Signaling Power of Open Market Share Repurchases in Germany

Fin Mkts Portfolio Mgmt (2006) 20:123–151DOI 10.1007/s11408-006-0011-9

Signaling power of open market share repurchasesin Germany

Andreas Hackethal · Alexandre Zdantchouk

Published online: 24 May 2006© Swiss Society for Financial Market Research 2006

Abstract This paper shows that abnormal stock price returns around the dateof open market repurchase announcements are four times higher in Germanythan in the USA (12 ver. 3%). We hypothesize that this observation can beexplained by national differences in repurchase regulations. Our empirical evi-dence indicates that German managers primarily buy back shares to signal anundervaluation of their firm. We demonstrate that the stringent repurchase pro-cess prescribed by German law attributes a higher credibility to undervaluationsignals than do the lax US regulations, and thereby corroborates our hypothesis.

Keywords Open market share repurchases · Event study ·Undervaluation signaling · Repurchase policy and regulation

JEL Classification Numbers G14 · G35

1 Introduction

In May 1998 the “Corporation Control and Transparency Act” (KonTraG)abolished the legislation restricting share repurchasing in Germany. In the

Financial support from the E-Finance Lab, Frankfurt am Main, and from Freitag & Co., Frankfurtam Main, is gratefully acknowledged. We would like to thank two anonymous referees for theirhelpful comments.

A. Hackethal (B)European Business School,65375 Oestrich-Winkel,Germanye-mail: [email protected]

A. ZdantchoukFreitag & Co.,Frankfurt am Main, Germany

124 A. Hackethal, A. Zdantchouk

subsequent 5 years, more than 180 German firms took advantage of this newfreedom to make over 240 share repurchase announcements.

The new German laws on share repurchase programs differ in importantaspects from the laws and regulations that govern share repurchase transac-tions in the USA. For example, managers of German firms must first obtainapproval for a share repurchase program at their annual shareholder meetingand then publicly announce the share repurchase. Management must then givea full statement at the next shareholder meeting on the reasons for the trans-action, its volume, and the price paid per share. In contrast, US firms needonly obtain approval from their company board and must publicly announcethe establishment of a repurchase program. They do not have to disclose thedetails of any actual repurchase transactions either to the authorities or totheir shareholders. The existing empirical literature on US open market sharerepurchases cited below provides strong evidence that many US firms announcerepurchase programs to convey their assessment to investors that their sharesare undervalued. A change in disclosure requirements would perhaps affect thequality of such a signal and thus have an effect on the motivation of firms toengage in share repurchases. An empirical analysis of share repurchases in Ger-many therefore promises to provide new insights into the role of specific legalprovisions in instigating corporate insiders to successfully disseminate privateinformation to capital markets.

This paper proceeds from this general conjecture and pursues two goals.In order to find out whether German share repurchase announcements have amarket impact different from the repurchase announcements in other countries,we measure the abnormal share price effects around the two major publicly ob-servable events in the context of share repurchases in Germany, namely a firm’sinitial voluntary statement to seek shareholder approval and the subsequentobligatory announcement to begin share repurchasing. We then test four pop-ular hypotheses on why managers repurchase shares and thereby investigatewhether motivations of German firms differ systematically from the motiva-tions of their US counterparts. Our main empirical finding is that although bothGerman and US firms seem primarily to repurchase shares in order to signalthat their firm is undervalued, announcement price effects tend to be consider-ably greater in Germany than in the USA. We discuss three possible explana-tions for this observation and conclude that German repurchase regulations arestricter than US regulations and therefore provide a more powerful signalingtool.

The paper unfolds as follows: the next section describes the German legalframework governing repurchase programs and compares this framework toUS regulations. Section 3 briefly reviews the empirical literature on repurchasetransactions in various countries. Section 4 presents the dataset, and Sect. 5describes the methodology. Section 6 reports and discusses the empirical results.The last section summarizes the paper and concludes with a side note to regula-tors. Throughout the paper we concentrate on open market programs becausethey have been by far the most popular vehicle in Germany.

Signaling power of open market share repurchases in Germany 125

2 Repurchase motivations

The management’s motivation to repurchase shares has already been discussedextensively in the literature.1 The following provides only a brief review. To thisextent we have grouped the most widely discussed motivations into two cate-gories. The five motivations from the first category are generally considered tobe in line with the shareholders’ interests. The second category contains threemotivations which tend to contravene the shareholders’ interests.

Motivations broadly in line with shareholder interests include the attempt bythe management to signal to investors that the true value of their corporation’sequity exceeds its current market value. Such a signal might be based on themanagement’s assessment that the true mean of the probability distribution ofthe firm’s future cash flows is actually higher than perceived by the market.Alternatively, the management might agree with the market on the mean of thedistribution but might believe that the variance of future cash flows around themean value is higher than expected by markets (Dann 1981). In the first case, allof the firm’s risky securities appear to be undervalued. In the latter case, onlyequity claims appear to be undervalued, whereas claims in the form of risky debtmight in fact be overvalued. Share repurchases might then lead to a reduction ofdebtholders’ wealth to the benefit of shareholders. It is typically assumed that afirm’s management is better informed about the firm’s true current competitiveposition and its future value creation prospects than its outside investors. How-ever, a straight public management announcement that it considers its firm’sshares to be undervalued generally lacks credibility. Since outside investors can-not distinguish between the true and deliberately misleading announcements,they will regard all undervaluation announcements as cheap talk unless thecost of producing false announcements is sufficiently high. Share repurchaseannouncements incur two types of costs for overvalued firms. Firstly, firms thatrepurchase overvalued shares must assume that the share price will soon declineto its true intrinsic value so that they incur a loss from such a transaction. Sec-ondly, firms which announce a share repurchase but subsequently decide notto do so must expect a damaging effect on their reputation for honest marketcommunication. Depending on the legal and regulatory provisions for sharerepurchase programs, such behavior can also attract the interest of the author-ities to investigate potential price manipulations. Given these potential costsas a deterrent for false announcements, share repurchase announcements canserve as a trustworthy device to enhance the credibility of an undervaluationsignal. Such credible signals should then lead to an appreciation in stock price.

Also in line with shareholders’ interests are share repurchases through whicha firm aims to distribute excess cash to its shareholders. Excess cash gives rise toagency conflicts as self-interested managers could use these funds for negativenet present value investments such as fringe benefit consumption or empirebuilding and thereby harm the firm owners (Jensen 1986). Agency difficulties

1 See for example Comment and Jarrell (1991), Stephens and Weisbach (1998) and Grullon andIkenberry (2000).


might also arise in situations where a firm’s management does not engagein outright value destruction but merely runs out of value-adding investmentopportunities and retains excess cash on the balance sheet. Shareholders arenot able to reallocate their capital to more productive uses in the economy.Both types of agency conflicts are ameliorated if excess cash is used to repur-chase the firm’s own shares. As a consequence, share price effects of such sharerepurchases should be ceteris paribus stronger for firms where manager andshareholder interests are less aligned. Alignment is typically attained throughincentive-based manager compensation contracts or through the concentrationof control rights in the hands of large blockholders, who have stronger incen-tives to monitor and discipline the management than dispersed owners (Shleiferand Vishny 1986).

A third motivation concerns the optimization of a firm’s capital structure.However, in many countries such as Germany, France, Italy, and Hong Kong,the volume of shares to be repurchased must not exceed 10% of total sharesoutstanding. Firms from such countries are therefore not able to implementrepurchase programs to substantially increase their debt-to-capital ratio andtransfer value from debt holders to shareholders. As Grullon and Ikenberry(2000) point out, firms might nevertheless use repurchase programs to fine-tune their capital structure in response to potential dilution effects from theiremployee and executive stock options incentive plans. If the subsequent changein the capital structure is only marginal, and if – as is typically the case – theannouncement does not provide new information on the structure of the com-pensation plan, price effects should be quite moderate.

In countries where the tax burden on an investor’s income is typically higherfor dividend income than for capital gains, shareholder-oriented firms have afourth motivation to repurchase shares. Such firms should substitute dividendpayouts with share repurchases in order to maximize after-tax proceeds fortheir shareholders. Although shareholders will prefer repurchases to dividendsin this scenario, price effects from individual tax-induced repurchase announce-ments need not necessarily be strong. If the management of a given firm hasa reputation for shareholder-oriented acting, the shareholders of that firm willexpect the management to return cash in the form of repurchases and not viadividends anyway. Repurchase announcements will then convey no extra tax-efficiency signal. Rather these firms should experience tax-induced abnormalshare price returns in the instance where the applicable tax regime is amendedin favor of capital gains. Following such a change in the tax regime, one shouldalso observe an increased ratio of repurchasing firms to dividend paying firms.

The last motivation from this category is also tax-induced and applies to firmswho plan to acquire another firm. In many jurisdictions, an exchange of shares ismore tax-efficient for the target firm’s shareholders than the receipt of cash. Theacquiring firm can therefore repurchase its own shares to obtain a tax-efficientcurrency for M&A transactions. Tax advantages for the target firm may thentranslate into a lower acquisition price and thereby benefit the shareholdersof the acquiring firm. If the characteristics of the acquisition deal are known


ex ante, the repurchase announcement will, however, not create new marketinformation and should therefore leave the acquirer’s market value unchanged.

Motivations that tend to contravene shareholders’ interests include manage-ment’s efforts to fend off a value-creating takeover attempt. Target managementmust fear that it will be ousted as a consequence of such a takeover and that itwill typically not be compensated for any losses in future income and privatebenefits. Target management might then use share repurchases as a defenseinstrument to reduce the amount of outstanding shares available to the raid-ing firm from dispersed outside shareholders as well as to deplete any cashreserves obtainable through the acquisition. Bagwell (1991) provides a moresubtle argument as to how share repurchases can deter a takeover. She arguesthat shareholders who tender in a repurchase typically have lower valuationsfor the firm than the average shareholder. This heterogeneity in valuationsmight leave the raider with a more “expensive” pool of shareholders after therepurchase, possibly deterring him from making a takeover offer. Given theabove-mentioned restrictions on the volume of shares to be repurchased inGermany, the chances of fending off a takeover by the sole means of sharerepurchases are probably quite low. However, share repurchase can be just oneof several defensive measures taken by incumbent target management and assuch might indeed contribute to shareholder value destruction.

A second situation in which share repurchases can violate the interests ofat least one group of shareholders arises when only inside shareholders areinformed of the exact timing of repurchase transactions. They can then use thisknowledge to dispose of their shares at a higher price than under normal marketconditions (Ikenberry and Vermaelen 1996). This would cause a wealth transferfrom outside shareholders to inside shareholders and, if anticipated by outsideshareholders, lead to a negative announcement effect.

Finally, managers who hold a substantial equity stake in their firm mightlaunch a repurchase program in an attempt to dilute the control rights of othershareholder groups. Even if these other shareholders are not willing to sell theirshares to the firm, the transaction costs associated with such an attempt couldhave a negative value impact.

If the market believes that a firm is driven by one or several motivationsfrom the first (second) category, one should observe positive (negative) priceeffects around the repurchase announcement date. Contrarily, if the markethas no prior awareness of the motivations of a given firm, its share price shouldnot react at all in response to an announcement. Repurchase firms are thenbelieved to pay out cash to shareholders that would have yielded a return oncapital equal to the cost of capital if kept inside the firm. Thus, unless the firmuses underproductive assets to finance the payout (equivalent to the excesscash distribution hypothesis above), total future earnings will decline by thesame proportion as the number of shares outstanding, and earnings per share(EPS) will remain unchanged. Repurchase firms that solely attempt to boostvaluations based on increased EPS should not be successful if the market isaware of such an attempt. The “EPS bump” argument, which is often advocated


by investment banks, analysts, and the firms themselves, can therefore be ruledout by simple logic (see Grullon and Ikenberry 2000).

In this paper we confine our testing to four out of the eight hypothesesregarding potential motivations of German firms to repurchase shares: the“undervaluation signal”, the “excess cash”, the “takeover defense”, and the“tax efficiency” hypotheses. As will be shown in the next section, the “rentseeking by insiders” motivation is largely ruled out by German law, so thatthe associated hypothesis can be falsified immediately. There are two reasonswhy we refrain from testing the remaining three hypotheses in this paper. Thefirst is methodological. Our empirical approach is based on regressing abnormalreturns from share buy-back announcements on firm characteristics. As we haveargued above, buy-back announcements that are driven either by the “capitalstructure optimization”, the “acquisition currency”, or the “control dilution”motivation would most likely result in very small share price reactions. Wetherefore have reason to believe that even covariates that ideally captured anyof these three motivations would carry statistically insignificant coefficients inour regression model. The second reason concerns data availability. Informa-tion on optimal capital structure, acquisition plans, and on the balance of poweramong shareholder and stakeholder groups is generally not available.

3 Share repurchase regulation in Germany and in the USA

The German Corporation Control and Transparency Act (KonTraG) becameeffective on May 1, 1998, permitting German firms to repurchase commonand preferred shares under the following conditions and subject to the fol-lowing requirements: a share repurchase program must first be authorized bythe firm’s annual general meeting of shareholders (AGM). The AGM decideson the maximum amount of shares to be repurchased. The amount must notexceed 10% of total shares outstanding2, and the repurchase must be made outof distributable profits. The AGM also determines the time frame during whichtransactions can take place. This period must not exceed 18 months. In addition,shareholders decide on the precise buy-back procedure unless the firm plansto repurchase shares over the open market.3 AGM approval is immediately tobe reported to Germany’s financial services authority (BaFin). AGM approvaldoes not oblige a firm to actually repurchase shares. In section 4 we show thatmore than two thirds of German firms did not act upon AGM approval. When

2 A 5% threshold applies to financial institutions trading for their own purposes [section 71 (1) Nr.7 AktG]. The law does not specify whether the thresholds apply to the total stock of repurchasedshares held in treasury or solely to one 18-month period. In the latter case, firms could in principlebuy back a substantially higher portion of own shares by obtaining AGM approval in subsequentyears. For a discussion of this ambiguity see Kraft and Altvater (1998) and Bosse (2000).3 Existing types of non-open-market buy-backs include fixed-price tender offers, where the corpo-ration offers to buy a specified amount of shares at a fixed price – typically exceeding current marketprices – during a pre-specified period; Dutch-auction tender-offers, which are similar to fixed-pricetender offers, except that prices are set in a book-building procedure and targeted buy-backs, wherethe corporation negotiates with a particular shareholder over the purchase of a block of shares.


a firm wishes to repurchase shares over the open market, it must make a pub-lic announcement in advance. This announcement does not have to state thevolume of shares to be repurchased and is not binding for the firm, i.e., thefirm can still choose not to transact. Should it repurchase shares, transactionsmust be conducted in such a way that all the firm’s shareholders are treatedequally. Moreover, transactions must not be used for the purpose of trading. Inthe AGM following a transaction, the firm has to issue a full statement as to thereason for the share repurchase, how many shares were involved, and at whatprice they were repurchased.

In their totality, these legal provisions strongly mitigate the motivations thatwe discussed in our introduction as not being commensurate with the interestsof (outside) shareholders. Firstly, the requirement for equal treatment of allshareholders in combination with the obligation to obtain an explicit AGMauthorization for tender offers or targeted buy-backs strongly impedes wealthtransfers from one shareholder group to another. In fact, a mere four out ofthe 237 buy-back announcements in our initial sample were not open marketprograms. Secondly, because the amount of equity to be repurchased is cappedat 10% of total nominal capital, the opportunities for substantially changing afirm’s capital structure or deterring a takeover are fairly limited. For example,an inside blockholder must ex ante hold more than 45% of total share capitalto achieve a majority stake in excess of 50% of share capital after a full 10%repurchase transaction.4 Moreover, the fact that repurchased shares held intreasury are not entitled to voting rights impedes control right dilution.

German regulations can also impose considerable costs on firms who make amisleading repurchase announcement. Firstly, the costs of establishing a repur-chase program are not trivial because AGM approval is required. Secondly,firms who publicly announce an imminent repurchase transaction, but then donot actually repurchase any or only a miniscule amount of shares, must fearthat regulators suspect price manipulation and initiate investigations. Share-holders of these firms will most likely also demand inquiries, possibly callinginto question the managers’ reputation for truthful disclosures. Repurchasingand immediately reselling shares is also not a viable option for this type of firmsince such behavior would certainly be viewed as trading in own shares, whichis explicitly prohibited by law. Generally, higher expected costs of misleadingrepurchase announcements will reduce the occurrence of such announcements.In turn, such costs increase the strength of the signal that honest managers cancommunicate to the market by means of a repurchase announcement. Arguablythen, the strictness of German repurchase regulations allows for fairly strongsignals.

4 In fact, Gerke et al. (2003) report that German firms have on average only bought back 3.2% oftheir outstanding shares. This figure is actually not too far off the corresponding US figure. Grullonand Michaely (2002) report that the median US firm sought to repurchase 5% of its outstandingshares and Stephens and Weisbach (1998) find that US firms completed roughly 75% of theirauthorized shares during a 3-year period after the announcement.


US regulations on open market share repurchasing differ in several impor-tant aspects from German laws.5 In the USA, share repurchase programs onlyrequire approval from the board of directors rather than the AGM. There isneither a restriction on the duration of the program nor on the number ofshares to be repurchased. Stephens and Weisbach (1998) report that it is indeednot uncommon that US open market programs are spread out over severalyears. Apart from making an initial public announcement, firms are not obligedto formally disclose any details of their repurchase programs or their individ-ual repurchase transactions. As a consequence, both the costs for establishinga repurchase program and the penalties for not buying back shares after anannouncement are lower in the US than in Germany. Moreover, US firms havemore flexibility in terms of transaction volume and timing. This leaves moreopportunities for corporate insiders to extract rents from corporate outsiders.

Two broad empirical implications follow from these discrepancies betweenthe German and US repurchase regulation. We would expect a smaller portionof negative abnormal announcement returns in Germany than in the US and, ifboth German and US firms are mainly motivated by signaling undervaluation,we would also expect that positive returns are on average higher in Germanythan in the USA.

4 Related empirical literature

The empirical literature on stock buy-backs has so far largely focused on USmarkets. A number of clear-cut results have emerged. Share repurchases lead tosignificant positive abnormal returns on average; however, stock price reactionsto tender offers tend to be at least twice as strong as reactions to open markettransactions. Masulis (1980), Dann (1981), Vermaelen (1981), and Commentand Jarrell (1991) find that abnormal returns from fixed price tender offers arewell in excess of 10% and that the average premium of the tender over the mar-ket price is more than 20%. According to Comment and Jarrell (1991), Dutchauction tender offers lead on average to an abnormal return of 8% during the3 days around the announcement. In contrast, open market transactions byUS corporations were found by virtually all studies to result in much smallerabnormal returns of approximately 3% (see Table 1).

The studies cited so far all provide strong evidence for the validity of theundervaluation signal hypothesis. Ikenberry et al. (1995) observe a strong neg-ative relationship between the market-to-book ratio prior to the buy-backannouncement and the magnitude of positive abnormal returns thereafter.Abnormal returns were also found to be larger for firms whose stocks un-derperformed in the market during the days before announcement (Stephensand Weisbach 1998; Comment and Jarrell 1991; Ikenberry et al. 1995). Bothresults reconcile neatly with the view that the higher the potential is for

5 For an international overview on the open market share repurchase regulations see Kim et al.(2004). The regulations in other European Union countries are fairly similar to those in Germany.


Table 1 Prior empirical results on abnormal returns from announcing open-market repurchaseprograms (OMR)

Country Study Abnormal returns Dataset

US McNally (1999) CAR [−1; +1]: 2.5% 702 OMR (1984–1988)Grullon and Michaley (2002) CAR [−1; +1]: 2.7% 4,443 OMR

(1980–1997)Vermaelen (1981) CAR [−1; +1]: 3.7% 243 OMR (1970–1978)Stephens and Weisbach (1998) CAR [−1; +2]: 2.7% 591 OMR (1981–1990)Ikenberry et al. (1995) CAR [−2; +2]: 3.5% 1,239 OMR

(1980–1990)Comment and Jarrell (1991) CAR [−1; +1]: 2.3% 1,197 OMR

(1984–1988)Australia Lamba and Ramsay (2000) CAR [−1; +1]: 3.3% 103 OMR

(1989–1998)Canada Li and McNally (1999) CAR [−2; +2]: 3.6% 183 OMR (1989–1992)

Ikenberry et al. (2000) CAR [−15; +15]: 0.9% 1,060 OMR(1989–1997)

France Ginglinger and L’Her (2006) CAR [0; +1]: 0.6% 363 OMR (1998–1999)Germany Schremper (2002) CAR [−1; +1]: 4.1% 112 (mostly) OMR

(1998–2000)Gerke et al. (2003) CAR [−1; +1]: 6.1% 120 OMR (1998–2000)Seifert and Stehle (2003) CAR [−1; +1]: 5.9% 192 OMR (1998–2003)

Japan Zhang (2002) CAR [−1; +2]: 6.0% 39 OMR (1995–1999)Korea Jung and Lee (2003) CAR [0; +5]: 2.8% 382 OMR (1994–1998)Switzerland Dumont et al. (2004) CAR [−2; +2]: 1.8% 10 OMR (1999–2003)UK Rau and Vermaelen (2002) CAR [−2; +2]: 1.1% 126 OMR (1985–1998)

Oswald and Young (2002) CAR [−1; +1]: 1.4% 266 (mostly) OMR(1995–2000)

Rees (1996) CAR [−2; +2]: 0.3% 882 OMR (1981–1990)

an actual undervaluation, the stronger are the signaling effects. Vermaelen(1981) shows evidence that the strength of the signal is also a function ofits credibility. He discovers that abnormal returns increase in the amount ofshares held by management as well as in the announced portion of outstand-ing equity to be repurchased (see also Comment and Jarrell 1991; Ikenberryet al. 1995). The assumption is that signal credibility increases if manager wealthis at risk. Moreover, the extent of information asymmetries between manage-ment and investors also seems to have a bearing on announcement effects.Ikenberry et al. (1995) document that price effects are inversely related tofirm size. Given that smaller firms disclose less information to capital marketsand are less researched by equity analysts, information asymmetries shoulddecrease with firm size. Taken together, buy-backs seem to serve as a crediblesignaling device for managers who seek to convey to investors that the marketcapitalization of their firm is lower than its true value.6

Among others, Shoven and Simon (1987), Bagwell and Shoven (1988), Evanset al. (2000), and Li and McNally (1999) have explicitly tested the validity of the

6 Further studies that underscore this insight include Netter and Mitchell (1989) and Bartov (1991).Wansley et al. (1989) directly assess buy-back motivations by means of questionnaires and foundthat perceived undervaluation was indeed one of the most frequently quoted motivations.


free cash flow hypothesis. They find a positive correlation between abnormalreturns and measures of excess funds at the discretion of management. Theyconclude that buy-backs can be an effective means of convincing the marketthat management does not invest excess funds into high-private-benefit, low-NPV projects. Moreover, Stephens and Weisbach (1998) observe that firms withmore excess cash ceteris paribus tend to buy back a larger volume of shares,indicating that repurchases are indeed used by some firms to reduce excess cash.

Bagwell and Shoven (1989) provide some evidence that is inconsistent withthe tax efficiency hypothesis. Cash distributions through repurchases have notdeclined in the US immediately subsequent to the 1986 Tax Reform Act al-though the act abolished the tax advantage of repurchases over dividends. Lieand Lie (1999) qualify this result by showing that a large number of US firmsdid indeed prefer regular dividend increases over open market repurchases inthe late eighties.

Some studies have measured market reactions to US repurchase announce-ments that were perhaps motivated at least to some extent by a firm’s objectiveto deter a takeover attempt. Dann and DeAngelo (1988), Davidson and Gar-rison (1989), and Denis (1990) observe negative abnormal stock price returnsand thereby corroborate the hypothesis that this type of buy-back transactionviolates shareholders’ interests.

In summary and not surprisingly, there is no single overriding motivation forUS firms to repurchase shares. However, “undervaluation signaling” seems tobe the most prominent motivation.

Empirical results for other countries are broadly in line with those for theUSA. Cumulative returns around the announcement day are on average strictlypositive (see Table 1), and evidence of the motivation to repurchase shares ismostly consistent with the undervaluation signaling hypothesis. The UK, how-ever, marks an exception. Rau and Vermaelen (2002) argue that legal restric-tions severely constrain UK firms in using repurchase announcements as anundervaluation signaling device. For example, UK firms are not permitted torepurchase shares in the 2-month period preceding earnings or when directorspossess price-sensitive but unpublished information about the firm. The authorssuggest that repurchases are largely motivated in the UK by tax consequencesfor institutional shareholders such as pension funds.

We are aware of three related empirical studies for Germany. Schremper(2002) analyses 112 buy-back announcements from the period between May1998 and December 2000 and finds significant abnormal returns of around2.6%. The sample of Gerke et al. (2003) comprises 120 buy-back announce-ments from May 1998 to December 2002, which are uncontaminated by othercoincidental news events. The authors find average abnormal returns on theannouncement day of 6.1%. They subdivide their sample to measure differ-ences in abnormal returns between a) firms that either belong to the DAX100 index (+2.7%), the Nemax index (+9.0%), or the small cap index (+4.8%),b) firms that either stated undervaluation (+8.9%) or the exchange of cash intoa superior acquisition currency (+5.2%) as their main repurchasing motivation,and c) announcements that occurred during the general upturn of German


equity markets between May 1998 and February 2000 (+3.7%) and announce-ments that occurred during the subsequent bear market (+7.1%). Seifert andStehle (2003) largely confirm the results of Gerke et al. (2003) but reject theirfindings on the effect of bull and bear markets on price effects. We extend theexisting empirical literature on German repurchase announcements by usinga larger sample size, by investigating price effects around the initial disclosureto seek AGM approval for a buy-back plan, and by conducting multivariateregression analyses on a richer set of independent variables.

5 Data

Data on AGM approvals of share repurchase plans can be downloaded from thewebsite of Germany’s financial services authority BaFin (http://www.bafin.de).The BaFin reports the names and addresses of the firms and the day on whichthey obtained AGM approval. From May 1998 until April 11, 2003, 483 corpo-rations sought AGM approval for a total of 785 buy-back plans. Roughly 70%of the approvals were granted in the three months May, June, and July, whenmost AGMs take place in Germany. However, we do not treat AGM approvalas a price-relevant event for two related reasons. Firstly, the information that afirm plans to establish a repurchase program has to be communicated to cap-ital markets well in advance of the AGM. Secondly, repurchase programs areapproved by the AGM in the vast majority of cases, so that “surprises” arevery rare. As a consequence, we conducted a key-word search on several newsdatabases such as Reuters, Bloomberg, and Factiva to find reports on initialstatements by German firms which were considering establishing a repurchaseprogram in the near future. This search strategy yielded 321 such statements.We matched initial statements with AGM approvals to measure the time thattypically elapsed between these two events. As shown in Figure 1 below, AGMapproval occurred on average 27 weeks after the initial statement.

Finally, we searched the ad hoc announcements database of the DeutscheBörse AG and the news databases mentioned above to find announcements byfirms with authorized repurchase programs that they intend to buy back shares.For our observation period from May 1, 1998, to April 11, 2003, we identi-fied 237 common share buy-back announcements by 181 different Germanfirms. Figure 2 shows the number of announcements per month. Repurchaseannouncements were made on average 21 weeks after the AGM took place(see Figure 1 above). Taking into account that 761 AGM approvals occurred

Intention to obtain AGMauthorization for buy-back plan

AGM authorizationof buy-back plan

Announcement(s) of imminentbuy-back transaction

time

3- 125 weeks(average: 27 weeks)

3- 77 weeks(average: 21 weeks)

Fig. 1 Time line of share repurchase programs in Germany


0 0

1

0 0 0

1

2

3

1

2

1 1

3

7

5

8

5

7

8

3

7

2

3

4

9

8

3

6

7

8

7

2

0

6

5

4

6

12

8 8

5

4

3

1 1

2

5

2

8

7 7

4 4

1 1

3

4

1 1

5

10

15M

ay-9

8

Jul-9

8

Sep

-98

Nov

-98

Jan-

99

Mar

-99

May

-99

Jul-9

9

Sep

-99

Nov

-99

Jan-

00

Mar

-00

May

-00

Jul-0

0

Sep

-00

Nov

-00

Jan-

01

Mar

-01

May

-01

Jul-0

1

Sep

-01

Nov

-01

Jan-

02

Mar

-02

May

-02

Jul-0

2

Sep

-02

Nov

-02

Jan-

03

Mar

-03

'98: 4 '00: 67'99: 51 '01: 63 '02: 43 '03: 9

Fig. 2 Share repurchase announcements in Germany (N = 237)

in the period from May 1, 1998, to Nov. 15, 2002 (22 weeks prior to the end ofour observation period), this implies that only three out of ten AGM approvalswere actually followed by a repurchase announcement.

For the purpose of our event study, we dropped all initial statements forwhich we were not able to pinpoint the exact day of the very first news release(e.g., because the statement was part of the invitation letter to the AGM) orfor which we identified coincident, confounding news (e.g., because the state-ment was buried in a more comprehensive disclosure of material information).This filtering technique significantly reduced the number of initial statementobservations to 111. Of the 237 observations on repurchase announcements,we dropped four because they concerned tender offers.7 We excluded anothernine observations due to coincident confounding events such as a board changeor an earnings disclosure. Thus, 224 repurchase announcement observationsremained in the sample.

For the sake of brevity and for the sake of comparability with US studies,we decided to conduct a regression analysis only for the abnormal returns fromrepurchase announcements and not for abnormal returns from initial state-ments. In the following we define the variables used as covariates. Table 2below shows the corresponding descriptive statistics.

• MTB: The Market-to-Book (MTB) ratio is defined as the market value ofequity 2 days before the announcement date divided by the book valueof equity as reported in the most recent financial statement prior to therepurchase announcement. The higher the MTB ratio of a firm, the higherinvestors’ expectations are with regards to the firm creating economic value

7 These are AGIV (4-Apr-00, fixed-price tender offer to common shareholders), Friedrich Gro-he (7-Oct-99, fixed-price tender offer to minority holders of preferred stock), Kögel Fahrzeuge(7-Dec-98, fixed-price tender to common shareholders) and Krones AG (18-Jan-99, Dutch-auctiontender offer to common shareholders).


Table 2 Descriptive statistics (N = 224)

Dependent variables Independent variables Independent dummy variables

AR CAR CAR MTB SIZE PAST CASH Averages[0] [−1; +1] [−1; +10] RETURN(%) (%) (%) (%) (%)

Max. 40.6 36.8 56.5 14.80 11.55 49.6 125.3Min. −13.6 −18.5 −36.6 0.18 1.76 −71.8 1.3 NMLISTING 0.43 TARGET 0.08Avg. 4.9 6.0 7.0 2.55 5.49 −11.1 39.3 UNDERVAL 0.43 CONT.25 0.33Median 3.1 4.6 5.0 1.71 5.06 −7.9 32.1 SERVICE 0.29 CONT.50 0.35Stdev. 7.8 9.4 14.3 2.78 2.10 21.2 30.4 FINANCIAL 0.13 CONT.75 0.11

in the future. Firms with high MTB ratios are therefore often referred to as“growth” firms. Expectations are more modest for so-called “value” firmswith low MTB ratios. As is done in most other related studies, we arguethat for value firms perceived undervaluation is a more likely factor in thedecision to repurchase than for growth firms. A negative relation betweenabnormal returns and MTB would then be consistent with the “undervalua-tion signaling” hypothesis.

• SIZE: Firm size is expressed as the logarithm of the firm’s enterprise valuein million Euros. Enterprise value is defined as the sum of the market valueof equity and the book value of interest bearing debt. Size is treated as aproxy for the extent of information asymmetries between a firm and capi-tal markets. The larger a firm, we argue, the more stringent the disclosurerequirements will be, and the more analysts will cover the firm resulting in agreater amount of firm-specific information being publicly available. Manag-ers of small firms thus have more potential in signaling private informationby means of repurchase announcements than managers of large firms.

• NMLISTING: This dummy variable is set to 1 if the firm was traded on theNeuer Markt – a by now abolished segment of the German stock exchange foryoung and innovative firms. Like SIZE, NMLISTING also serves as a proxyfor information asymmetries between the company and its investors. Firmslisted on the Neuer Markt are typically characterized by shorter track recordsand a higher degree of uncertainty regarding future industry prospects thanmore mature firms listed on other exchange segments. As a consequence,signals should be more powerful and abnormal returns should be higher forNeuer Markt firms.

• PASTRETURN: This variable measures the cumulated returns of a firm’sstock over the 30 day-period prior to our event window [−31; −2].8 Like thevariable MTB, it attempts to capture the potential for undervaluation of afirm’s stock. Of course, a poor stock price performance prior to a repurchaseannouncement may simply reflect a true decline in firm value. However,

8 We have used both absolute and abnormal returns in our regressions. The coefficients and theirstatistical significance levels are virtually the same for the two alternative specifications.


Table 3 Motivations to repurchase shares as declared by management (N = 185)

Number of declarations Percent of 185 sample firms

Acquisition currency 107 58Undervaluation 96 52Employee participation programs 32 17Cancellations/pay-outs to shareholders 27 15Other 5 3Sum 267

there is ample empirical evidence that markets tend to overshoot in bothdirections in their assessments of firms’ fundamental value (Shiller 2000).Arguably then, the potential for undervaluation is larger for firms with apoor prior stock performance than for those with positive past returns. Weexpect a negative relation between PASTRETURN and price effects aroundthe announcement. While not shown in this paper, we also ran regressionswith PASTRETURN variables for the 45-day and the 60-day periods preced-ing the announcement event. We found that results are not affected by thechoice of period length.

• UNDERVAL: This dummy variable is set to 1 if a firm states “undervalua-tion” as a main motivation for repurchasing own shares. Although Germanfirms are not legally obliged to disclose their motivations for share buy-backsex ante, it is common practice that they provide such information voluntar-ily as part of the announcement. Out of the 224 firms in our final sample,185 disclosed their motivations.9 In many cases more than one motivationwas stated. Table 3 below reports the total number of declarations and thepercentage of firms per type of motivation. 96 or roughly one half of thefirms stated a perceived undervaluation of their stock as a primary reason tobuy back shares.10 If such an undervaluation were credible, we would expectabnormal returns for repurchase firms that state undervaluation as their mainmotivation to be ceteris paribus higher. However, because there is no penaltyfor wrong statements, also firms that are in fact not undervalued would facestrong incentives to declare undervaluation as one of their motivations. Wetherefore assume that rational investors treat voluntary statements regardingrepurchase motivations as cheap talk, and we expect to observe no differencein announcement effects between firms stating different motivations.

• CASH: This variable is defined as the amount of liquid assets over the bookvalue of equity.11 It is used as a proxy for the amount of free cash that isat the disposal of managers in relation to the amount of capital provided byshareholders. If the cash position is large, and if the firm does not possess

9 UNDERVAL was set to zero for the 39 firms that did not specify their motivations.10 In the Canadian sample of Li and McNally (1999), more than two-thirds of the 183 firms statedundervaluation as their main motivation.11 In section 6 we also discuss empirical results for an alternative specification that carries themarket value instead of the book value of equity in the denominator (CASH/MTB).


enough profitable investment opportunities, investors will welcome sharerepurchases as a means to avoid negative-NPV projects in general and pri-vate benefit consumption by managers in particular.

• CONTROL25, CONTROL50, and CONTROL75: These dummy variablesare set to 1 if the aggregate holdings of the two largest shareholders exceedcertain threshold values.12 Control25 is 1 for holdings greater or equal to 25%and below 50% of total shares outstanding. CONTROL50 is 1 if holdings aregreater or equal to 50% and smaller than 75%. CONTROL75 is 1 if holdingsare 75% or greater. We attempt to measure any price effects that might arisefrom a firm’s specific governance structure. If a firm is controlled by only afew large blockholders, minority outside shareholders may have to fear thatinside blockholders will exercise their power in their own interest, e.g., byinducing the firm’s management (which in the case of manager- and family-controlled firms might actually be identical with or at least closely related toblockholders) to transact with them at favorable terms or to invest in pro-jects that one-sidedly benefit them.13 If the extraction of private benefits byinside shareholders is indeed prevalent, one should expect greater positiveprice effects for firms with highly concentrated ownership. In these cases, abuy-back implies an unexpected payout of cash that outsiders might interpretas a signal for a closer alignment of insider and outsider interests.

• TARGET: In section 1 we argued that managers can destroy shareholdervalue by repurchasing shares solely in an effort to fend-off a value-creatinghostile takeover attempt by a raider. Because the threat of such an attemptmay have only been known to target firm managers but not to the publicat the announcement date, it posed a big challenge to identify target repur-chase firms in our sample. We therefore searched for situations in which itwas more likely that target firm managers indeed used repurchases to pro-tect their control rights against potential raiders. TARGET only takes on thevalue 1 if both of the following conditions are satisfied: Firstly, repurchas-ing shares must bear the potential of reducing the free float held by outsideshareholders to a level that would make it difficult for raiders to acquire acontrolling stake in the firm over the open market. Consequently, we requirethat the free float on the announcement date be below 25%.14 Secondly,we require that managers or family owners have a substantial but not the

12 Alternative specifications based on the holdings of the largest single shareholders and theaggregate shareholding of the three largest shareholders leave empirical results regarding controlstructure and announcement effects unchanged.13 Ehrhardt and Nowak (2003) show in their empirical analysis that private benefits for familyblockholders can indeed be very large in German firms and that stocks of firms where foundingfamilies own more than 75% underperformed their peers significantly over a three-year period.Nenova (2003) finds that the value of corporate voting rights, which can be interpreted as lowerbound for actual private benefits of the controlling shareholders, was more than twice as high inGermany than in the US in 1997.14 While not reported in this paper we also ran regressions based on alternative threshold valuesfor the free float (30 or 35%). Empirical results remained unchanged.


majority equity stake in the firm.15 Only then must inside shareholders fearthat raiders can acquire sufficient control in the firm and subsequently curbexisting opportunities for insiders to extract private benefits from the firm.This situation is most likely if the combined stake of managers and familyowners is between 25 and 50%.

• FINANCIAL and SERVICE: We introduced two industry dummies to con-trol for industry effects. FINANCIAL is 1 if a firm belongs to the financialservices sector, and SERVICE is 1 if a firm belongs to all other service indus-tries. For firms from the manufacturing industry, both dummies are set to zero.In an extended model, we have also controlled for time effects by includingdummies for the years 1999–2003. Because no trend emerged, and becausethe results for the other variables remained unchanged, we omit the resultsfor such an extended model in this paper.

If the “undervaluation signal” hypothesis holds, the coefficients for MTB,SIZE, and PASTRETURN should all carry negative signs, and NMLISTINGshould carry a positive sign. If managers aim to reduce agency difficulties byliberating excess cash, the coefficient for CASH should carry a positive sign.Positive coefficients for CONTROL50 and CONTROL75 would provide addi-tional evidence for a reduction of conflicts of interests between corporate insid-ers and outsiders. Finally, a negative and statistically significant coefficient ofTARGET would be consistent with the “takeover deterrence” hypothesis.

6 Methodology

We conduct a standard market-model event study to measure price effects frombuy-back announcements. Price effects correspond to abnormal returns on afirm’s stock on the announcement day [0] or over a short time window sur-rounding that date (e.g., days [−1; +1]). Expected “normal” daily returns R∗

itare derived from a CAPM market model. We use daily share price returns dur-ing the time window [−270; −60] and the ordinary least square (OLS) model in(1) to estimate the parameters for the market model.

Rit = αi + βiRmt + εit for t = −270, −269, . . . , −60

with E(εit) = 0 and Var(εit) = σ 2(εit). (1)

Share price data are taken from Datastream, and daily returns for day t arecomputed by subtracting the natural logarithm of the share price at the end ofday t-1 from the natural logarithm of the share price at the end of day t. Marketreturns are based on the broadly defined Composite DAX (CDAX) index.

15 Because managers affiliated with the owner family could well have a different surname, we werenot able to distinguish between managing families and pure owner families.


Coefficient estimates from (1) and daily market returns are entered into (2)to obtain expected returns R∗

it for the shares of firm i.

R∗it = ai + bi × Rmt

for t = −30, −29, . . ., +29, +30. (2)

Abnormal returns for day t are defined as the difference between the observedreturn and the expected return for that day [see (3)].

ARit = Rit − R∗it

for t = −30, −29, . . ., +29, +30. (3)

The t statistics in equations (4) and (5) are used to test the null hypothesis thatabnormal returns on a particular day and cumulative abnormal returns for agiven period [t; t+n] are not different from zero.

T = ARit

σ(ARi)with σ(ARi) equal to the standard error of the

estimate from (1) (4)

T = CARt+ni

(√

n · σ(ARi))with CARt+n

i =t+n∑

j=t

ARi,j. (5)

We use OLS-regression with White-corrected standard errors to directly testthree hypotheses regarding the motivations of German firms to repurchaseshares. As noted above, the “tax efficiency” hypothesis cannot be tested thisway. However, because Germany experienced a substantial shift in its tax regimein the middle of our observation period, we are able to provide indirect evidenceon the validity of this fourth hypothesis.

7 Empirical results

7.1 Abnormal returns

Figure 3 below plots average cumulative abnormal stock returns for the 121-dayperiod surrounding the 111 initial statements by German firms that they seekAGM approval for a share repurchase plan.

Cumulative abnormal returns are virtually zero until 5 days before the event,but significantly positive thereafter. While the average price effect on the eventday itself is only a moderate 1.5%, cumulative price effects jump to over 5% forthe 11 days surrounding the event day. There are two mutually non-exclusiveexplanations why positive abnormal returns can already be observed shortlybefore day zero. Firstly, although we have carefully searched for the very first


-5%

0%

5%

10%

-60 -50 -40 -30 -20 -10 0 10 20 30 40 50 60

CAR, %

Days

AR [0] = 1.47% (4.86)

CAR [-1;+1] = 2.53% (4.84)CAR [-1;+5] = 2.87% (3.59)CAR [-5;+5] = 5.21% (5.19)CAR [-15;+15] = 6.89% (4.10)

Fig. 3 Average cumulative abnormal returns from statements to seek AGM approval for a buy-back plan (N = 111) Note: Parentheses show t-statistics. Corresponding median values are 0.89%for AR[0], 1.75% for CAR[−1; +1], 1.69% for CAR[−1; +5], 4.43% for CAR[−5; +5] and−5.22% for CAR[−15; +15]

point in time when information on the initial statement was publicly available,we may have missed some items of public information that had been disclosed tothe markets in advance. Secondly, positive pre-announcement abnormal returnscan indicate that no public information was available prior to the firm’s state-ment, but that firm insiders traded on their private information regarding suchan imminent statement.

Given that cumulative abnormal returns around the initial statement to seekAGM approval total around 5%, we can roughly infer the expected averagemagnitude of abnormal returns from a subsequent repurchase announcement.The price reaction (RA = 1 + rA) around date A of the initial statement shouldbe a positive function of the expected abnormal share price appreciation onthe later repurchase announcement date B(RB = 1 + rB) and the probability pthat investors assign to the actual occurrence of a buy-back announcement.16

16 We use the same value for p for all firms and thereby implicitly assume that investors do notestimate that probability conditional on some observable firm characteristics. Would they do so,abnormal returns around the initial announcement to seek AGM approval should be different forfirms with a high probability and for firms with a low probability to subsequently repurchase shares.To test for differences in abnormal returns from the initial announcement we subdivided the sampleof 111 firms into a subsample of repurchase firms and a subsample of non-repurchase firms. Wefound that average abnormal returns do not differ in a statistically significant way between the twosubsamples. Therefore, one can largely rule out the possibility that investors are ex ante able tosystematically discriminate between firms that subsequently repurchase shares and those that donot.


Taking on the perspective of a risk neutral investor who, at date A, determinesthe maximum share price appreciation RA at which it is no longer worthwhileto buy the stock in question, we can write:17

RA = p × (RA × RB) + (1 − p)1. (6)

Collecting terms and solving for RB yields

RB = 1p

− (1 − p)

(p × RA). (7)

From section 4 we know that the unconditional probability that AGM approvalof a share repurchase plan is followed by a repurchase announcement is 237/761=31.1%. Inserting p = 0.311 and rA = 5.21% (=CAR[−5; +5]) into equa-tion (7) yields an expected abnormal return of rB = 11.0% for the repurchaseannouncement. If we insert rA = 2.87% (=CAR[−1; +5]), rB drops to 6.2%.

Figure 4 below shows average cumulative abnormal returns around the 224dates in our sample when German firms announced their intention to repur-chase their own shares over the open market. Average abnormal returns on theevent day are positive at 4.9%, with 83% of the sample firms showing positiveabnormal returns. Abnormal returns are also positive on average over thesubsequent days. Cumulative abnormal returns are roughly 6% for the eventwindow [−1; +1] and almost 7% for the event windows [−1; +5] and [−1; +10].All return figures are significantly different from zero at the 1% level. Figure 4also shows that share prices of German sample firms experienced a conspicuousabnormal downward trend over the thirty trading days before the announce-ment date. Comment and Jarrell (1991) document a very similar pattern intheir analysis of some 1,200 US open market repurchase programs. Announce-ments are preceded in their study by negative net-of-market stock performance,and positive excess price effects reverse about half of this underperformance.In Germany, abnormal returns for the thirty post announcement days reversealmost the entire 30-day pre announcement underperformance. However, ifone looks at a wider event window from day −60 to day +60, in fact the verysame pattern as in the USA emerges.

Can observed price effects around date B be reconciled with the expectedprice effects implied by the abnormal returns around date A? In fact, observedpost announcement returns (roughly 7%) lie within the expected range (roughly6% to 11%). However, observed cumulative returns for a wider event windowof, say, 60 days before date B to 60 days after date B are negative on averageand therefore clearly lie outside the expected range. Given that positive postannouncement returns typically do not fully compensate for preannouncementunderperformance, rational investors should factor in such a pattern into theirassessments of a firm’s value on date A. In such a scenario, we would expect

17 For the sake of simplicity, we assumed a discount rate equal to zero.


-10%

-5%

0%

5%

-60 -50 -40 -30 -20 -10 0 10 20 30 40 50 60

CAR, %

Days

AR [0] = 4.90% (19.58)

CAR [-1;+1] = 5.97% (13.77)

CAR [-1;+5] = 6.82% (10.30)

CAR [-1;+10] = 6.99% (8.07)

CAR [-30;-2] = -7.54% (-5.59)

Fig. 4 Average cumulative abnormal returns from buy-back announcements (N=224) Note:Parentheses show t-statistics. Corresponding median values are 3.06% for AR[0], 4.55% forCAR[−1; +1], 5.62% for CAR[−1; +5], 4.97% for CAR[−1; +10] and −6.40% for CAR[−30; −2]

insignificant or even negative average abnormal returns on date A. The magni-tude of positive announcement returns around date A poses a puzzle for whichwe have no full answer.18

Taken together, we find aggregate abnormal returns for the two announce-ments to establish and to conduct a repurchase plan of between 10 and 12%for Germany. These effects are considerably higher than those documented forother countries and in particular higher than the 3% typically found for the USmarket (see Table 1 above). We now turn to the analysis of the motivations ofGerman firms to repurchase shares.

7.2 Regression analysis

Table 4 below presents the results of six OLS regressions. The full model includes12 independent variables. For the reduced model we have dropped three vari-ables that are strongly correlated (correlation coefficients exceeding 0.25) to

18 Possibly, the market assigns to all firms, irrespective of whether they establish a repurchaseplan or not, a certain probability that stock prices will experience a prolonged period of negativeabnormal returns in the future. However, only those firms with an approved repurchase plan havethe opportunity to respond to a potential undervaluation by means of a repurchase transaction.All other undervalued firms lack such an instrument to signal undervaluation. On date A, rationalinvestors will then only partially factor in the likely negative price effects prior to date B since theex ante probability for such an underperformance does not necessarily depend on the existence ofa repurchase plan. As a consequence, investors might perceive the establishment of such a plan as asignal that the firm in question finds future undervaluation likely and that its management endowsitself with an instrument to combat information asymmetries.


Table 4 OLS regression results (N = 224)

AR[0] CAR [−1; +1] CAR [−1; +10] AR[0] CAR [−1; +1] CAR [−1; +10]

CONSTANT 11.53%*** 11.63%*** 15.02%*** 11.60%*** 12.69%*** 13.53%***(0.000) (0.001) (0.003) (0.000) (0.000) (0.000)

MTB −0.51%*** −0.45%*** −0.69%** −0.44%*** −0.38%*** −0.63%**(0.000) (0.003) (0.013) (0.001) (0.009) (0.020)

SIZE −0.93%*** −1.00%*** −1.29%*** −1.01%*** −1.13%*** −1.26%***(0.001) (0.001) (0.003) (0.000) (0.000) (0.001)

NMLISTING 2.95%*** 2.43%* 2.31%(0.006) (0.058) (0.266)

PASTRETURN −3.13% −7.430%** −13.56%*** −3.94% −7.95%** −14.09%***(0.243) (0.033) (0.001) (0.155) (0.026) (0.001)

UNDERVAL 0.91% 1.83% 5.12%*** 1.35% 2.10%* 5.59%***(0.369) (0.125) (0.007) (0.172) (0.068) (0.002)

CASH 0.09% 1.56% −0.45%(0.957) (0.406) (0.871)

TARGET 0.21% −3.25% −5.81% 0.36% −3.25% −5.80%(0.922) (0.141) (0.127) (0.873) (0.131) (0.113)

CONTROL25 −2.09% −1.42% −2.98%(0.183) (0.414) (0.227)

CONTROL50 −1.41% −1.26% −2.00%(0.340) (0.490) (0.437)

CONTROL75 0.23% 3.05% 2.40% 0.91% 3.39%* 3.84%(0.902) (0.155) (0.500) (0.571) (0.065) (0.230)

SERVICE −2.78%** −3.60%** −4.63%** −2.65%** −3.44%*** −4.67%**(0.030) (0.013) (0.048) (0.034) (0.018) (0.042)

FINANCIAL −2.24%* −2.98%* −3.65%* −3.02%** −3.55%** −4.28%**(0.079) (0.054) (0.080) (0.018) (0.020) (0.041)

R2 0.218 0.216 0.216 0.184 0.196 0.177F Stat 4.51 3.98 4.53 5.21 5.426 6.998Significance F 0.000 0.000 0.000 0.000 0.000 0.000

Reported results are OLS regression coefficients with White-corrected standard errors (p-valuesin parentheses)

* Significantly different from zero at the 10%-level

** Significantly different from zero at the 5%-level

*** Significantly different from zero at the 1%-level

one or more of the other variables. Independent variables are abnormal returnson the announcement date and cumulative abnormal returns for the two eventwindows [−1; +1] and [−1; +10], respectively.

All six models show unequivocally that abnormal price effects from repur-chase announcements are on average greater for firms with lower MTB, forfirms with a smaller enterprise value (SIZE), for firms that experienced lowershare price returns prior to announcement (PASTRETURN), and for firmsthat stated undervaluation as a motivation for the share repurchase (UNDER-VAL). For the other variables results are more ambiguous. NMLISTING ishighly correlated with firm size and does not seem to provide too much addi-tional explanatory power as compared to SIZE for wider event windows. Thecoefficients of both CONTROL25 and CONTROL50 are negative but not sig-nificant. The coefficient of CONTROL75 always carries a positive sign but is


only weakly significant for the reduced model and CAR[−1; 1]. Price effectsfrom repurchases that are potentially perceived by investors as a takeoverdefense device (TARGET) are virtually zero on the announcement day butstrongly negative (and slightly significant) when measured over a 2- or 11-dayobservation period. Finally, the amount of cash on a firm’s books does not affectshare price reactions at all (CASH).

We interpret these results as providing strong evidence in favor of the sig-naling hypothesis. Investors seem to be more willing to update their beliefsregarding a firm’s current market position and its future prospects if there isa bigger potential for repurchase announcements to provide new informationto market participants and if undervaluation of the firm’s equity is more likely.We have argued above that an undervaluation signal carries more informationcontent for smaller firms because information asymmetries between managersand investors are presumably larger. We have also argued that undervaluationtends to be more likely if the firm is a value stock with a low MTB ratio and ifshare prices experienced a dip, given that uncertainty is high enough to indeedjustify this dip.

The fact that past share price returns, whether measured in absolute or abnor-mal terms, are inversely related to announcement effects indicates a deliberatetiming of the announcement by management; and deliberate timing is consis-tent with the view that firms use buy-backs to signal undervaluation. Over the30-day period prior to announcement 162 of the 224 sample firms experiencednegative cumulative stock returns.

For these observations, CAR[−1; +10] is significantly higher at 7.6% as com-pared to 5.4% for the 62 firms with a positive value for the variable PASTRE-TURN.19 We conduct a simple check for deliberate timing. If the repurchasedecision of most of the 162 firms with negative values for PASTRETURN wasindeed triggered by poor stock price performance, we should not observe toomany prior instances in their recent stock price history where prices experi-enced a dip of similar magnitude. For each of these 162 firms, we thereforecounted the number of trading days over the period between AGM approvaland repurchase announcement on which cumulative returns over the prior 30trading days had been similarly bad or even worse than the 30-day performanceprior to announcement. For 59 firms we could not find a single day wherethis had been the case. However, for the remaining 103 firms, we counted onaverage 28 days with a similar or worse history of declining stock prices. Thesefigures change somewhat if we only consider observations with values for PAST-RETURN smaller than −15% (−30%). For 44 (22) out of 81 (39) firms, thecumulative stock price performance prior to announcement had indeed beenthe worst since AGM approval, and the average number of days with a similaror worse 30-day history drops to 20 (18) for the other firms. Taken together,we find that over half of the sample firms with a particularly poor stock price

19 Cumulative abnormal returns during the thirty days prior to announcement were negative for159 firms. The average CAR[-1; +10] for this sample was 7.4% as opposed to 5.8% for the 65observations with positive abnormal past returns.


performance over the 30 pre-announcement trading days had not experienced asimilar dip since AGM approval. For these firms, cumulative abnormal returnsaround the repurchase announcement were especially high at CAR[−1; +10] =11.9%. We conclude that a number of repurchase announcements in our samplehad at least been partially triggered by a poor and, according to the firm’s man-agement, apparently unjustified performance and that investors were factoringin past performance in their assessment of the credibility of an undervaluationsignal.

Coincident voluntary statements by managers that they view their firm asundervalued also seem to have measurable effects on abnormal announcementreturns. Table 4 shows that cumulative abnormal returns between day −1 andday +10 are on average more than 5% points higher if a firm made such a state-ment. This result is surprising given the fact that the statement itself is virtuallyfree of cost and should therefore lack any credibility. Apparently, however, themarket assigns some credibility to it, perhaps since managers have to justifysuch a statement in the subsequent AGM. Interestingly, Ikenberry et al. (1995)document a similar phenomenon for their sample of 1,239 US open marketrepurchase announcements which took place over the course of the 1990s. Fif-teen percent of the firms stated a reason for the repurchase in the associatedpress release and 38 firms stated undervaluation as their primary motivation.Average abnormal returns for this subsample of 38 firms are 5.3% versus 3.5%for the full sample of 1,239 observations. We conclude that a voluntary underval-uation statement might invigorate an undervaluation signal in a similar fashionas a poor recent stock performance.

We find no evidence corroborating the “excess cash” hypothesis. In section 1we argued that low MTB ratios in conjunction with large cash positions couldindicate financial slack on a firm’s books. Cash-based share repurchases reducefinancial slack and thereby mitigate agency problems between managers andowners, potentially leading to higher valuations. Although the coefficient forMTB carries the expected sign in Table 4, we do not observe any clear rela-tionship between abnormal returns and a firm’s cash position. While not shownin Table 4, we also ran a regression on an extended model specification thatincluded the interaction term CASH/MTB. For all three independent variablesthe coefficients for this interaction term were statistically insignificant. We alsotested whether the effect of large cash positions on abnormal announcementreturns sets in only for specific control structures, i.e., for cases where the com-pany is largely controlled by inside blockholders who might be in a positionto extract rents from outside minority stockholders. It emerged that the coeffi-cients for the relevant interaction terms with CONTROL50 and CONTROL75variables were statistically insignificant. On a more general level, however,we detected some indication that agency conflicts between inside blockhold-ers and outside minority shareholders are mitigated through repurchases. Thecoefficients for CONTROL75 are always positive and in one instance also sta-tistically significant. If a firm is controlled by two shareholders with a combinedownership stake in excess of 75% (25 firms in the sample), cumulative abnor-mal returns were on average roughly 3% higher than for firms with dispersed


ownership (49 firms in the sample). We assume that minority investors of firmswith highly concentrated ownership infer from a repurchase announcementthat manager interests are now closer aligned to their own interests. Hence,we cannot fully rule out that some German firms repurchase shares to reduceagency problems.

The negative coefficients of the TARGET variable provides some, albeitweak, evidence for a negative relationship between abnormal returns and theperception by investors that a buy-back transaction aims at deterring a hostiletakeover. The fact that coefficients for TARGET were negative only for widerevent windows and very weakly significant only for the widest event windowCAR[−1; +10] may point to a large degree of initial uncertainty regarding thetrue motivations of management that can only be resolved after (time-consum-ing) further investigations.

As already stated above, we are not able to test the validity of thetax-efficiency hypothesis directly. This hypothesis presumes that a firm shouldsubstitute share repurchases for dividend payments if capital gains are tax-advantaged over dividend income for the most important shareholders. In Ger-many, capital gains have in fact enjoyed a more investor-friendly tax treatmentthan dividend income over the entire period from 1998 to 2003. However, theGerman Tax Reform 2000 substantially increased the relative attractiveness ofrepurchases by cutting the corporate tax rate on retained earnings.20 In 2000the marginal tax imposed on dividend income was 61.3% for investors in thehighest tax bracket, and the marginal tax imposed on capital gains was 51.6%.In 2001, taxes on dividends were 59.1% for the same clientele, and taxes oncapital gains were only 38.4%, implying a tax differential of over 20% points. Ifrepurchases by German firms were strongly motivated by tax efficiency, firms’propensity to substitute share repurchases for dividend payments should haveincreased substantially from 2000 to 2001.

Pan (2005) documents that the Tax Reform 2000 indeed brought substantialchanges in the dividend payout behavior of publicly listed German firms. Whilethe fraction of dividend payers among all listed German firms had hoveredaround 75% between 1990 and 2000, it dropped sharply to 56% in 2001, furtherdown to 53% in 2002, and finally to 47% in 2003. And while the aggregateamount of dividend payments by listed German firms had been growing annu-ally and in real terms for seven straight years between 1993 and 2000, it declinedby over 20% between 2000 and 2001. Pan (2005) concludes that taxation is theleading factor to explain the decline in dividend payments in Germany. Whiledividend payout behavior of German firms thus appears to be largely tax driven,share repurchase behavior of German firms does not appear to have been pos-itively affected by the Tax Reform 2000. Figure 2 shows that the number ofrepurchase announcements declined from 67 in 2000 to 63 in 2001 and furtherdown to 43 in 2002. We are inclined to conclude that taxation-induced dividend

20 At the same time, the minimum holding period after which capital gains from selling sharesare tax-exempt was extended from six months to twelve months. Our example above applies tolong-term investors with an investment horizon greater than 1 year.


substitution is unlikely to be the key motivation for German firms to repurchaseown shares.

In summary, our empirical results show that the overriding motivation ofGerman firms to repurchase shares seems to be undervaluation signaling. Wethereby confirm for Germany one of the main findings of the existing inter-national literature. However, we find that average announcement effects onshare prices are of much greater magnitude in Germany than in other countries(5% + 7% = 12% as compared to roughly 3% in the US). There are threepotential explanations for this phenomenon. The first is based on internationaldifferences in general disclosure requirements for firms. The other two arebased on international differences in the laws governing share repurchases.

Leuz and Wüstemann (2004) demonstrate that it is not the primary aim of theGerman accounting system to maximize information dissemination to outsideinvestors, but rather to support private information channels to privileged insidecapital providers such as “Hausbanks” and large blockholders. Empirical stud-ies show that the information content of financial statements is indeed lessvalue-relevant and less timely than in the US and the UK.21 As a consequence,information asymmetries between managers and (outside) investors might gen-erally be larger in Germany than in the two capital market based financial sys-tems. An additional public disclosure by a German firm would then potentiallyembody more new material information than the very same disclosure by aUS or a UK firm. If different degrees of general information asymmetries areindeed the main determinant of country differences in repurchase announce-ment effects, one should observe similar differences in abnormal returns fromother types of announcements. However, Gebhardt (2001) documents in hisliterature survey that this is not the case. Neither announcements of changes individend payouts nor of seasoned equity offerings lead to stronger price effectsin Germany than in the USA. We therefore tend to reject this first explanation.

The second explanation proceeds from the conjecture that certain Germanlaws for repurchase programs largely rule out repurchases that harm (groupsof) shareholders. The obligation to obtain AGM approval and to treat all share-holders equally in a repurchase transaction can be very effective in deterringmanagers from transferring wealth from one shareholder group to anotherand from all shareholders to themselves. In contrast, US managers theoreti-cally have much more leeway in designing and timing their repurchase plans.We therefore expect to observe a larger portion of repurchase announcementswith negative announcement effects in the USA than in Germany. Bhattach-arya and Dittmar (2002) report that abnormal returns for the 2,405 US openmarket repurchase announcements in their sample are on average 3.6% forthe event window [−1; +1] and have a standard deviation equal to 9.21%. Ifwe assume that these abnormal returns are normally distributed, we obtain anestimated portion of announcements with negative abnormal returns of approx-imately 35%. For our German sample we observe a portion of only 17%. Legal

21 See, e.g., Joos and Lang (1994) and Harris et al. (1994).


deterrence of repurchases that are not in line with shareholders’ interests canthus be assumed to explain international differences in average announcementeffects to some extent.

The third explanation holds that the strictness of German repurchase regula-tion allows for credible undervaluation signals. In section 2 we have argued thatthe prescribed repurchase process makes misleading announcements costly.These costs would make repurchase announcements less attractive for over-valued firms and would thereby increase the portion of undervalued firms inthe pool of repurchasing firms. Investors might then be induced to attribute ahigher credibility to undervaluation signals wrapped in repurchase announce-ments. Grullon and Ikenberry (2000, p. 44) summarize the situation in theUSA: “A primary concern of both regulators and investors is that, because ofthe flexibility and modest disclosure requirements associated with open marketrepurchase programs, there is some potential for companies to mislead inves-tors by announcing repurchase programs while having no intention of buyingstock”. As a consequence, investors should treat repurchase announcements byUS firms with more skepticism. Ikenberry et al. (1995) show that repurchasefirms outperform a control portfolio of non-repurchase firms by an average of12% over the four post-announcement years. Repurchase firms with very lowMTB ratios outperform the control portfolio by even 45%. The authors con-clude that markets largely ignore the information conveyed by US open marketrepurchases and, as a consequence, systematically under-react to undervalua-tion signals. We argue that, given the much stricter repurchase regulations inGermany, markets take more notice of the information contained in repur-chase announcements by German firms and, as a consequence, under-reactionto repurchase announcements is less of an issue in Germany than in the USA.22

Given that undervaluation signaling is the major repurchase motivation of bothUS and German firms, differences in under-reaction patterns could largelyexplain the differences in repurchase announcement effects.

8 Conclusion

In this paper we have analyzed abnormal share price returns from Germanopen market repurchase announcements. Share repurchases are only a recentphenomenon in Germany and are governed by fairly strict laws. We aimedto provide new insights into the role the legal and regulatory environmentplays in inducing corporate insiders to successfully reveal private informationon true firm value. With this aim we measured abnormal price effects aroundtwo events, namely the initial statement by a German firm that it plans toestablish a repurchase plan and the firm’s subsequent announcement that itintends to repurchase shares over the open market. Total abnormal returnsaround these two events add up to roughly 12% and are much higher than

22 Because long-term post-announcement returns are not yet available for a sufficient number ofGerman repurchase firms, we cannot test this hypothesis in this paper.


the 3% typically reported for US repurchase announcements. We tested fourhypotheses regarding the primary motivations of German repurchase firms.Our empirical evidence strongly favors the undervaluation signal hypothesisand thereby confirms the results obtained for other countries. We discussedthree possible explanations for the large discrepancy in average announcementeffects between German and US firms and identified differences in repurchaseregulations as the most likely explanation. The stringent repurchase processprescribed by German law attributes a much higher credibility to undervalua-tion signals than do the lax US regulations.

We conclude this paper with a side note for regulators. The patterns of abnor-mal returns around German repurchase announcements reinforce the legalrequirement that German firms must report an imminent repurchase transac-tion through a public ad hoc disclosure. Since the preceding initial statementby managers to seek AGM approval is also accompanied on average by con-siderable price effects, we are inclined to suggest that such a statement shouldalso be subject to ad hoc disclosure requirements. Otherwise, opportunitiesremain for insider trading – and insider trading is illegal in Germany since 1994.Figure 3 above shows that substantial abnormal returns occurred during the5 days before the respective voluntary statement, thus indicating that insidertrading in the context of repurchase transactions might still be an issue inGermany.

References

Bagwell, L.S.: Share repurchases and takeover deterrence. RAND J. Econ. 22, 72–88 (1991)Bagwell, L.S., Shoven, J.: Share repurchases and acquisitions: an analysis of which firms partici-

pate. In: Auerbach, A.J. (ed.): Corporate Takeovers: Causes and Consequences pp. 191–213. TheUniversity of Chicago Press, Chicago (1988)

Bartov, E.: Open-market repurchases as signals for earnings and risk changes. J. Account. Econ.14, 275–294 (1991)

Bhattacharya, U., Dittmar, A.: Costless versus costly signaling in capital markets: theory and evi-dence from share repurchases, Working Paper, Indiana University (2002)

Bosse, C.: Handel in eigenen Aktien durch die Aktiengesellschaft. Z. Wirtsch. Bankrecht 54, 806–809 (2000)

Comment, R., Jarrell, G.: The relative signaling power of Dutch-auction and fixed-price self-tenderoffers and open-market repurchases. J. Financ. 46, 1243–1271 (1991)

Dann, L.: Common stock repurchases: an analysis of return to bondholders and stockholders.J. Financ. Econ. 9, 113–138 (1981)

Dann, L., DeAngelo, H.: Corporate financial policy and corporate control: a study of defensiveadjustments in asset and ownership structure. J. Financ. Econ. 20, 87–127 (1988)

Davidson, W., Garrison, S.: The stock market reaction to significant tender offer repurchases ofstock: size and purpose perspective. Financ. Rev. 24, 93–107 (1989)

Denis, D.: Defensive changes in corporate payout policy: share repurchases and special dividends.J. Financ. 45, 1433–1456 (1990)

Dumont, P.-A., Isakov, D., Pérignon, C.: Stock repurchases on a second trading line, mimeo, HEC-University of Geneva (2004)

Ehrhardt, O., Nowak, E.: The effect of IPOs on German family-owned firms: governance changes,ownership structure, and performance. J. Small Bus. Manag. 41, 222–232 (2003)

Evans, J., Evans, R., Gentry, J.: The decision to repurchase shares: a cash flow story, Working Paper,Curtin University of Technology (2000)


Gebhardt, G.: Announcement effects of financing decisions by German companies: synthesis of anempirical research programme, Working Paper, Goethe-University, Frankfurt (2001)

Gerke, W., Fleischer, J., Langer, M.: Kurseffekte durch Aktienrückkäufe - eine empirischeUntersuchung für den deutschen Kapitalmarkt. In: Börsig, C., Coenenberg, A.G. (eds.) Bewer-tung von Unternehmen: Strategie - Markt - Risiko, Kongress-Dokumentation, 56, pp. 275–304.Deutscher Betriebswirtschafter-Tag 2002, Stuttgart (2003)

Grullon, G., Michaely, R.: Dividends, share repurchases, and the substitution hypothesis. J. Financ.57, 1649–1684 (2002)

Ginglinger, E., L’Her, J.-F.: Ownership structure and open market stock repurchases in France. Eur.J. Financ. 12, 77–94 (2006)

Grullon, G., Ikenberry, R.: What do we know about stock repurchases? J. Appl. Corp. Financ. 13,31–51 (2000)

Harris, T., Lang, M., Möller, H.: The value relevance of German accounting measures: an empiricalanalysis. J. Account. Res. 32, 187–209 (1994)

Ikenberry, D., Vermaelen, T.: The option to repurchase stock. Financ. Manag. 25, 9–24 (1996)Ikenberry, D., Lakonishok, J., Vermaelen, T.: Market underreaction to open-market share repur-

chases. J. Financ. Econ. 39, 181–208 (1995)Ikenberry, D., Lakonishok, J., Vermaelen, T.: Stock repurchases in Canada: performance and stra-

tegic trading. J. Financ. 55, 2373–2397 (2000)Jensen, M.: Agency costs of free cash flow, corporate finance, and takeovers. Am. Econ. Assoc. Pap.

Proc. 76, 323–329 (1986)Joos, P., Lang M.: The effects of accounting diversity: evidence from the European Union.

J. Account. Res. 32, 141–168 (1994)Jung, S.-C., Lee, Y.-G.: Stock repurchases in a developing market: evidence from Korea, Working

Paper. Chonnam National University, University of Houston Victoria (2003)Kim, J., Schremper, R., Varaiya, N.: Survey on open market repurchases regulations: cross-country

examination of the ten largest stock markets. San Diego State University, Mimeo (2004)Kraft, G., Altvater, C.: Die zivilrechtliche, bilanzielle und steuerliche Behandlung des Rückkaufs

eigener Aktien. Neue Zeitsch. Gesellsch. 1, 448–452 (1998)Lamba, A., Ramsay, I.: Share buy-backs: an empirical investigation. Working Paper, The University

of Melbourne (2000)Leuz, C., Wüstemann, J.: The role of accounting in the German financial system. In: Krahnen,

J.P., Schmidt, R.H. (eds.) The German Financial System, pp. 447–478. Oxford University Press,Oxford (2004)

Li, K., McNally, W.: Information signaling or agency conflicts: what explains Canadian open marketshare repurchases? Working Paper, University of British Columbia (1999)

Lie, E., Lie, H.J.: The role of personal taxes in corporate decisions: an empirical analysis of sharerepurchases and dividends. J. Financ. Quant. Anal. 34, 533–552 (1995)

Masulis, R.: Stock repurchase by tender offer: an analysis of the causes of common stock pricechanges. J. Financ. 35, 305–321 (1980)

McNally, W.: Open market stock repurchase signaling. Financ. Manag. 28, 55–67 (1999)Nenova, T.: The value of corporate voting rights and control: a cross-country analysis. J. Financ.

Econ. 68, 325–351 (2003)Netter, J.M., Mitchell, M.: Stock-repurchase announcements and insider transactions after the

October 1987 stock market crash. Financ. Manag. 18, 84–96 (1989)Oswald, D., Young, S.: What role for taxes and regulation? A second look at open market share

repurchase activity in the United Kingdom. Working Paper, London Business School (2002)Pan, C.: Why did firms stop paying dividends around the globe? Working Paper, State University,

Ohio (2005)Rau, P., Vermaelen, T.: Regulation, taxes, and share repurchases in the United Kingdom. J. Bus. 75,

245–282 (2002)Rees, W.: The impact of open market repurchases on UK equity prices. Eur. J. Financ. 2, 353–370

(1996)Schremper, R.: Aktienrückkauf und Kapitalmarkt: eine theoretische und empirische Analyse deut-

scher Aktienrückkaufprogramme. Peter Lang, Frankfurt (2002)Seifert, U., Stehle, R.: Stock performance around share repurchase announcements in Germany.

Working Paper, Humboldt University, Berlin (2003)Shiller, R.J.: Irrational Exuberance. Princeton University Press, Princeton (2000)


Shleifer, A., Vishny, R.W.: Large shareholders and corporate control. J. Polit. Econ. 94, 461–488(1986)

Shoven, J., Simon, L.: Share repurchase and acquisitions: an analysis of which firms participate.NBER Working Paper No. 2243 (1987)

Stephens, C., Weisbach, M.: Actual share re-acquisitions in open-market repurchase programs.J. Financ. 53, 313–333 (1998)

Vermaelen, T.: Common stock repurchases and market signaling: an empirical study. J. Financ.Econ. 9, 139–183 (1981)

Wansley, J., Lane, W., Sarkar, S.: Management’s view on share repurchases and tender offer premi-ums. Financ. Manag. 18, 97–110 (1989)

Zhang, H.: Share repurchases under the Commercial Law 212–2 in Japan: market reaction andactual implementation. Pac. Basin Financ. J. 10, 287–305 (2002)

Author’s Biography

Andreas Hackethal holds the HCI endowed chair forFinancial Services Sales & Distribution at the EuropeanBusiness School in Oestrich-Winkel, Germany, and is adirector of the E-Finance Lab, Frankfurt/Main. He earnedhis doctorate degree and his postdoctoral lecturer qualifi-cation from Johann Wolfgang Goethe-Universität, Frank-furt/Main, and has a ten-year working experience witha large German bank and a global consulting company.His principal fields of research are comparative financialsystems, financial institutions and services, and corporatefinance.

Alexandre Zdantchouk is an Associate at Freitag & Co, anindependent German mergers & acquisitions advisory firm.He focuses on the execution of complex national and interna-tional transactions. Prior to his time at Freitag & Co he heldpositions at the asset management and advisory firm Lazardand the consulting firm Tillinghast-Towers Perrin. He holdsa degree in Business Administration from Johann WolfgangGoethe-Universität, Frankfurt/Main, and a B.A. degree inInternational Economics from Belarus Economics Univer-sity, Minsk.

Documents

Signaling Power of Open Market Share Repurchases in Germany