Bidder Gains and Losses of Firms Involved in Many Acquisitions. Durham Business School University of Durham Durham, DH1 3LB, UK. March 2005 ABSTRACT

Transcription

1 Bdder Gans and Losses of Frms Involved n Many Acqusons Antonos Antonou, * Dmrs Petmezas, and Huanan Zhao Durham Busness School Unversy of Durham Durham, DH1 3LB, UK March 2005 ABSTRACT We examne shareholders wealth effects (both n short- and long-run) of UK frequent bdders acqurng publc, prvate, and/or subsdary targets wh alternatve methods of payment between 1985 and We fnd that, n the short-run, bdders lose when acqurng publc targets and gan when purchasng prvate and subsdary targets. Ths result s robust after controllng for bdder s book-to-market rato (value/glamour), core-ndustry (dversfed/nondversfed), and target orgn (domestc/foregn). Our long-run evdence, however, reveals that acqurers experence sgnfcant wealth loss regardless of the target type acqured ndcatng that markets may overreact at the acquson announcement. As a consequence we argue, n contrary to Fuller et al. (2002), that a fruful concluson on the wealth effect of bdders acqurng prvate and subsdary targets can only be drawn under both short and long run nvestgaton. JEL Classfcaton: G11; G14; G34. Keywords: Mergers & Acqusons, Corporate Takeovers, Frequent Bdders, Method of Payment, Publc/Prvate/Subsdary Targets, Abnormal Returns. We would lke to thank Abhay Abhyankar, George Alexandrds, Phl Arbour, John Doukas, Krshna Paudyal, and semnar partcpants at the Durham Busness School for ther useful comments on prevous drafts of ths paper. * Correspondng author. Tel: +44-(0) ; Fax: +44-(0) Emal address: antonos.antonou@durham.ac.uk

2 1. Introducton The examnaton of shareholders wealth effects (value creaton or destructon) of Mergers and Acqusons (M&A) s one of the most coveted research areas n fnance. To date, a large amount of research has focused on examnng the short-wndow stock returns earned by targets and bdders around merger announcements. The stylzed fact emergng from ths strand of studes s undvded n that target frm shareholders earn a sgnfcant and posve abnormal return n a few days surroundng the takeover announcements, a fndng that s rather unsurprsng gven the hefty premums pad to the targets. Acqurng frms, on the other hand, are found to break even whle the combned enty (target and acqurer) earns a posve abnormal return around the announcement date 2. Gven these fndngs, a smple but nterestng queston arses: Do these observed abnormal returns solely reflect the expectaton of future cash flows resultng from the takeover events? Hetala, Kaplan, and Robnson (2000) argue that acquson announcement reveals not only the value of the acquson self but also the stand-alone value of the bdders, the potental synerges of the combnaton, and possbly the bdder overpayment. Hence, s often mpossble to solate the above effects from the observed abnormal returns. Fuller, Netter, and Stegemoller (2002) apply a sophstcated research desgn to control for (much of) the nformaton about bdder characterstcs contaned n stock returns at the acquson announcement. 3 They nvestgate the returns to US frequent bdders makng fve or more bds whn a three-year tme horzon. As they argue, the sample of frequent bdders 2 For evdence on acqurers short-run stock returns see, for example, Dodd and Ruback (1977), Asquh, Bruner and Mullns (1983), Denns and McConnell (1986), Bradley, Desa, and Km (1988), Franks and Harrs (1989). For evdence of combned frms see, for example, Bradley, Desa, and Km (1988), Mulhern and Boone (2000), Andrade, Mchell, and Stafford (2001). 1

3 allows to hold bdder characterstcs constant when examnng the pattern of announcement returns 4. In general, the authors conclude that bdders experence sgnfcant wealth loss when buyng publc targets, whle earn substantal gans when prvate and subsdary targets are purchased. Ths s, however, a premature concluson as short-run event study conclusons rely strctly on the assumpton of market effcency. Nevertheless, s possble that stock prces temporarly devate from ther fundamental values due to nvestors systematc over or under-reacton to acquson announcements. In such case, serous doubts arse towards short-run wndow s ably to dstngush real economc gans from market neffcency. Accordngly, Healy, Palepu, and Ruback (1992) pos that: From a stock prce perspectve, the antcpaton of real economc gans s observatonally equvalent to market msprcng. Ths vew ndcates that, ndeed, short-run systematc under- or over-reacton to an event has gradually become accepted n the lerature. Fama hmself, the father of the effcent market hypothess, has recently conceded that stock prces could become somewhat rratonal. 5 In a nutshell, the volumnous lerature related to behavoural fnance emphaszes that results generated by short-run event studes need to be nterpreted wh further skeptcsm. We thus beleve that Fuller et al. s (2002) concluson needs to be braced wh certan cauton. In ths case, we argue that a complementary long-run analyss n ths context s consdered essental n order to reach a relatvely thorough nvestgaton of shareholders wealth effects. If the long-run results mrror the short-run fndngs, we can then be more confdent to accept ther short-run conclusons. However, f the short-run evdence s not supported by the long- 3 Fuller et al. (2002) s the frst major attempt n examnng takeover announcement returns of multple bdders nvolved n acqusons of publc, prvate, and subsdary targets wh alternatve methods of payment between 1990 and Fuller et al. (2002, p. 1792) argue Snce we control for acqurer characterstcs n that the same bdder wll often choose to acqure targets wh varyng ownershp status, and wh dfferent payment methods, we can examne the varaton n acqurer returns as a functon of these bd characterstcs. 5 As two economsts debate markets, the tde shfts. Belef n effcent valuaton yelds ground to role of rratonal nvestors Mr. Thaler takes on Mr. Fama. The Wall Street Journal, October 18,

4 run results, we can then cast doubt on whether Fuller et al. s (2002) suggeston s economcally sound and ntuve or merely a potental product of short-run market neffcency. In addon, of course, such fndngs have not been tested n other countes apart from the US. We therefore conduct, n ths paper, a UK study by examnng the stock returns (both n short- and long-run) of frequent bdders that successfully acqured three or more publc, prvate, or subsdary targets usng alternatve methods of payment whn a three-year perod between 1985 and Our comprehensve sample constutes of 4173 UK takeovers takng place over a 20-year perod. A pont that s worth mentonng s that a sgnfcant proporton of UK frms appear to engage n multple acqusons over ths perod (more than 40% of the entre populaton) whle, most mportantly, prvate targets and subsdares are major components of the UK takeover market (approxmately 90%), a fact that very few studes have taken nto account. In general, our results demonstrate that posve abnormal returns are present only n the short-run (.e., the takeover announcements). Bdders gan when buyng prvate or subsdary targets and lose when purchasng publc targets. Ths fndng s fully consstent wh Fuller et al. (2002). In addon, we add further evdence to the short-run study by takng nto account bdder s book-to-market rato (value glamour), core-ndustry (dversfed/non-dversfed), and target orgn (foregn/domestc). On the other hand, our long-run results show that bdders experence sgnfcant losses regardless of the type of target acqured. Ths fndng mples that the stock market may overreact n the short-run and s prces are gradually corrected n the long run. Hence, our evdence rases a bg queston mark towards Fuller et 3

5 al. s (2002) concluson as the short-run economc gans (.e., the reflecton of the acquson synerges) of buyng prvate and subsdary targets cannot be materalzed n the long run. The remander of ths paper s organzed as follows: Secton 2 descrbes the data and the methodology. Sectons 3 and 4 report and dscuss the emprcal fndngs. Secton 5 concludes our analyss. 2. Data and Methodology 2.1. Data We examne a sample of successful takeovers by U.K. publc companes that acqured both domestc and foregn targets, announced between January 1, 1985 and May 6, The sample acqusons are drawn from the Secures Data Corporaton s (SDC) Mergers and Acqusons Database whle the perod selected s drven by the total avalably of the database and the defnon of multple bdder we have set (acqurng 3 targets whn a 3-year perod). 6 The followng crera are used n selectng our fnal sample: 1. Acqurers are U.K. frms publcly traded on the London Stock Exchange (LSE) and have fve days of return data around the takeover announcement and one to three year return data lsted on the DataStream Database. 2. The acqurer completes three or more bds n any three-year wndow durng the sample perod. 3. The bdder acqures at least 50% of the target s shares as a result of the takeover. 6 SDC s a commercal database that ncludes nformaton on U.K. Takeover Bds snce However, the frst multple bdder appears to do the frst bd n

6 4. The target s a publc, prvate, or subsdary frm The deal value s one mllon dollars or more We om fnancal and utly frms (followng Fama and French 1992) for both bdders and targets. We also exclude clustered acqusons where the bdder acqures two or more frms whn fve days n order to solate the overlappng effect among the bds. Our fnal sample conssts of 618 unque acqurers proceedng to 4173 bds. The full sample s then dvded nto three groups based on the method of payment for the acquson,.e., pure cash, pure stock, and combned. The combned payment sub-sample ncludes all acqusons n whch the payment method s neher pure cash nor pure stock. As we use Dmson, Nagel, and Qugley (2003) UK 3-factors to account for UK book-to-market peculares, we nclude n our long run analyss bds carred out between for 3-year analyss (2607 frms), bds up to 1999 for 2-year analyss (2995 frms) and takeovers from (3383 frms) for 1-year analyss respectvely. Table 1 presents the summary statstcs for acqurers makng multple acqusons and ther targets. Panels A, B, C, and D report the annual mean and medan acqurer and target sze for all bds, only publc bds, only prvate bds, and only subsdary bds, respectvely. The mean and medan sze for each acqurer and each target s the frm sze at the year the deal was announced. The acqurer s and publc target s market capalzaton equals the prce per share one-month pror to the bd announcement tmes the number of common shares outstandng For prvate and subsdary targets, the frm sze s measured as the deal value of the bd. The 7 We examne subsdary targets, as they are one of the three man categores of the market for corporate control. All subsdary targets are unlsted companes after checkng the Target Publc Md Code from the SDC database. 5

7 fnal row of each panel presents the mean and medan sze for each unque acqurer and target (.e., counted only once for each frm). Accordngly, for the entre sample n Panel A, the mean (medan) sze of the acqurer s 488 mllon pounds (77 mllon pounds) for 618 unque acqurers, whle for 4173 unque targets the mean (medan) sze s 37 mllon pounds (6 mllon pounds). Table 1 also presents a general upward trend n both merger actvy and sze of acqusons for publc, prvate, and subsdary targets, droppng slghtly by Panels B, C, and D present the dstrbuton of mean and medan sze of frms based on target ownershp status,.e., Publc (Panel B), Prvate (Panel C), and Subsdary (Panel D). Panel B llustrates that the mean (medan) sze s 159 mllon pounds (42 mllon pounds) for 195 unque publc targets. Panel C shows that the prvate targets mean (medan) sze s much smaller than that of publc targets, 15.8 mllon pounds (4.75 mllon pounds) for 2459 unque prvate targets. Panel D reports that the mean (medan) sze of 1519 unque subsdary targets s also smaller (56 mllon pounds (8.7 mllon pounds)) than that of publc targets. In sum, Table 1 shows that the sze of publc acqusons s sgnfcantly greater than prvate and subsdary ones Methodology We calculate Cumulatve Average Resduals (CARs) for the fve-day perod [-2, +2] 10 around the announcement date suppled by SDC. More specfcally, we estmate the abnormal returns by usng a modfed market-adjusted model: 8 We employ a one mllon dollars cut-off pont to avod results beng generated by very small deals. Smlarly, studes lke Fuller, Netter, and Stegemoller (2002), Moeller, Schlngemann, and Stulz (2004) n the US use a cut-off pont of one mllon dollars. 9 Despe the decrease n number of deals after 2000 the total value of transactons has sgnfcantly ncreased. As an ndcaton of the latest data evdence, the total value of takeovers n the frst quarter of 2004 s almost double than of frst quarter of We choose the fve-day perod because Fuller et al. (2002) fnd that a fve-day wndow around the merger announcement gven by SDC s wde enough to capture the frst menton of a merger every tme for a sample of about 500 announcements. 6

8 AR = R R (1) mt where, R s the return on frm [ln ( P t )- ln ( P t 1 )] and R mt s the value-weghed market ndex return (.e., the FT-All Share measured as the frst dfference of the log of the Market Index). We then calculate the CAR for each frm over the 5-day event wndow. The t- statstcs are estmated usng the cross-sectonal varaton of abnormal returns. 11 It s obvous that n our long run analyss a subsequent acquson wll occur whn less than 36 months after a precedng acquson, snce our sample conssts of multple acqurers. We therefore use Calendar Tme Portfolo Regressons (CTPR) to sdestep the problem of aggregatng daly returns to obtan a long-term return, and allow nferences that are not based by cross-sectonal dependence. 12 In each calendar month, a portfolo s formed by ncludng all stocks wh an acquson event durng the past 12, 24, or 36 months. The portfolo s rebalanced every month by ncludng new event frms executed a transacton n the prevous month and droppng the ones whose latest acquson event falls out of the one to three-year holdng perod. The average monthly abnormal return durng the one to threeyear post-event perod s the ntercept from the tme-seres regresson of the calendar portfolo return on the Fama and French three-factor model. The FF 3-factor model are 11 We do not estmate market parameters based on a tme perod before each bd, snce for frequent acqurers, there s a hgh probably that prevous takeover attempts would be ncluded n the estmaton perod, hence makng beta estmatons less meanngful. Addonally, has been shown that for short wndow event studes, weghtng the market return by the frm s beta does not sgnfcantly mprove estmaton. (Brown and Warner, (1980)). 12 Cross-sectonal dependence caused by overlappng observatons leads to downwards-based standard errors and therefore causes t-statstcs to be based upwards. In addon, accordng to Mchell and Stafford (2000), due to the number of frms beng dfferent for each month, heteroskedastc resduals are lkely to be present when regressng calendar tme average portfolo returns n excess of the rsk free rate aganst the factors of an asset-prcng model. Hence, we use Andrews (1991) heteroskedastcy and autocorrelaton consstent standard errors so as to realstcally assess the valdy of our results. 7

9 estmated by usng the UK 3-factor of Dmson et al. s (2001) to account for the UK B/M rato peculares : 13 R pt R = a + β ( R R ) + s SMB + h HML + ε (2) ft mt ft t t where R pt s the average monthly return of the calendar portfolo, R ft s the monthly rsk free return, R mt s the monthly return of the value-weghted market ndex, SMB t the valueweghted return on small frms mnus the value-weghted returns on large frms, and HML t the value-weghted return on hgh book-to-market frms mnus the value-weghted return on low book-to-market frms. In addon, β, s and h are the regresson parameters and ε s the error term. The α (ntercept) s nterpreted as the average of the ndvdual, frm-specfc ntercepts. 3. Emprcal Results for the Short-run Analyss 3.1. Bdder Abnormal Returns by Target Type and Method of Payment In Table 2, Panel A, we present fve-day CARs for the full sample classfed by target publc status and method of payment. For all bds, the CAR s posve (0.74%) and statstcally sgnfcant at 1% sgnfcance level. When focusng on publc targets we obtan a sgnfcant negatve CAR of 1.95%. Ths s consstent wh UK studes of Frth (1980), Draper and Paudyal (1999, 2004), Sudarsanam Holl and Salam (1996) and Sudarsanam and Mahate (2003) who fnd negatve and sgnfcant bdder abnormal returns surroundng the announcement. When we further dfferentate on the bass of method of payment CARs are, surprsngly, all negatve rrespectve of the mode of fnancng used, wh stock payment 13 Dmson, Nagel, and Qugley (2003) use dfferent breakponts to those of Fama-French (1993) to construct Sze and Book-to-Market portfolos manly due to sze and B/M rato beng negatvely correlated n the UK and large frms (small frms) beng concentrated n the low (hgh) BE/ME quartle. 8

10 generatng the largest negatve and hghly sgnfcant CAR of 4.05%. Ths s consstent wh Myers and Majluf (1984) hypothess and Moeller, Schlngemann and Stulz (2004) fndng, suggestng that the greater nformaton asymmetry assocated wh stock payments leads to more negatve performance. 14 For cash payments CARs are stll negatve but margnally sgnfcant. 15 Ths can be attrbuted to hgher offers (premum) for cash exchanges to compensate target shareholders for the mmedate payment of taxes. For prvate targets, CARs are posve (0.73%) and n most cases sgnfcant. Ths s n lne wh Chang (1998) 16 and Ang and Kohers (2001) who document substantal gans for acqusons of prvately held frms. Accordng to the explanatons offered, prvate frms exhb concentrated ownershp, whch leads to less agency conflcts, allevate the publc pressure from outsde nvestors and therefore have the opportuny to avod hubrs-motvated takeovers. The nature of prvate targets self auto-protects the acqurng company from managers empre buldng ncentves, snce such acqusons do not offer, n most cases, the prestge they pursu. Prvate frms confront the problem of lqudy, meanng that they cannot be bought and sold as easly as publc frms. Therefore, n order to create an attractve mage for ther company and a plausble ncentve as a profable nvestment opportuny for potental acqurers, they offer ther shares at a dscount (lqudy). Ths strategy of lqudy 14 Myers and Majluf (1984) argue that the premse of nformaton asymmetry rases the proposon that managers wh prvate nformaton that ther frm s shares are overvalued offer these shares as consderaton n takeover bds. Outsde nvestors, recognzng the adverse selecton problem, consequently revse ther estmate of the offer s value downwards, a plausble explanaton for the negatve performance of stock deals. Moeller et al. (2004) test ths hypothess and fnd that hgher growth uncertanty due to nformatonal asymmetry n the case of equy fnancng leads to more negatve returns. 15 Better performance of bdders used cash fnancng s consstent to the lerature. See for example, Travlos (1987), Fshman (1989) and Martn (1996) who fnd that bdders makng cash offers have greater abnormal returns at the bd announcement than do those makng stock offers. 16 Chang (1998) fnds posve abnormal returns for stock offers to prvate targets and attrbutes ths fndng to the fact that prvately held frms are not oblged to release value relevant nformaton to the publc, thus generatng a hgh cost of obtanng nformaton (nformaton hypothess). Such cost s very lkely to be assocated wh larger returns for acqurng frms snce they capture a greater proporton of the expected gans, partcularly when there are only few frms wh whch the target may reap synergstc gans. It s fnally lkely that posve returns are attrbuted to the lmed competon hypothess that predcts hgh lkelhood of underpayment. 9

11 dscount becomes even more essental due to the lack of aucton-lke atmosphere whn prvate frms, oppose to the aucton-lke nature and, obvously, lqudy of publc frms, enhancng by the presence of rsk arbrageurs. Further, we confrm Fuller et al. (2002) that publc frms acqurng subsdary targets experence the greatest gans (1.09%). Wh respect to medum of exchange, Facco and Masuls (2004) pos that, when a subsdary acquson takes place, cash s preferred snce corporatons sellng subsdares are often motvated by fnancal dstress concerns or a desre to restructure towards ther core competency. Consequently, there s strong preference for cash consderaton n order to realze these fnancal or asset restructurng goals and also due to the fact that bdders are frequently motvated to dvest subsdares to fnance new acqusons or reduce ther tax burden. Such preference for cash payments s lkely to lead to sgnfcant posve returns. 17 A last but not least notceable pont from Table 2 s that both prvate and subsdary targets exhb substantally lower levels of stock fnancng (3.5% and 1.8%, respectvely) approvng the above cash preference explanaton. To argue at ths pont, we note that unlsted companes are very closely held. Consequently, accordng to Martn (1996), snce stockfnanced acqusons typcally reduce the wealth of acqurng frm s shareholders, the lkelhood of acqusons beng fnanced n ths manner should be lower when block-holdngs are hgher. 17 Fuller et al. (2002) however document hgher returns for subsdary targets when stock s used as a method of payment. 10

12 3.2. Bdder Abnormal Returns by Relatve Sze and Method of Payment A very mportant component affectng bdder returns s the relatve sze of target to acqurer. Due to the fact that prvate targets are, on average, much smaller than publc targets we expect the mpact on the bdder of a prvate acquson to be smaller than a publc acquson. Therefore we control for the effect of target sze on bdder returns n order to be able to compare n a relatvely better way publc and prvate takeovers. We use the relatve sze of target to bdder by defnng as target market value (when the target s publc) or the deal value (when the target s a prvate frm or subsdary) dvded by bdder market value. Table 3, Panel A, dsplays the results for the overall sample. We fnd that CARs are n general posvely related wh the relatve sze regardless the method of payment. Accordngly, Asquh et al. (1983), Jensen and Ruback (1983) and Kang (1993) fnd greater abnormal returns for large publc targets n the 1970s. In addon, Fuller et al. (2002) dentfed a smlar pattern to our evdence for a sample of US takeovers. Ths s lnked to the suggeston made by Loderer and Martn (1990), who clam that large frms seem to pay too much for ther targets and large bds seem to be overprced on average, facts that deterorate share prce performance. In Panel B, for publc targets, we fnd CARs are negatvely related wh relatve sze though ths result s manly drven by stock offers. However, ths pattern s reversed for prvate targets and subsdares (Panels C and D). Ang and Kohers (2001) addonally suggest that the acqurng return when bddng a publc target s sgnfcantly smaller than the return when bddng a prvate target. In partcular, the larger the target s relatve to the bdder, the stronger the target s negotatng power and ably to benef from the transacton. 11

13 Alternatvely, bddng frms may fnd more dffcult to ntegrate larger publc targets nto ther busness due to hgher regulatory costs nvolved. Fnally, another plausble explanaton s that there are fundamental dfferences n the dvson of gans and/or synerges between takeovers nvolvng publc and prvate targets, and these dfferences are magnfed the greater the relatve sze of the merger. Accordng to Fuller et al. (2002), ths may be to an extent due to a lqudy effect. Fnally, for all Panels (A, B, C, and D) we observe when the relatve sze s lower (.e., <5% where most of large frms are ncluded) cash offer s the domnant method of payment. Along these lnes Myers and Majluf, (1984) and DeAngelo et al., (1984) argue that the larger the sze of the target frm the more lkely the acqurer to use share fnancng n M&A deals. Ths evdence also collaborates wh Facco and Masuls (2004) who suggest that cash fnancng s more preferable to larger acqurers due to s ease of use and ther better access to debt markets, s ably to avod sgnfcant costs of obtanng shareholder approval of pre-emptve rghts exemptons and stock authorzatons and the hgher regulatory costs of stock offers Abnormal Returns by Book-to-Market Rato and Method of Payment Rau and Vermaelen (1998) suggest that glamour acqurers (.e., wh low book-to-market rato) outperform value ones after a merger rrespectve of the payment method used. 18 In some ways the market fals to understand that past manageral performance s not necessarly a good ndcator of future performance, at least n the case of acqusons. Ths result s n contrast to ther fndngs for the long-run performance of bddng frms. They also report a 18 The man argument s the extrapolaton hypothess that explans the dfferental performance of glamour and value acqurers. Acqurers commandng a hgh market ratng due to ther recent performance and expected future performance (glamour acqurers) may act out of overconfdence or hubrs n makng acqusons. The stocks of such companes may also be overvalued and although the managers may be aware of such overvaluaton, the stock market may be not. 12

14 sgnfcant tendency of glamour acqurers to fnance ther acqusons wh ther own stock 19 and ths tendency s stronger n mergers than n tender offers. 20 In Table 4, Panel A, we report the CARs of glamour acqurers. We fnd a sgnfcant posve CAR 0.87% for all acqurers. When the sample s dvded accordng to method of payment all CARs are posve and statstcally sgnfcant (except for stock payments). For publc, prvate, and subsdary target sub-samples, we obtan the same return pattern reported for the full sample (Table 2). Bdders, on average, lose when buyng publc targets (-2.75%) and gan when buyng prvate (0.92%) and subsdary (1.26%) targets. Takng nto account the alternatve methods of payment does not alter our results. Panel B reports the CARs of value acqurers and ndcates the same return pattern wh Panel A. As a consequence, our fndngs are robust for all book-to-market groups of acqurers, enhancng the full sample evdence (Table 2). Fnally, consstent to Rau and Vermaelen (1998), we document that glamour acqurers sgnfcantly outperform value acqurers rrespectve target type except for publc acqusons Abnormal Returns by Domestc/Foregn Targets and Method of Payment Snce the UK s a leadng player n nternatonal acqusons, the study of UK acqusons abroad appears as an mportant aspect n determnng the overall success of FDI by acquson. 22 In addon, n respect to the mpact on returns of bddng frms that engage n such acqusons, the lerature suggests dfferng performance to domestc acqusons, 19 Consstent to the nformaton asymmetry argument, glamour acqurers tend to have hgh past share prce returns, whle the oppose s true for value acqurers. Hence, seems plausble glamour acqurers to use ther overvalued equy as a method of payment and value acqurers to use cash for the oppose reasons. 20 In our study, we wll not nvestgate the dfferental performance of mergers and tender offers, snce n the UK the vast majory of offers are tender offers. 21 Ths fndng follows Sudarsanam and Mahate (2003) who use a sample of UK publc takeovers and fnd an nverse relatonshp to RV (1998) fndngs n the short-run. 22 Healy and Palepu (1993) note that, over the late 1980s, the UK was the lead acqurng naton n nternatonal acqusons accountng for almost 30 per cent of nternatonal corporate nvestments over that perod. 13

15 although no clear concluson can be drawn concernng the drecton of the results. 23 Doukas and Travlos (1988) argue that acqusons of non-domestc targets serve as a dversfcaton vehcle enablng the expanson of the boundary of the acqurng frm and therefore s better performance. Ths expanson perms the nternalzaton of synerges that would otherwse be lost because of varous market falures. As far as the method of payment s concerned, untl very recently foregn takeovers by UK companes almost unversally nvolved cash, as the targets were frequently unwllng to accept foregn equy (Gaughan, 2002). Table 5 presents the CARs of bdders acqurng domestc (UK) or foregn (non-uk) targets. Panel A reports the results for domestc acqusons, whch mrror the prevous fndng obtaned n the full sample of Table 2. The CARs for publc targets are sgnfcantly negatve (-4.27%) under stock payment and margnally sgnfcant for joned payment. However, CARs are posve and sgnfcant for prvate targets and subsdares regardless of the means of payment. 24 For cross-border acqusons, Panel B vrtually reports the same pattern as Panel A although CARs for publc targets are not sgnfcant. Gven that the sample sze for publc targets s small, s not ntuve to draw fruful conclusons from these results. Overall, results reported n Panel A and B confrm to a major extent the return pattern documented n Table 2. Ths emprcal evdence s consdered crcal snce one could argue that our results are contamnated by the nal selecton of the sample ncludng both domestc and foregn targets. In a nutshell, the general pattern holds even after target orgn s taken nto consderaton. 23 See for example, Doukas and Travlos (1988), Kang (1993), Eun, Kolodny and Scheraga (1996), Fatem and Furtedo (1998), and Goergen and Renneboog (2004). 24 Domestc acqusons outperform foregn ones only for prvate targets, possbly due to more mperfect nformaton n non-domestc deals. The larger profs for stock versus cash payments (4.80% vs 0.69%) n foregn acqusons could be explaned by the attempt of bdders to offset the greater uncertanty connected wh the nformaton problems assocated wh acqurng abroad. Ths comes along wh the fndngs of Goergen and 14

16 3.5. Abnormal Returns by Dversfed/Non-Dversfed Targets and Method of Payment Prevous emprcal evdence suggests that corporate dversfcaton may ndeed affect shareholders wealth. Jensen and Ruback (1983), Bradley, Desa and Km (1988) found that the announcement of a dversfyng acquson was generally assocated wh a small posve mpact on shareholder performance. 25 However, there s a large body n the lerature provdng evdence that dversfcaton may dmnsh shareholders wealth (e.g., Lang and Stulz, 1994; Berger and Ofek, 1995; Servaes, 1996). Doukas and Kan (2004) argue that bdders that acqure unrelated targets experence greater excess cash flow declnes and valuaton dscounts than do bdders nvolved n related acqusons. In addon, Fuller et al. (2002) examne only the dversfcaton wealth effect of a bdder acqurng a subsdary target that s core or non-core-related wh the bddng company. They argue that the reason why a frm sells a subsdary s the gan from the ncreased focus, however, they fnd weak evdence that dversfed frms wll sell subsdares at a dscount relatve to non-dversfed companes. Table 6 reports the results of bdders acqurng publc, prvate and/or subsdary targets that are dversfed or non-dversfed from the bdder s ndustry. A dversfed company s defned as a frm whose 3-dg SIC code s dfferent from that of the target frm. 26 Panel A presents, for dversfyng acqusons, a smlar fndng as the one obtaned from the overall sample n Table 2. The CARs are posve and sgnfcant for the full sample (0.77%) and for prvate targets and subsdares (0.80% and 1.01% respectvely), whle sgnfcantly negatve abnormal returns are experenced for publc targets (-1.32%). Bdders buyng publc targets Renneboog (2004), who mply that the choce of means of payment does not act as a sgnal to the market about the over/undervaluaton of the bdder s equy. 25 For more recent evdence of posve abnomal returns from dversfyng acqusons see: Bllett and Mauer (2000) and Hadlock, Ryngaert and Thomas (2001). 26 Servaes (1996) ponts out that a straghtforward examnaton of the 4-dg SIC codes of the segments of the frm does not necessarly reveal the degree of dversfcaton of the frm. He argues that the use of the 4-dg SIC code would be too wde to dentfy the ndustral structure of the frm. Smlarly, Kahle and Walklng (1996) demonstrate how a 4-dg SIC code frm assgned to a frm mght be msleadng wh regard to the most reasonable 2- or 3-dg classfcatons. 15

17 generate sgnfcant losses regardless of the method of payment used (cash or stock), whle prvate targets earn sgnfcant gans when they purchase by cash. Panel B dsplays our results for non-dversfyng acqusons, whch are relatvely smlar to Panel A. More specfcally, we obtan sgnfcantly posve abnormal returns for the overall sample and for prvate targets and subsdares, and negatve CARs for publc acqusons. Therefore, as a whole, we conclude that our fndngs are robust even after cross-ndustry effect s taken nto consderaton. 4. Emprcal Results for Long-run Analyss Our short-run analyss by usng an exhaustve UK sample unequvocally confrms the general pattern dentfed by Fuller et al (2002) (.e., acqurers gan when buyng prvate targets and subsdares but lose when acqurng a publc targets). Our unque emprcal evdence, after controllng for book-to-market, core-ndustry, and target orgn characterstcs, further enhances these fndngs. In general, the above results lead to a natural concluson that buyng prvate and subsdary frms creates value for acqurng frm shareholders but makes them worse off when acqurng publc targets. However, our major doubts le on the ably of 5- day event study to be an adequate tme nterval for solatng the real economc gans from market msprcng, as both are observatonally equvalent. To testfy the above short-run concluson, we must nvestgate whether the same pattern sustans n the long run. We therefore, n ths paper, take a further step to examne long-run (one to three years) post acquson share prce performance n order to verfy the short-run fndngs. 16

18 4.1. Bdder Abnormal Returns by Target Type and Method of Payment Table 7, Panel A, represents the one-year post acquson average monthly abnormal returns. 27 We obtan sgnfcant and negatve monthly abnormal returns for the entre sample (-0.70%) of 3383 acqusons and three subgroups of publc, prvate and subsdary targets regardless of ther payment method. 28 Ths fndng suggests that acqurers lose, on average, for one year after the acquson rrespectve of whether cash or stock s used as a form of fnancng. Snce we came across strong evdence that bdders generate losses when they purchase publc targets, whle they earn sgnfcant profs when they acqure prvate targets and subsdares n the short-run, we are nterested n nvestgatng whether bdders exhb the same return pattern n the post-event perod. Alternatvely, the one-year negatve monthly average abnormal return for the overall portfolo could be drven solely by publc targets. When examnng whether the short-run wealth gans and losses reman n the long run, we fnd for publc targets, large and sgnfcant negatve monthly abnormal returns for the full sample (-1.50%) and three payment subgroups respectvely. Ths s n general consstent wh the short-run results. However, we document that monthly abnormal returns are all negatve and mostly sgnfcant for both prvate targets and subsdares (-0.55% and 0.70% respectvely) even when alternatve methods of payment are consdered. Ths emprcal evdence, hence, stands n sharp contrast wh the short-run results dentfed, whch ndcate that prvate targets and subsdares gan from acqusons. Our results for the two-year 27 For space purposes we nclude n our dscusson only 1-year share prce performance results, as return patterns for 2 and 3 years respectvely (as shown n Table 6) appear dentcal. 28 For US emprcal evdence on acqurers long run stock returns, see for example: Asquh (1983), Malatesta (1983), Jensen and Ruback (1983), Magenhem and Mueller (1988), Agrawal, Jaffe, and Mandelker (1992), Loderer and Martn (1992), Loughran and Vjh (1997), Rau and Vermaelen (1998), Agrawal and Jaffe (2000), and Meggnson, Morgan, and Nal (2004). For evdence from the UK, see for example: Frth (1980), Franks and Harrs (1989), Kennedy and Lmnack (1996), and Gregory (1997). There are, however, other studes [e.g., Bradley and Jarrell (1988), and Franks, Harrs and Tman (1991)] that do not fnd sgnfcant underperformance n the three years followng the merger. We are aware of very few papers examnng post-acquson performance of prvately held and subsdary frms (Moeller, Schlngemann and Stulz (2004) fnd nsgnfcantly posve 3-year post-acquson abnormal returns for prvate targets and zero abnormal returns for subsdary targets). 17

19 (Panel B) and the three-year (Panel C) post event wndows report vrtually the same pattern wh the one-year fndng (Panel A). As a whole, Table 7 makes clear that, n general, frequent acqurers destroy shareholders value n the long run rrespectve of the target ownershp status and the method of payment used n the transacton Robustness Test We subsequently conduct a robustness test to further evaluate the above evdence. Multple acqurers were nally defned as bdders that acqure three or more publc and/or prvate targets and/or subsdares whn a 3-year perod. Therefore, one could argue that, for example, a 36-month return seres may be determned by nter-effects sourcng from the same bdder acqurng both, publc and prvate or subsdary targets. In other words, the results we obtan for prvate targets or subsdares may be drven by the exstence of bdder returns from publc acqusons. In order to control from the effect of publc targets, we solate a sample of acqurers who bought Only Prvate or Only Subsdary targets and examne ther long run performance. Table 8 reports the one to three year post-acquson monthly average abnormal returns for only prvate and only subsdary subgroups. For only prvate, one to three-year monthly abnormal returns are negatve and statstcally sgnfcant at the 1% level though for stock payments are n general nsgnfcant. For only subsdares, one to three-year monthly abnormal returns are negatve but statstcally nsgnfcant except when stock s used as payment method. However, gven the small sample sze for only subsdares, we are not able to establsh valuable nferences from ths evdence. As a whole, s ndcated that even for the only prvate and subsdary groups, abnormal returns are negatve and mostly sgnfcant. Ths evdence further confrms the fndngs reported n Table 7. 18

20 4.3. Prce Reversals Fnally, we examne whether our results are just a manfestaton of long-term reversals as suggested by Jegadeesh and Tman (1993). In partcular, our fndng that acqurers buyng prvate targets and/or subsdares earn posve abnormal returns surroundng the announcement date but lose n the long run can be attrbuted to short-run persstence followed by long-term reversals. If the frms nvolved n prvate or subsdary acqusons experenced posve returns n the few months pror to the acquson announcement, then the stock prces of these acqurers may be subject to a bref persstence followed by long-term reversals. Frstly, the pre-event (pre-announcement) performance of each bdder acqurng prvate targets and/or subsdares s measured. Specfcally, for each acqurer, we calculate average returns for the sx months precedng the announcement of the acquson. Acqusons of prvate and subsdary frms are ranked accordng to ther pre-event returns and placed nto quntles. Subsequently we focus on acqusons of prvate/subsdary targets that le n the top and bottom quntles of pre-event monthly average returns. As a result we sort our sample nto four categores: ) Prvately-held acqusons that experenced the hghest pre-event returns, ) Prvately-held acqusons that exhbed the lowest pre-event returns, ) Acqusons of subsdary targets that generated the hghest pre-event returns, and v) Acqusons of subsdary targets that exhbed the lowest pre-event returns. Results for ths analyss are dsplayed n Table 9. We observe that acqurers of prvate targets who ganed hgh pre-event returns (5.24% on average) have sgnfcant 1-year postannouncement monthly average abnormal returns of 0.47%. Smlar results are obtaned for 2- and 3-year analyss respectvely. Ths fndng s consstent wh long-term reversal and s 19

21 not possble to determne whether the long-term abnormal performance s solely due to reversals or whether the qualy of the acquson s a contrbutng factor. Notceably however, acqurers of prvate targets who experenced negatve pre-event returns (-1.83%) also do poorly n the long run (-1.06%). The negatve average abnormal returns cannot be attrbuted to long-term reversals of stock returns snce the acqurers had negatve returns pror to the merger announcement. Moreover, bdders of subsdary targets who earned negatve pre-event returns have 1-year average abnormal returns of 1.13%. Ths fndng also cannot be attrbuted to prce reversals. As a consequence, we suggest that our results are not smply a manfestaton of momentum and therefore they are not just capturng long run stock-prce reversals. 5. Concluson Ths paper examnes shareholders wealth effects of UK frequent bdders acqurng publc, prvate, and subsdary targets wh eher cash or stock. In the short-run our fndngs are n lne wh Fuller et al. (2002) and confrm that acqurers experence sgnfcant loss when buyng publc targets and substantal gans when purchasng prvate and subsdary targets. However, gven the fact that short-run event study results can be drven by market msprcng we therefore take a step further to nvestgate whether the above concluson can stand up both n the long run and to a UK sample of acqusons. We nvestgate an exhaustve sample of UK frequent bdders (4173 acqusons between ) and corroborate Fuller et al s (2002) short-run evdence after further controllng for the value/glamour, domestc/foregn, and dversfcaton effect. Nevertheless, our longrun results unambguously ndcate that all frequent acqurers experence wealth losses durng 20

22 the three-year perod after the acquson, rrespectve of the type of target acqured. Ths fndng contrasts sharply to the short-run evdence that acqurers gan when buyng prvate and subsdary targets mplyng a possble market overreacton at the acquson announcement. We therefore beleve s premature to accept Fuller et al. s (2002) concluson based solely on the short-run fndngs. In ths respect, gven the nconsstency between the short- and long-run evdence, we consder that no frm concluson can so far be drawn on whether acqurng prvate and/or subsdary targets creates real economc gans to shareholders or ndeed the short run gans are merely an lluson of market msprcng. 21

23 References Agrawal, A., J. F. Jaffe and G. N. Mandelker (1992), The post-merger performance of acqurng frms: A re-examnaton of an anomaly, Journal of Fnance, Vol. 47, pp Andrade, G., M. Mchell and E. Stafford (2001), New evdence and perspectves on mergers, Journal of Economc Perspectves, Vol. 15, pp Andrews, W. K. Donald (1991), Heteroskedastcy and Autocorrelaton Consstent Covarance Matrx Estmaton, Econometrca, Vol. 59, pp Ang, J, and N. Kohers (2001), The Take-over Market for Prvately Held Companes: the US Experence, Cambrdge Journal of Economcs, Vol. 25, pp Asquh, P. (1983), Merger bds, uncertanty, and stockholder returns, Journal of Fnancal Economcs, Vol. 11, pp , R. Bruner, and D. Mullns (1983), The gans to bddng frms from merger, Journal of Fnancal Economcs, Vol. 11, pp Berger, P. G. and E. Ofek (1995), Dversfcaton s Effect on Frm Value, Journal of Fnancal Economcs, Vol. 37, pp Bllett, M. T. and Davd C. Mauer (2000), Dversfcaton and the Value of Internal Capal Markets: The Case of Trackng Stock, Journal of Bankng and Fnance, Vol. 24, pp Bradley, M. and G. A. Jarrell (1988), Comment, n J. C. Coffee, L. Lowensten, and S. Rose- Ackerman, eds., Knghts, Raders and Targets: The Impact of the Hostle Takeover, Oxford Unversy Press., A. Desa and E. H. Km (1988), Synergstc gans from corporate acqusons and ther dvson between the stockholders of target and acqurng frms, Journal of Fnancal Economcs, Vol. 21, pp Brown, S. J., and J. B. Warner (1980), Measurng Secury Prce Performance, Journal of Fnancal Economcs, Vol. 8, pp Chang, S. (1998), Takeovers of Prvately Held Targets, Methods of Payment, and Bdder Returns, Journal of Fnance, Vol. 53, pp DeAngelo, H., L. DeAngelo and E. M. Rce (1984), Gong Prvate: Mnory Freezeouts and Stockholder Wealth, Journal of Law and Economcs, Vol. 27, pp Denns, D. K. and J. J. McConnell (1986), Corporate mergers and secury returns, Journal of Fnancal Economcs, Vol. 16, pp Dmson E, S. Nagel and G. Qugley (2003), Capturng the Value Premum n the U.K , Fnancal Analysts Journal, Vol. 59, pp Doukas, J. and N. G. Travlos (1988), The Effect of Corporate Multnatonalsm on Shareholders Wealth: Evdence from Internatonal Acqusons, Journal of Fnance Vol. 43, pp and O. B. Kan (2004), Excess Cash Flows and Dversfcaton Dscount, Fnancal Management, (Summer 2004), pp Draper, P. and K. Paudyal (1999), Corporate Takeovers: Mode of Payment, Returns and Tradng Actvy, Journal of Busness, Fnance & Accountng, Vol. 26, pp (2004), Acqusons: Prvate versus Publc, European Fnancal Management Journal, forthcomng. Eun, C. S., R. Kolodny and C. Scheraga (1996), Cross-border Acqusons and Shareholder Wealth: Tests of the Synergy and Internatonalsaton Hypothess, Journal of Bankng and Fnance, Vol. 20, pp Facco, M. and R. W. Masuls (2004), The Choce of Payment Method n European Mergers and Acqusons, Journal of Fnance, forthcomng. 22

24 Fama, E. F. and K. French (1992), The cross-secton of expected stock returns, Journal of Fnance, Vol. 47, pp (1993), Common rsk factors n the returns on stocks and bonds, Journal of Fnancal Economcs, Vol. 33, pp Fatem, A. and E. P. Furtedo (1988), An Emprcal Investgaton of the Wealth Effects of Foregn Acqusons n S. Kour and A. Ghosh (eds) Recent Developments n Internatonal Bankng and Fnance, No. 2, Lexngton MA, Lexngton Books. Frth, M. (1980), Takeovers: Shareholder Returns and the Theory of the Frm, Quarterly Journal of Economcs, Vol. 94, pp Fshman, M. (1989), Preemtve bddng and the role of the medum of exchange n acqusons, Journal of Fnance, Vol. 44, pp Franks, J. R. and R. S. Harrs (1989), Shareholder wealth effects of corporate takeovers: the U.K. experence , Journal of Fnancal Economcs, Vol. 23, pp and S. Tman (1991), The postmerger share-prce performance of acqurng frms, Journal of Fnancal Economcs, Vol. 29, pp Fuller, K., J. Netter and M. Stegemoller (2002), What Do Returns to Acqurng Frms Tell Us? Evdence from Frms that Make Many Acqusons, Journal of Fnance, Vol. 57, pp Gaughan, P. A. (2002), Mergers, Acqusons, and Corporate Restructurngs, 3 rd Edon, John Wley and Sons, New York. Goergen, M. and L. Renneboog (2004), Shareholder Wealth Effects of European Domestc and Cross-border Takeover Bds, European Fnancal Management, Vol. 10, pp Gregory, A. (1997), An Examnaton of the Long Run Performance of UK Acqurng Frms, Journal of Busness Fnance & Accountng, Vol. 24, pp Hadlock, C., M. Ryngaert and S. Thomas (2001), Corporate Structure and Equy Offerngs: Are there Benefs to Dversfcaton?, Journal of Busness, Vol. 74, pp Hansen, R. S. (1987), A Theory for the Choce of Exchange Medum n Mergers and Acqusons, Journal of Busness, Vol. 60, pp Healy, P. M. and K. G. Palepu (1993), Internatonal Corporate Equy Acqusons: Who, Where and Why? n K. A. Froot (ed.) Foregn Drect Investment, Unversy of Chcago Press, Chcago. and R. S. Ruback (1992), Does corporate performance mprove after mergers? Journal of Fnancal Economcs, Vol. 31, pp Hetala, P., S. Kaplan and D. Robnson (2000), What s the prce of hubrs? Usng takeover battles to nfer overpayments and synerges, Workng Paper, INSEAD and Unversy of Chcago. Jegadeesh, N. and S. Tman (1993), Returns to buyng wnners and sellng losers, Journal of Fnance, Vol. 48, pp Jensen, M. C. and R. Ruback (1983), The Market for Corporate Control: The Scentfc Evdence, Journal of Fnancal Economcs, Vol. 11, pp Kang, J. (1993), The Internatonal Market for Corporate Control - Mergers and Acqusons of US Frms by Japanese Frms, Journal of Fnancal Economcs, Vol. 34, pp Kennedy, V. A. and R. J. Lmmack (1996), Takeover actvy, CEO turnover, and the market for corporate control, Journal of Busness Fnance & Accountng, Vol. 23, pp Lang, L. H. P. and R. Stulz (1994), Tobn s Q, Corporate Dversfcaton and Frm Performance, Journal of Polcal Economy, Vol. 102, pp Loderer, C. and K. Martn (1990), Corporate Acqusons by Lsted Frms: The Experence of a Comprehensve Sample, Fnancal Management, (Wnter), pp (1992), Post acquson performance of acqurng frms, Fnancal Management, Vol. 21, pp

25 Loughran, T. and A. M. Vjh (1997), Do Long-Term Shareholders Benef from Corporate Acqusons, Journal of Fnance, Vol. 52, pp Magenhem, E. B. and D. C. Mueller (1988), Are acqurng frm shareholders better off after an acquson, n J. C. Coffee, Jr., L. Lowensten, and S. Rose-Ackerman, edors, Knghts, Raders and Targets: The Impact of the Hostle Takeover, New York, Oxford Unversy Press. Malatesta, P. H. (1983), The wealth effect of merger actvy and the objectve functons of mergng frms, Journal of Fnancal Economcs, Vol. 11, pp Martn, K. J. (1996), The method of payment n corporate acqusons, nvestment opportunes, and management ownershp, Journal of Fnance, Vol. 51, pp Meggnson, W. L., A. Morgan and L. Nal (2004), The determnants of posve long-term performance n strategc mergers: corporate focus and cash, Journal of Bankng and Fnance, Vol. 28, pp Mchell, L. M. and E. Stafford (2000), Manageral Decsons and Long-Term Stock Prce Performance, Journal of Busness, Vol. 73, pp Moeller, S. B, F. P. Schlngemann and R. M. Stulz (2004), Frm sze and the gans from acqusons, Journal of Fnancal Economcs, forthcomng. (2004), Do Acqurers wh Uncertan Growth Prospects Gan Less from Acqusons?, Workng Paper (Fsher School of Busness). Mulhern, J. H. and A. L. Boone (2000), Comparng acqusons and dvestures, Journal of Corporate Fnance, Vol. 6, pp Myers, S. C. and N. S. Majluf (1984), Corporate Fnancng and Investment Decsons when Frms have Informaton that Investors Do Not Have, Journal of Fnancal Economcs, Vol. 13, pp Rau, P. R. and T. Vermaelen (1998), Glamour, Value and the Post Acquson Performance of Acqurng Frms, Journal of Fnancal Economcs, Vol. 49, pp Servaes, H. (1996), The Value of Dversfcaton Durng the Conglomerate Merger Wave, Journal of Fnance, Vol. 51, pp Sudarsanam, P. S., P. Holl and A. Salam (1996), Shareholder Wealth Gans n Mergers: Effect of Synergy and Ownershp Structure, Journal of Busness, Fnance & Accountng, Vol. 23, pp Sudarsanam, P. S. and A. A. Mahate (2003), Glamour Acqurers, Method of Payment and Post-Acquson Performance: The UK Evdence, Journal of Busness, Fnance & Accountng, Vol. 30, pp Travlos, G. N. (1987), Corporate Takeover Bds, Methods of Payment, and Bddng Frms Stock Returns, Journal of Fnance, Vol. 42, pp UNCTAD (2000), World Investment Report 2000: Cross-Border Mergers and Acqusons and Development, UNCTAD, Geneva. 24