nonrepurchasing firms, we do not put muchemphasis on this finding.

Our findings indicate that the signal send by firms with the undervaluationmotive is stronger than the one sent by other motives firms and the repurchase



reason is an important variable for the market. Overall, our results suggest thatmanagement statements have value.

Ali AkyolUniversity of MelbourneParkvilleVictoria [email protected]

REFERENCES

Asquith, P., and D. W. Mullins (1986), ‘Signaling with Dividends, Stock Repurchases,and Equity Issues’, Financial Management, 15, 27–44.

Banyi, M., E. Dyl, and K. Kahle (2008), ‘Errors in Estimating Share Repurchases’, Journalof Corporate Finance, 14, 460–74.

Barber, B. M., and J. D. Lyon (1997), ‘Detecting Long-Run Abnormal Stock Returns:The Empirical Power and Specification of Test Statistics’, Journal of Financial Eco-nomics, 43, 341–72.

Ben-Rephael, A., J. Oded, and A. Wohl (2011), ‘Do Firms Buy Their Stock at BargainPrices? Evidence from Actual Stock Repurchase Disclosures’, Available at SSRN:http://ssrn.com/abstract=1738922

Bhattacharya, S. (1979), ‘Imperfect Information, Dividend Policy, and ‘the Bird in theHand’ Fallacy’, Bell Journal of Economics, 10, 259–70.

Billett, M. T., and H. Xue (2007), ‘The Takeover Deterrent Effect of Open Market ShareRepurchases’, Journal of Finance, 62, 1827–50.

Bonaimé, A. A. (2010), ‘Mandatory Disclosure and Firm Behaviour: Evidence fromShare Repurchases’, Available at SSRN: http://ssrn.com/abstract=1585459

Brailsford, T., D. Marchesi, A. Simon, and I. Tutticci (2008), ‘The Determinants of ShareRepurchase Decisions in on-Market Buy-Backs’, The University of QueenslandWorking Paper.

Brav, A., J. Graham, C. Harvey, and R. Michaely (2005), ‘Payout Policy in the 21stCentury’, Journal of Financial Economics, 77, 483–527.

Brockman, P., and D. Y. Chung (2001), ‘Managerial Timing and Corporate Liquidity:Evidence from Actual Share Repurchases’, Journal of Financial Economics, 61, 417–48.

Brown, C. (2007), ‘The Announcement Effects of off-Market Share Repurchases inAustralia’, Australian Journal of Management, 32, 369–85.

Brown, S. J., and J. B. Warner (1985), ‘Using Daily Stock Returns – The Case of EventStudies’, Journal of Financial Economics, 14, 3–31.

Campbell, J. Y., A. W. Lo, and A. C. MacKinlay (1997), Chapter 4: Event-Study Analysis,The Econometrics of Financial Markets. Princeton, NJ: Princeton University Press.

Chan, K., D. L. Ikenberry, and I. Lee (2007), ‘Do Managers Time the Market? Evidencefrom Open-Market Share Repurchases’, Journal of Banking & Finance, 31, 2673–94.

Comment, R., and G. A. Jarrell (1991), ‘The Relative Signaling Power of Dutch-Auctionand Fixed-Price Self-Tender Offers and Open-Market Share Repurchases’, Journal ofFinance, 46, 1243–71.

Cook, D. O., L. Krigman, and J. C. Leach (2004), ‘On the Timing and Execution ofOpen Market Repurchases’, Review of Financial Studies, 17, 463–98.

Dann, L. Y. (1981), ‘Common-Stock Repurchases – An Analysis of Returns to Bond-holders and Stockholders’, Journal of Financial Economics, 9, 113–38.



Dharmawan, G., and J. Mitchell (2001), ‘Australian Buy-Back Regulations – A Cross-Country Comparison’, Australian Journal of Corporate Law, 12, 246–81.

Dittmar, A. (2000), ‘Why Do Firms Repurchase Stock?’, Journal of Business, 73,331–55.

Douthit, J. D., L. W. Kearney, and D. E. Stevens (2011), ‘Can Cheap Talk Be As Effectiveat Mitigating Agency Problems as Reputation Building? An Experimental Test in thePresence of a Noisy Performance Measure. Florida State University Working Paper’,Available at SSRN: http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1868986

Graham, J. R. ‘Quarter 2’, 1999. FEI Survey. http://faculty.fuqua.duke.edu/~jgraham/Grullon, G., and D. Ikenberry (2000), ‘What Do We Know about Stock Repurchases?’,

Journal of Applied Corporate Finance, 13, 31–51.Grullon, G., and R. Michaely (2004), ‘The Information Content of Share Repurchase

Programs’, Journal of Finance, 59, 651–80.Ikenberry, D., J. Lakonishok, and T. Vermaelen (1995), ‘Market Underreaction to Open

Market Share Repurchases’, Journal of Finance, 50, 982–3.Ikenberry, D., J. Lakonishok, and T. Vermaelen (2000), ‘Stock Market Repurchases in

Canada: Performance and Strategic Trading’, Journal of Finance, 55, 2372–97.Ikenberry, D., and T. Vermaelen (1996), ‘The Option to Repurchase Stock’, Financial

Management, 25, 9–24.Konan, C., D. Ikenberry, and L. Inmoo (2004), ‘Economic Sources of Gain in Stock

Repurchases’, Journal of Financial & Quantitative Analysis, 39, 461–79.Lamba, A., and I. Ramsay (2005), ‘Comparing Share Buybacks in Highly Regulated

and Less Regulated Market Environments’, Australian Journal of Corporate Law, 17,261–80.

Lie, E. (2005), ‘Operating Performance Following Open Market Share RepurchaseAnnouncements’, Journal of Accounting & Economics, 39, 411–36.

Loughran, T., and J. R. Ritter (1995), ‘The New Issues Puzzle’, Journal of Finance, 50,23–51.

McNally, W. J., and B. F. Smith (2007), ‘Long-Run Returns Following Open MarketShare Repurchases’, Journal of Banking & Finance, 31, 703–17.

McNally, W. J., B. F. Smith, and T. Barnes (2006), ‘The Price Impacts of Open MarketRepurchase Trades’, Journal of Business Finance & Accounting, 33, 735–52.

Miller, M., and K. Rock (1985), ‘Dividend Policy under Asymmetric Information’,Journal of Finance, 40, 1031–51.

Mitchell, J., and P. Robinson (1999), ‘Motivations of Australian Listed CompaniesEffecting Share Buy-Backs’, Abacus, 35, 91–119.

Mitchell, J., and G. Dharmawan (2007), ‘Incentives for on-Market Buy-Backs: Evidencefrom A Transparent Buy-Back Regime’, Journal of Corporate Finance, 13, 146–69.

Mitchell, J., G. Dharmawan, and A. Clarke (2001), ‘Managements’ Views on ShareBuy-Backs: An Australian Survey’, Accounting and Finance, 41, 93–129.

Mitchell, J., H. Izan, and R. Lim (2006), ‘Australian on-Market Buy-Backs: An Exami-nation of Valuation Issues’, Multinational Finance Journal, 10, 43–79.

Oded, J. (2005), ‘Why Do Firms Announce Open-Market Repurchase Programs?’,Review of Financial Studies, 18, 271–300.

Otchere, I., and M. Ross (2002), ‘Do Share Buy Back Announcements Convey Firm-Specific or Industry-Wide Information?: A Test of the Undervaluation Hypothesis’,International Review of Financial Analysis, 11, 511–31.

Peyer, U., and T. Vermaelen (2009), ‘The Nature and Persistence of Buyback Anoma-lies’, Review of Financial Studies, 22, 1693–745.

Purnanandam, A. K., and B. Swaminathan (2004), ‘Are IPOs Really Underpriced?’,Review of Financial Studies, 17, 811–48.

Rees, W. (1996), ‘The Impact of Open Market Equity Repurchases on UK Equity Prices’,European Journal of Finance, 2, 353–70.



Stephens, C. P., and M. S. Weisbach (1998), ‘Actual Share Reacquisitions in Open-Market Repurchase Programs’, Journal of Finance, 53, 313–33.

Vermaelen, T. (1981), ‘Common Stock Repurchases and Market Signalling: An Empiri-cal Study’, Journal of Financial Economics, 9, 139–83.

Vermaelen, T. (1984), ‘Repurchase Tender Offers, Signalling and Managerial Incen-tives’, Journal of Financial and Quantitative Analysis, 19, 163–81.

Wang, C.-S., N. C. Strong, S. Tung, and S. W. Lin (2009), ‘Share Repurchases, theClustering Problem, and the Free Cash Flow Hypothesis’, Financial Management, 38,487–505. Forthcoming.

Zhang, H. (2005), ‘Share Price Performance Following Actual Share Repurchases’,Journal of Banking & Finance, 29, 1887–901.

APPENDIX A – EXAMPLE OF A BUYBACK ANNOUNCEMENT

AUTRON CORPORATION LIMITED 2001-09-25 ASX-SIGNAL-G HOMEX –Melbourne

APPENDIX 3C

ANNOUNCEMENT OF BUY-BACK

(EXCEPT MINIMUM HOLDING BUY-BACK)

Name of Entity: Autron Corporation Limited ACN or ARBN: 002,876,182We (the entity) give ASX the following information.

INFORMATION ABOUT BUY-BACK1. Type of buy-back On market2. Class of shares which is subject of the

buy back (eg, ordinary/preference)Ordinary

3. Voting rights (eg, one for one) One for one4. Fully paid/partly paid (and if partly

paid, details of how much has been paidand how much is outstanding)

Fully paid

5. Number of shares in the class on issue 519,146,4536. Whether shareholder approval is

required for buy-backNo

7. Reason for buy-back Recent adverse movements in the Com-pany’s share prices are regarded by theCompany as not reflecting truly theunderlying value of the shares particularlyhaving regard to the recently announcedresults and dividend declaration on 13 &24 September 2001 respectively. Therefore,the Company believes that it is in theinterest of the body of shareholders at thistime to commence the process of anon-market buy-back of ordinary shares.

8. Any other information material to ashareholder’s decision whether to acceptthe offer (eg, details of any proposedtake-over offer or announcement)

See announcements of 13 & 24 September2001



ON-MARKET BUY-BACK9. Name of broker who will act on the

company’s behalfTo be advised

10. Name of each director and relatedparty of a director who reserves theright to sell shares, and number ofshares in respect of which that directoror related party reserves the right

Messre Y C Leong and E Loh have sharessubject to margin facilities and accordinglythe relevant financial instructions canmake margin calls. Those Directors mayincur forced selling of their shares andtherefore must reserve their right to sellshares up to maximum of five millionshares each.

11. If the company intends to buy back amaximum number of shares – thatnumberNote: This requires a figure to beincluded, not a percentage.

Up to 25 million shares

12. If the company intends to buy backshares within a period of time – thatperiod of time; if the company intendsthat the buy-back be of unlimitedduration – that intention

Unlimited duration

13. If the company intends to buy backshares if conditions are met – thoseconditions

N/A

EMPLOYEE SHARE SCHEME BUY-BACK: Items 14 to 15 are Not ApplicableSELECTIVE BUY-BACK: Items 16 to 18 are Not ApplicableEQUAL ACCESS Scheme: Items 19 to 22 are Not ApplicableCOMPLIANCE STATEMENT

1. The company is in compliance with all Corporations Law require-ments relevant to this buy-back.

2. There is no information that the listing rules require to be disclosedthat has not already been disclosed, or is not contained in, or attachedto, this form.

M GarbuttCOMPANY SECRETARY25/09/2001



Documents

Share Repurchase Reasons and the Market Reaction to Actual Share Repurchases: Evidence from Australia