Is the market underreacting or overreacting to open market share repurchases? A UK perspective

Research in International Business and Finance 26 (2012) 26– 46

Contents lists available at ScienceDirect

Research in International Businessand Finance

journal homepage: www.elsevier.com/locate/r ibaf

Is the market underreacting or overreacting to open marketshare repurchases? A UK perspective

Ian Crawforda,∗, Zhiqi Wangb,1

a School of Management, University of Bath, Bath, Claverton Down, BA2 7JX, UKb Department of Management & Organisational Psychology, Faculty of Business School, Birkbeck, University of London, London, UK

a r t i c l e i n f o

Article history:Received 26 January 2011Received in revised form 17 May 2011Accepted 17 May 2011

Available online 26 May 2011

JEL classification:G14G35

Keywords:Open market share repurchasesMarket underreaction or overreactionLong run abnormal returns

a b s t r a c t

Using UK open market repurchases, we reject the market under-reaction hypothesis and the market overreaction hypothesisproposed by Ikenberry et al. (1995) and Peyer and Vermaelen(2009), respectively. The evidence suggests that the UK marketreacts slowly to actual repurchases made by value firms. UK repur-chases on average do not suffer from share undervaluation priorto the announcement. Value firms perform just as well as glam-our firms during the authorisation period but outperform glamourfirms significantly 2 years following the announcement. It turnsout that value firms repurchase over 6% more shares than glamourfirms during the authorisation period.

© 2011 Elsevier B.V. All rights reserved.

1. Introduction

Share performance of US open market share repurchases has been extensively studied and twomain explanations are offered to explain the seemingly die-hard existence of abnormal returns fol-lowing the announcement. Ikenberry et al. (1995) propose the market underreaction hypothesis whichsuggests that the market treats repurchase announcements with scepticism, leading prices to adjustslowly over time (4 years). Some years late, Peyer and Vermaelen (2009) find evidence consistentwith the overreaction hypothesis which suggests that open market share repurchases are a responseto a market overreaction to bad news prior to the repurchase. Long run abnormal returns are in

∗ Corresponding author. Tel.: +44 012 2538 4399; fax: +44 012 2538 6473.E-mail addresses: [email protected] (I. Crawford), [email protected] (Z. Wang).

1 Tel.: +44 020 7631 6467.

0275-5319/$ – see front matter © 2011 Elsevier B.V. All rights reserved.doi:10.1016/j.ribaf.2011.05.002

dx.doi.org/10.1016/j.ribaf.2011.05.002

http://www.sciencedirect.com/science/journal/02755319

http://www.elsevier.com/locate/ribaf

mailto:[email protected]

mailto:[email protected]

dx.doi.org/10.1016/j.ribaf.2011.05.002

I. Crawford, Z. Wang / Research in International Business and Finance 26 (2012) 26– 46 27

particular persistent among value firms (high book-to-market ratios) and are strongly related to pre-announcement share undervaluation. Will this phenomenon exist in the UK where characteristics ofand regulations governing open market repurchases differ significantly from the US?

Despite the similarities in law and language between US and UK markets, UK open market sharerepurchases have significantly different characteristics and are governed by different law and report-ing regulations. According to the listing rules of the London Stock Exchange, listed firms must publishactual repurchases on the regulatory news services (RNS) which are viewable to the market andinvestors the next business day. In addition, the UK company law (Companies Act 1985) required firmsto report actual repurchases (the amount and quantity of share repurchases) in Annual Reports. Basedon the repurchase information provided by Annual Reports and the RNS, the market and investors caneasily gauge the commitment level of managers towards announced repurchase programmes. UnlikeUS repurchases, UK repurchases have an authorisational duration between two consecutive AnnualGeneral Meetings (usually 12 months). Thus, long run share performance is subject to less influencefrom other corporate events or economic situations which have taken place following the announce-ment. As a result, UK open market share repurchases present an excellent opportunity to understandthe market reaction to the announcement outside of the US markets.

The aims of this paper are not to challenge the market efficiency theory or to investigate the moti-vations of repurchases, but to answer the following questions. First, are open market repurchasesannounced by UK firms on average preceded by share undervaluation at the announcement? Second,is the initial market reaction of UK open market repurchases a response by the market to correctits overreaction to bad news prior to the announcement? Third, are open market share repurchasesattracting a positive and significant return following the announcement? In particular, how does book-to-market affect long run abnormal returns? Finally, would the impact of book-to-market on long runabnormal returns suggest market underreaction to repurchase news or overreaction to bad news?

Relying on a sample of 468 repurchases announced between January 1999 and December 2004,we examine share performance surrounding and following the announcement with considerationof characteristics of firms and repurchases. We first consider short-term returns surrounding theannouncement. The average market return of sample firms prior to the announcement is staticallypositive and reveals limited evidence of share undervaluation. The average market return surroundingthe announcement is positive and statistically significant, consistent with the literature. Further, whenthe sample is segmented by size quartiles and book-to-market quintiles, small firms and glamour firms(low book-to-market ratios) appear to earn higher 5-day initial market returns than big firms and valuefirms. However, announcement abnormal returns of small firms can hardly compensate the shareundervaluation these firms suffer before the announcement. Firms in other size quartiles and book-to-market quintiles show no sign of pre-announcement share undervaluation. Finally, the regressionanalysis shows that the initial market return is unrelated to characteristics of firms and repurchasessuch as size, book-to-market and the percentage of repurchase. We find no evidence to suggest thatthe initial market reaction is a response to share undervaluation prior to the announcement.

We then investigate the long run share performance. Long run abnormal returns can be influencedheavily by different measurement methods (Barber and Lyon, 1997; Lyon et al., 1999; Kothari andJerold, 1997; Mitchell and Stafford, 2000). Here, two different methods are used to determine whetherour sample firms on the whole exhibit abnormal share performance 1 and 2 years following theannouncement. It is evident that our sample firms outperform matched size and book-to-marketportfolios in the long run using any of these two measurement methods. One year buy-and-holdabnormal returns of value firms are only slightly higher than these of glamour firms. Further analysisreveals that the relationship between book-to-market and 1-year buy-and-hold abnormal returnsdisappears after controlling size, percentage of shares sought, actual repurchases and other factors.On the other hand, book-to-market is a dominant factor in determining 2-year buy-and-hold abnormalreturns. The average 2-year buy-and-hold abnormal return of value firms is 15.86% (p = 0.004) while it is6.74% for glamour firms (p = 0.154). Firm size does not have significant influence on long run abnormalreturns in spite of the fact that small firms suffer from significant share undervaluation prior to theannouncement. The results suggest that long run abnormal returns can hardly be explained by themarket overreaction hypothesis or the market underreaction hypothesis. We also notice that actualrepurchases during the authorisation period are not influenced by share price movement. However, the

28 I. Crawford, Z. Wang / Research in International Business and Finance 26 (2012) 26– 46

analysis of actual repurchases shows that large firms repurchase over 4% more shares than small firmswhile value firms actually repurchase 6% more shares than glamour firms during the authorisationperiod. Evidently, share undervaluation is not the motivation for high actual repurchase activity amonglarge and value firms. Moreover, long run abnormal returns are not affected by the percentage of sharesought at the announcement. These results suggest that long run abnormal returns among value firmsare likely to be the market’s delayed reaction to actual repurchases.

We find that share performance of UK repurchases is similar in many ways to that of US repur-chases. For example, UK repurchases generate positive initial market returns in the short-term andhave outperformed their matched size and book-to-market portfolios in the long-term. In particular,long run abnormal returns are significantly high among UK value firms. On the other hand, we noticeseveral differences in share performance between US and UK repurchases. In the short-term, we areunable to find the evidence to suggest that UK repurchase on average suffer from share undervalua-tion prior to the announcement. There is no evidence to indicate that the market reacts excessivelyto bad news prior to the announcement. In the long run, abnormal returns among repurchases areonly related to book-to-market. It seems that actual repurchases conducted by value firms promotesignificantly high abnormal returns 2 years following the announcement. More critically, we are thefirst to examine the statistical impact of different announcement dates on share performance. Previouspapers recognise that UK repurchases do not have a US style of announcement dates, but the impact ofdifferent definitions of announcement dates on share performance has never been examined (Rau andVermaelen, 2002; Oswald and Young, 2004). Our results reveal that different definitions of announce-ment dates do not have significant influence on share performance. Our final result is surprising. Weinterpret the relationship between book-to-market and abnormal returns as an indication of marketunderreaction to actual repurchases by value firms which are different from the results of Ikenberryet al. (1995) and Peyer and Vermaelen (2009).

The findings of this work provide an alternative explanation to the findings reported by prior stud-ies using American, Canadian and Hong Kong data. Previous papers largely explain long run abnormalreturns among value firms as evidence to support the share undervaluation hypothesis or the marketdelayed reaction to the announcement. The main problem in using open market share repurchases tosignal share undervaluation is that open market share repurchases are not firm commitments and thatmanagers have discrete power to implement or discard the repurchase programme and the marketand shareholders are fully aware of the flexible nature of repurchase programmes. If managers feel sostrongly about share undervaluation, there are plenty of other costly payout signals such as dividendincreases and fixed price tender offers which can be adopted by managers to make an immediate andstrong statement regarding perceived share undervaluation. It is also at odds to relate long run abnor-mal returns to a corporate event, a share repurchase programme, which was announced 3 or 4 yearsearlier, not to mention the fact that there must be plenty of interim and final results announcementsmade during the 3 or 4 year period following repurchase announcements. These papers also seem toignore the possibility that long run abnormal returns are the market reaction to actual repurchasesmade by these firms. Our findings link long run abnormal returns earned by value firms with actualrepurchases.

In addition, Von Eije and Megginson (2008) reveal that UK firms are the largest and most prolificshare repurchasers in Europe and share repurchases in the UK have been triggered by privatizationsduring the 1980s. Open market share repurchases were first legalised in the UK in 1981 while mostEuropean countries prohibited share repurchases until 1998 (Lasfer, 2001). Even though many Euro-pean countries have civil law systems compared with the common law system in the UK and Ireland,the findings of UK papers would be helpful to law makers in European countries. The RNS news pub-lished on the London Stock Exchange are open to the market and shareholders which would help themmake sell or non-sell decisions when companies offer to buy back their shares. If shareholders in thelong run would benefit from holding shares in firms which purchased all or part of shares as promised,shareholders and the market would place a great emphasis on actual repurchase reporting which canbe viewed by the public. Recently, it has become relatively common for US firms to report their actualrepurchases in their IQ-Qs and 10-Ks, though the exact timing remains unreported (Zhang, 2005). Itis clear evidence that a good practice in the UK can spread to other countries. Finally, since a lot ofEnglish speaking countries and regions around the world have been significantly influenced by English


corporate law and stock trading regulations, the findings of this paper would motivate researchers tolook into the relationship between actual share repurchases and share performance of repurchaseannouncements in these countries.

This study is organised as follows. Section 2 reviews the literature on share repurchases. Sec-tion 3 describes the data, share performance measurement, significance tests and variables. Section 4presents empirical results and Section 5 concludes.

2. Review of the literature

Prior studies provide a wide range of theories, such as the share undervaluation hypothesis (Dann,1981; Vermaelen, 1981; Comment and Jarrell, 1991; Ikenberry et al., 1995), the free cash flow hypoth-esis (Jensen, 1986; Grullon and Michaely, 2004), the dividend substitution hypothesis (Jagannathanet al., 2000; Grullon and Michaely, 2002; Brav et al., 2005; Skinner, 2008), the personal taxation sav-ing hypothesis (Lie and Lie, 1999), the takeover deterrence hypothesis (Stulz, 1988; Bagwell, 1991;Billett and Xue, 2007) and the capital structure hypothesis (Hovakimian et al., 2001; Hovakimian,2004; Grullon and Michaely, 2004), to explain the average positive announcement return surround-ing repurchase announcements in the United States. The two most popular and dominant explanationsare the share undervaluation hypothesis and the free cash flow hypothesis. Dann (1981), Vermaelen(1981) and Hertzel and Jain (1991) suggest that firms use repurchase announcements to signal to themarket that their shares are undervalued. On the other hand, Jagannathan and Stephens (2003) andGrullon and Michaely (2004) find evidence to support open market share repurchases being used todistribute excess cash.

Regarding long run share performance following repurchases, the share undervaluation hypothesisseems to be more plausible than the free cash flow hypothesis because actual repurchase behaviourof repurchasing firms indicates that shares are most likely to be repurchased when share prices havedeclined following repurchase announcements (Stephens and Weisbach, 1998; Ikenberry et al., 2000).Based on the share undervaluation hypothesis, two additional hypotheses are proposed by Ikenberryet al. (1995) and Peyer and Vermaelen (2009) to explain long run abnormal returns earned by repur-chasing firms. The market underreaction hypothesis suggests that the market’s initial reaction to therepurchase announcement is incomplete and it takes several years for the market to finally realiseeconomic benefits of share repurchase programmes (Ikenberry et al., 1995, 2000; Chan et al., 2004).Peyer and Vermaelen (2009) attribute the abnormal share performance of US repurchases to the mar-ket’s overreaction to bad news prior to repurchase news (called the market overreaction hypothesis).Similar findings are reported by Zhang (2005), employing a sample of 3,628 repurchase executionannouncements made by 135 Hong Kong firms between September 1993 and August 1997. Zhang(2005) finds that the Hong Kong market responds more favourably to the repurchase execution newsmade by small and value (high book-to-market) firms. He describes this finding as the evidence of shareunderpricing. It is not surprising because regardless of why firms repurchase shares, managers wouldonly repurchase shares when shares are properly priced or undervalued. If share prices are quicklyreadjusted to the level which is above the expectations of managers upon repurchase announcements,it is unlikely that managers would be motivated to carry out the actual repurchase programmes.

Within the UK context, there are only a handful of papers which investigate the market reaction torepurchases announcements in the short term and the long term. Rees (1996) is the first to examine themarket reaction to 882 actual repurchase news made by UK firms between 1981 and 1990. He reportsa small 0.25% abnormal return in the 5 days surrounding the actual repurchase news. Of course, hisresult is not compatible with an average 3% announcement abnormal return reported by US studiesbecause US studies examine the initial market response to repurchase announcements not to actualrepurchase news releases. Rau and Vermaelen (2002) are the first to study the initial market reactionand the long run returns of 264 share repurchase announcements made by UK firms between January1985 and June 1998. Their sample includes open market share repurchases, private repurchases andtender offers reported by Security Data Corporation (SDC) in the sample period. They find a statisticallysignificant average abnormal return of 1.14% in the 11-day window surrounding the announcement,which they consider to be significantly smaller than the initial market return reported by US studies.In the long run, UK firms announcing open market share repurchases earn insignificant and negative


abnormal returns of −2.01% in the year following the repurchase announcement. They conclude thatUK firms are not repurchasing shares to take advantage of undervalued share prices.

The results of Rau and Vermaelen (2002) are questioned by Oswald and Young (2004). Oswaldand Young (2004) examine repurchase announcements made by UK firms between January 1995 andDecember 2000 using several databases, including the SDC, the Financial Times and the London StockExchange Regulatory News Service. Their findings suggest that the results of Rau and Vermaelen (2002)might be caused by sampling biases in the SDC data. In addition, they report statistically significantnegative (positive) abnormal returns measured over the 1 year period prior (subsequent) to repur-chase announcements. Actual repurchase behaviour of UK managers seems to suggest that like theirUS counterparts, UK managers buy more shares when prices fall. They interpret these findings as beingdue to the fact that UK share repurchase announcements are partly motivated by share undervaluation.However, their conclusions are not always supported by the findings in other papers. Lee et al. (2010)report that unlike repurchases in Italy and Germany, open market share repurchases announced by UKfirms appear to earn smaller and insignificant announcement abnormal returns upon the announce-ment. They conclude that special UK characteristics such as tax treatment are the reason for such alow average announcement return and a high and statistically significant pre-announcement return.An empirical survey of the motivation of UK share repurchases finds that UK share repurchases areprimarily used to achieve an optimal capital structure (Dixon et al., 2008).

The further investigation of UK actual repurchases reveals evidence to support the free cash flowhypothesis. Wang et al. (2009) examine the market reaction to the announcement of UK actual sharerepurchases between September 1997 and July 2003. Using a sample of 9,020 buyback announcements,they find abnormal returns upon buyback announcements and abnormal returns are related to firmswhich suffer from overinvestment problems and are unrelated to signalling costs. They explain thatthe findings largely support the view that share repurchases reduce the agency costs of excess freecash flow. By collecting UK actual repurchase data between January 1995 and April 2003, Oswald andYoung (2008) reveal that share repurchases can alleviate the agency costs of free cash flow and bettermanagerial incentive alignment and closer monitoring by external shareholders are important factorsstimulating such payouts.

The present work is motivated by the lack of an examination of long run abnormal returns 2 yearsfollowing repurchase announcements in prior studies regarding UK share repurchases. That is vitalin the sense that US, Canadian and Hong Kong studies all find long run abnormal returns a few yearsfollowing repurchase announcements, which include the whole authorisation period. In particular,Peyer and Vermaelen (2009) mention that most anomalies disappeared after being reported by aca-demic studies while repurchase anomalies have not gone away-long run abnormal returns followingopen market share repurchase announcements are still large and significant especially among valuefirms. It is difficult for US studies to collect actual repurchase data since US firms are not requiredby the regulations to report their actual buyback activity. When actual repurchases can be traced,Ikenberry et al. (2000) show that value firms tend to purchase more shares than glamour firms. Toour knowledge, no UK studies investigate the relationship between value firms and the level of actualrepurchases. In addition, extant papers reveal that UK repurchases are different from repurchases inthe US and some of the European countries. First, the average announcement abnormal return of UKrepurchases is smaller than that of repurchases in the US, Italy and Germany. Second, several stud-ies find that UK repurchases attract a positively significant announcement abnormal return prior torepurchase announcements (Lasfer, 2001; Lee et al., 2010). Thus, the examination of long run shareperformance of UK repurchases is a must.

Previous papers using American, Canadian and Hong Kong data largely explain long run abnormalreturns among value firms as the evidence to support the share undervaluation hypothesis or themarket delayed reaction to the announcement. The difficulties in this explanation are obvious. Openmarket repurchases are not always followed by actual repurchases (Stephens and Weisbach, 1998;Jagannathan et al., 2000; Lie, 2005). It is rather unreasonable to relate long run abnormal returns toopen market share repurchase programmes, which were announced 3 or 4 years before, just to considerhow many corporate events would have taken place during such a long time period. Moreover, thereare many other payout options, such as dividend increases and fixed price tender offers, which can beused by managers to lift up share prices more quickly upon the announcement than open market share


repurchases. Thus, we propose that small announcement abnormal returns earned by UK repurchasesare not the market reaction to perceived share undervaluation beforehand, but the market responsetowards the propensity of actual repurchases. Since actual repurchases can happen any time duringthe 12-month period following repurchase announcements, the market would have to wait for a whileto readjust share prices based on the level of actual repurchase commitment, resulting in a delay inthe market’s reaction to actual repurchases. In other words, long run abnormal returns are influencedby the level of actual repurchases rather than repurchase announcements which are a mere promisenot a firm commitment to buy back shares.

3. Data, performance measurement and variables

3.1. Data

As there is no UK equivalent of the Wall Street Journal Index, our sample is collected from variousresources, such as the Financial Times, FAME, Regulatory News Service (RNS) provided by Hemscottand companies’ annual reports. The use of various resources ensures that the sample includes butdoes not overestimate open market share repurchases announced in the time period. First, all repur-chase announcements published by the Financial Times for the sample period are collected by usingkeywords like “share buybacks” and “share buy-back”, while repurchase announcements made byclose-fund investment trusts as well as announcements of tender offers and preferred share repur-chases are excluded from the sample. This process yields 219 open-market share announcements.Second, as the Financial Times focuses on the medium to large companies, the FAME database is usedto supplement the sample. FAME keeps all stock data of all UK publicly quoted firms (including liveand dead firms). A search of the FAME UK publicly listed companies’ stock data yields 317 cases whosereported year-end outstanding shares decreased at least once on a yearly basis between 1999 and2005. Third, companies’ annual reports and the RNS (Hemscott) are used to check the nature of theserepurchases. With the aid of companies’ annual reports, 162 out of 317 firms announcements are foundto be unrelated to open market share repurchases, while the rest of the firms announced 213 openmarket repurchases during the sample period. 36 more open market share repurchase announcementsare added to the sample after checking these companies’ RNS (Hemscott). A direct search of open mar-ket share repurchases through the RNS could result in overestimation of sample size. As required bythe Listing Rules of the London Stock Exchange, UK companies that repurchase shares in the marketneed to provide the news containing the number of shares repurchased and the prices paid for sharesthe following day to the London Stock Exchange, which then publishes the repurchase information onthe RNS. British Petroleum (BP), for example, could have hundreds of share repurchase news items onthe RNS in a single financial year, but as a matter of fact all of these are executed under one authori-sation right. Therefore, unlike in Oswald and Young (2004) the RNS is not used here directly to searchfor open market repurchase news. Our initial sample consists of 468 open market share repurchases.Panel A of Table 1 shows the number of repurchase announcements collected from each of the abovesources.

Using Australian open market share repurchases announced between 1996 and 2009, Farrugiaet al. (2011) find evidence to suggest that the number of repurchase programmes in a HIGH economicstate is statistically significantly higher than that in both LOW and MEDIUM economics states. Theydefine these economic states using the Australian seasonally adjusted quarterly industrial productionindex. To determine the economic conditions for the sample period, we use the British seasonallyadjusted quarterly and monthly industrial production index (2006 = 100) reported by the Office forNational Statistics to define the trend. The quarterly and monthly adjusted production indices arefirst logarithm transferred, and then regressed on a constant and a time trend from January 1999 toDecember 2004. The estimated constants are 2.015 (quarterly) and 2.014 (monthly) and the slopesare −0.00055 (quarterly) and −0.00018 (monthly). The constants here are very different from 0.013reported by Farrugia et al. (2011). It is possible that they use a longer sample period (13 years) whileour sample period is 6 years. Thus, we do not follow their method to develop the trend for the naturallog of the quarterly or monthly UK industrial production since the regression results suggest that thereare no up and/or down market cycles, but a static economic condition, a rather straight line (see Fig. 1).

32I.

Crawford,

Z. W

ang /

Research

in International

Business and

Finance 26 (2012) 26– 46

Table 1Descriptive statistics for open market share repurchases between January 1999 and December 2004.

Keywords or sections search Raw results Repurchases

Panel AThe FTs Share buy-back or share buybacks 1,163 219FAME Report sections: stock Data-annual stock data 317 213RNS Select companies: AGM/EGM; share buybacks; repurchase of own shares 189 36

Total 1,669 468

Year N Size quartiles Book-to-market quintiles Missing Negative

Small Large Lowest Highest BTMV BTMVSize 1 Size 2 Size 3 Size 4 BTMV 1 BTMV 2 BTMV 3 BTMV 4 BTMV 5

Panel B1999 65 4 9 19 29 5 5 14 10 27 2 22000 73 6 8 18 35 2 2 11 18 34 4 22001 69 3 15 19 30 4 5 13 17 28 1 12002 93 6 16 24 41 11 12 14 23 27 2 42003 80 4 14 11 49 13 10 13 14 28 0 22004 88 3 10 11 57 22 15 10 17 17 2 5

Total 468 26 72 102 241 57 49 75 99 161 11 13

Panel A summarizes the number of repurchases collected from the Financial Times (FTs), FAME and Regulatory News Service (RNS) for the period between January 1999 and December2004. The table also reports the keywords and key sections that are used to search for repurchase announcements from the Financial Times; FAME and Regulatory News Service (RNS).Panel B summarises the number of open market repurchases by year, size quartiles and book-to-market quintiles. At the end of June each year, all UK listed companies with above or equalto zero book-to-market ratios on DataStream are initially divided into four size quartiles and each size quartile contains 25% of all stocks. Size quartile 1 consists of the top 25% smallestcompanies while size quartile 4 constitutes the bottom 25% largest companies. Then each size quartile is further divided into 5 quintiles on the basis of book-to-market ratios and eachquintile contains 20% of all companies in a given size quartile. The 20% of all companies with the lowest book-to-market ratios in a given size quartile are assigned to quintile 1 whilequintile 5 contains the 20% of all companies with the highest book-to-market ratios in the given size quartile. This allocation procedure results in 20 reference portfolios each month. Sizeand book-to-market ranks of sample firms are determined against these 20 portfolios at the announcement.

I. Craw

ford, Z.

Wang

/ R

esearch in

International Business

and Finance

26 (2012) 26– 4633

Fig. 1. Natural log of monthly industrial production (LOP) and the number of repurchases (No of Announcements) announced every month between January 1999 and December 2004.


However, it is reasonable to see whether there is a link between the natural log of monthly industrialproduction and the number of repurchases announced every month. Fig. 1 shows how the natural logof industrial production and the number of share repurchase announcements made by UK firms ineach month of the sample period have evolved. It is clear that there is no relationship between thenatural log of industrial production and the number of repurchase announcements made by UK firms.Instead, the number of repurchase announcements is randomly distributed throughout the sampleperiod. It is possible that the lack of pattern is caused by the static economic state during the sampleperiod. Therefore, our findings should be interpreted in the context of a stable economic state.

3.2. Performance measurement

We examine both short-term returns surrounding the announcement and long run share per-formance following the announcement. Short-term share performance is calculated over variouswindows from 40 days before to 10 days following the announcement. As suggested by Ikenberryet al. (1995), abnormal returns which are calculated over such short intervals are not overly sensitiveto the benchmark used. Thus, we report results using a market adjusted model, measuring abnormalreturns in relation to the FTSE non-financial index. For each sample firm, the market adjusted model isestimated by using share mid-prices in the 51-day event window to subtract the corresponding shareprices of the FTSE non-financial index. The conventional t-tests are used to interpret significant level.

Care must be taken when measuring long run share performance since long run abnormal returnsare very sensitive to the benchmarks and the procedures used (Barber and Lyon, 1997; Lyon et al.,1999; Mitchell and Stafford, 2000). In this paper, we adopt two different approaches. The first is acalendar time technique recommended by Mitchell and Stafford (2000). The second is an equallyweighted buy-and-hold strategy used by Ikenberry et al. (1995) and recommended by Lyon et al.(1999). For both of these approaches, abnormal returns are calculated relative to equally weightedsize and book-to-market portfolios formed every June using all UK listed firms on the London StockExchange and AIM with zero or above book-to-market ratios. 20 size and book-to-market referenceportfolios are built according to size and book-to-market ratios. Size quartile 1 consists of the top 25%smallest firms while size quartile 4 constitutes the bottom 25% largest firms. Then each size quartile isfurther divided into 5 quintiles on the basis of book-to-market ratios and each quintile contains 20%of all firms in a given size quartile. The 20% of all companies with the lowest book-to-market ratios ina given size quartile are assigned to quintile 1 while quintile 5 contains the 20% of all firms with thehighest book-to-market ratios in the given size quartile.

Under equally weighted calendar time approach, abnormal returns (CTARs) are calculated eachmonth relative to monthly returns of matched size and book-to-market reference portfolios. Thisprocedure assumes monthly rebalancing with sample firms receiving equal portfolio weights eachmonth. Returns of size and book-to-market reference portfolios are calculated over 2 years assumingequal weighting. These returns are then used as benchmarks to measure abnormal share performance.Each month following the month of the announcement, abnormal returns of each sample firm aremeasured relative to its size and book-to-market portfolio matched at the announcement. The monthlyequally weighted abnormal returns of the portfolio of sample companies are then standardized byestimates of the portfolio standard deviation. Mitchell and Stafford (2000) mentioned two reasonsto standardize the monthly CTARs: to control heteroskedasticity and to give more weight to periodsof heavy repurchase announcements since the portfolio residual variance is decreasing in portfoliosize, ceteris paribus Statistical inference of each standardized monthly CTAR is the same as in Mitchelland Stafford (2000), which is based on the time-series mean of the monthly standardized CTARs andstandard error of the mean.

The calendar time approach is rather descriptive in nature and does not resemble a realistic invest-ment strategy, so the second buy-and-hold approach is used to represent investment experience ofpassive investors. We assume an equally weighted buy-and-hold investment in all sample firms begin-ning in the month following the announcement and continuing for 12 months. If a firm delists from theexchanges, we assume the investment is sold at the last available price and the proceeds from the saleare reinvested for the rest of the investment horizon in the firm’s matched size and book-to-marketportfolio. At the end of 12 months, the portfolio is rebalanced and contains only surviving firms. In


this paper, five firms depart from the sample 12-month following the announcement. Firms used tocalculate returns of matched size and book-to-market portfolios were treated in the same way. The1-year buy-and-hold return for each firm in a given size and book-to-market portfolio is calculatedfrom the next month following each June. The equally weighted average of all annual returns in thegiven size and book-to-market portfolio is then used as a benchmark return for sample firms ranked inthat particular size and book-to-market rank at the month of the announcement. Statistical inferenceof the 1-year and 2-year equally weighted buy-and-hold abnormal returns is done via bootstrappingand empirical p values as shown in Ikenberry et al. (1995).

3.3. Independent variables

3.3.1. Pre-announcement returns (CAR −40, −3)Prior studies (Ikenberry et al., 1995; Chan et al., 2004; Peyer and Vermaelen, 2009) show that

repurchasing firms tend to suffer from a recent share undervaluation. Pre-announcement share under-valuation is often considered to be the evidence of the market overreaction to bad news. To controlthe impact of mean reversion arising from negative returns prior to the announcements, we includepre-announcement returns into the regression analysis. Pre-announcement returns are measured inthe same manner as the initial market return.

3.3.2. Size and BTMV rankingAs explained earlier, 20 size and book-to-market portfolios are formed each June using listed firms

on the London Stock Exchange and AIM. The size quartile and BTMV quintile cut-off values are thendetermined at the end of each June. All sample firms are sorted into size quartiles and BTMV quintilesrelative to the cut-off values at the announcement. The largest firms are in size quartile 1 while valuefirms are in BTMV quintile 5. Panel B of Table 1 reports the number of open market repurchases bysize and book-to-market ranks.

3.3.3. PercentThe percentage of shares sought by managers at the announcement date. Following Comment

and Jarrell (1991) and Ikenberry et al. (1995), we use this variable to examine the possible impact ofpercentage shares announcement on the initial market returns and long run abnormal returns.

3.3.4. Log(1 + % completion rate)Completion rate of an open market share repurchase is calculated as a percentage, representing the

amount or the number of shares repurchased in the authorisation period (12 months) relative to thesize of the repurchase announced. It is relatively easy to identify the number of shares repurchasedin the authorisation period for a UK open market share repurchase programme because of law andLSE regulations. As required by the Companies Act 1985, UK listed firms need to report the amount ofshares repurchased during the financial year in their Annual Reports. In addition, all actual repurchasesare also reported in detail on RNS (Hemscott). However, there are some impediments for calculationsof completion rate. If firms repurchase shares in several consecutive years, repurchases reported inAnnual reports are sometimes overlapping. In this case, we resort to the RNS to clarify the amountor numbers of repurchases for each authorisation period. Another problem is that some firms’ annualreports have gone missing while repurchase news are not recorded on the RNS (Hemscott) becausethese firms are delisted. Due to these obstacles, we are unable to compute completion rate for 84cases. After computing completion rate, we add 100% to completion rate and then log transform(1 + completion rate) to control the boundary nature of completion rate (0–100%).

Finally, two dummy variables are constructed to accommodate the law change in December 2003and two types of announcement dates used in the paper.

3.3.5. Year dummyIn December 2003, the Companies Act 1985 was changed to allow repurchased shares to be treated

as treasury shares which could then be sold in the market later on. As changes in tax and regulationshave played a huge role in the swing between dividend and share repurchases both in the US (Bagwell


and Shoven, 1989; Fama and French, 2001) and the UK (Rau and Vermaelen, 2002), it is necessary toanalyse the nature of announcement abnormal returns for these two time periods, from January 1999to December 2003 and from January 2004 to December 2004. The year dummy variable is designedto capture the impact of the law change in December 2003 and is equal to 0 for the repurchasesannounced between January 1999 and December 2003 and 1 for the repurchases announced betweenJanuary 2004 and December 2004.

3.3.6. Announcement type dummyUnlike US share repurchases, UK repurchases do not have a universally defined announcement date.

Previous studies (Rees, 1996; Rau and Vermaelen, 2002) suggest several definitions of announcementdates for UK open market share repurchases. Since this paper focuses on the analysis of announce-ment period return upon the announcement of a repurchase, the repurchase completion date isexcluded from the study. We analyse two kinds of announcements. Following Rau and Vermaelen(2002), we regard the repurchase intention date as the announcement date, the date when a repur-chase announcement is first published on the RNS (Hemscott) and/or the Financial Times. When therepurchase intention date is untraceable, we use an alternative announcement date which is theshareholders’ general meeting (Annual General Meeting or Extraordinary General Meeting) date. Anauthorisation right of a repurchase for a UK company is granted by shareholders at the Annual GeneralMeeting or the Extraordinary General Meeting. The announcement type dummy variable is designedto differentiate the impact of these two kinds of announcement dates and is equal 1 for the repurchaseintention date and 0 for the AGM and EGM date.

4. Empirical results

4.1. Short-term abnormal returns

Table 2 provides a comprehensive analysis of short-term abnormal returns for our sample firms.Looking at repurchases by year, there are positive and statistically significant abnormal returns, mea-suring from 40 to 3 days prior to the announcement from 2000 to 2003. The only negative andsignificant abnormal return appears in the year 1999, −3.41%. These returns hardly suggest thatrepurchase announcements overall suffer from pre-announcement share undervaluation. The averagemarket return, measuring from 2 days before through two days following the announcement, is 1.33%which appears on average to be quite similar to these of previous studies Rau and Vermaelen (2002).The initial market reaction changes slightly between 1999 and 2003, from 2.20% in 2002 to −0.35% in2003.

Ikenberry et al. (1995) and Comment and Jarrell (1991) suggest that undervalued firms are likelyto announce larger share repurchase programmes and both papers find that larger share repurchaseprogrammes are received more favourably by the market. Short-term returns are reported accordingto the percentage of shares sought in Table 2. The initial market return pattern shows no support forshare undervaluation. For example, the mean announcement period return is 2.15% for repurchaseswhich are for less than 5% of outstanding shares while for repurchases which are for more than 10%of outstanding shares, the initial market reaction is 1.50%. It appears that small share repurchaseprogrammes in our sample are received more favourably by the market than large share repurchaseprogrammes, which contrast with the results reported by Ikenberry et al. (1995) and Comment andJarrell (1991).

Table 2 also shows short-term announcement returns by size quartiles and BTMV quintiles. Firmsranked in the smallest size quartile show the highest abnormal returns on average at the announce-ment, 2.22%, just slightly higher than 1.15% earned by firms ranked in the biggest size quartile. Whenlooking closely, we notice that the high initial market return of small firms can barely compensatethe pre-announcement share decline, −2.35% while big firms enjoy a positive and significant return(2.41%) prior to the announcement. Looking at book-to-market quintiles, the average market reactionfor value firms (quintile 5) is 1.27% while it is 1.70% for glamour firms (quintile 1). In addition, thereis no indication of pre-announcement share undervaluation across the five book-to-market quintiles.


Table 2Short-term abnormal returns surrounding the announcement of open market share repurchases.

n Days relative to repurchase announcement

−40 to −3 −2 to +2 +3 to +10

Time period1999 65 −3.41 (−1.944*) 1.91 (2.785***) −1.35 (−1.680*)2000 73 4.02 (2.154**) 1.41 (1.179) −0.12 (−0.075)2001 69 4.77 (3.001***) 1.57 (2.512**) 0.10 (0.144)2002 93 4.29 (3.833***) 2.20 (4.147***) 0.81 (1.422)2003 80 2.81 (2.226**) −0.35 (−0.616) 0.11 (0.232)2004 88 0.37 (0.367) 1.32 (2.549**) −0.59 (−1.752*)

Total 468 2.27 (3.872***) 1.33 (4.6147***) −0.11 (−0.346)Percent of shares sought

≤5% 45 2.84 (2.0241**) 2.15 (2.5868**) 1.69 (1.7284*)>5%, but ≤10% 239 2.40 (2.8879***) 1.26 (3.2982***) −0.22 (−0.428)>10% 150 3.60 (3.3905***) 1.50 (2.7442***) −0.48 (−0.943)No information 34 −5.24 (−2.58**) −0.04 (−0.036) −0.13 (−0.129)

Size quartilesSize 1 (small) 26 −2.35 (−0.800) 2.22 (3.0298***) −0.29 (−0.345)Size 2 72 3.40 (1.9988**) 0.90 (1.3747) 0.22 (0.2964)Size 3 102 3.39 (2.454**) 1.89 (3.7197***) −0.14 (−0.21)Size 4 (large) 241 2.41 (3.2262***) 1.15 (2.6577***) −0.31 (−0.598)

Book-to-market quintilesBTMV 1 (lowest BTMV) 57 3.61 (1.8879*) 1.70 (2.5294**) 1.39 (2.3679**)BTMV 2 49 0.74 (0.5267) −0.13 (−0.162) −0.51 (−0.785)BTMV 3 75 2.82 (1.9755*) 1.97 (3.0568***) −0.29 (−0.546)BTMV 4 99 3.20 (2.1786**) 1.51 (2.6081**) −0.07 (−0.114)BTMV 5 (highest BTMV) 161 2.12 (2.18**) 1.27 (2.3527**) −0.66 (−0.847)

Announcement typeRepurchase intentions

1999 37 −6.48 (−2.751***) 2.03 (2.146**) −1.40 (−1.222)2000 30 2.91 (1.039) 0.44 (0.213) 0.03 (0.010)2001 22 1.00 (0.408) 2.82 (2.162**) 1.09 (0.832)2002 26 2.88 (1.318) 1.81 (1.775*) 2.65 (2.580**)2003 25 −3.20 (−2.648**) −0.16 (−0.125) 0.56 (0.784)2004 31 0.05 (0.032) 2.77 (2.457**) −0.97 (−1.843*)

Total 171 −0.80 (−0.857) 1.62 (2.951***) 0.17 (0.230)AGM/EGM

1999 28 0.65 (0.265) 1.76 (1.743*) −1.29 (−1.153)2000 43 4.80 (1.909*) 2.09 (1.455) −0.23 (−0.190)2001 47 6.53 (3.273***) 0.98 (1.450) −0.36 (−0.432)2002 67 4.84 (3.704***) 2.36 (3.762***) 0.09 (0.135)2003 55 5.55 (3.400***) −0.44 (−0.734) −0.09 (−0.145)2004 57 0.55 (0.405) 0.53 (1.082) −0.39 (−0.886)

Total 297 4.04 (5.495***) 1.16 (3.563***) −0.28 (−0.866)

This table reports abnormal returns (in percent) measured with respect to the FTSE non-financial index over days −40 to −3,−2 to +2 and +3 to +10 relative to the announcement of an open market share repurchase announced between January 1999and December 2004 (t-statistics reported in parentheses). Abnormal returns are reported for all sample firms and by year, thepercentage of shares sought at the announcement, size quartiles and book-to-market quintiles. As explained in Table 1, sizequartiles and book-to-market quintiles of sample firms are determined again all UK listed firms at the announcement.

* Denotes significance levels of 10% using two-tailed t-test.** Denotes significance levels of 5% using two-tailed t-test.

*** Denotes significance levels of 1% using two-tailed t-test.

Value firms and glamour firms earn positive and significant announcement returns (3.61% and 2.12%,respectively) before the announcement.

Short-term abnormal returns by year, percentage of shares sought, size quartiles and book-to-market quintiles reveal limited evidence of share undervaluation prior to the announcement. Eventhough pre-announcement negative abnormal returns appear in two occasions (smallest firms-sizequartile 1; repurchases announced in 1999), the announcement abnormal returns of these two sub-groups are not large enough to recover the size of share undervaluation. If repurchases are used to


Table 3Cross-sectional regressions of announcement abnormal returns.

Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7

Constant 0.0137 0.018 0.0169 0.0108 0.0116 0.0133 0.0241p-Value (0.0000) (0.0588) (0.0520) (0.1203) (0.0004) (0.0001) (0.2451)Percent of shares sought −0.0329 −0.0319p-Value (0.7075) (0.7358)PreCAR (−40 to −3) −0.0169 −0.0018p-Value (0.4797) (0.4652)Size ranking −0.0012 −0.0023p-Value (0.6640) (0.4809)BTMV ranking 0.0006 0.0003p-Value (0.7667) (0.8926)Announcement type 0.0046 0.0033p-Value (0.4682) (0.6168)Year dummy −0.0002 −0.0017p-Value (0.9846) (0.8046)

Adjusted R-squared −0.0009 −0.0019 −0.0019 −0.0021 −0.0009 −0.0021 −0.0106Prob F-test 0.4571 0.6551 0.6966 0.7747 0.4387 0.9872 0.9455Number of observations 468 434 441 441 468 468 413

This table reports cross-sectional results of the initial market returns on various explanatory variables. The dependent variableis the difference between share returns of a repurchase over days −2 to +2 relative to the announcement and share returnsof the market model, the FTSE non-financial index over the same period. Size ranking (one being the smallest) is based onthe market value of equity at the announcement while BTMV ranking (one being the lowest book-to-market ratios) is basedon the book-to-market ratio at the announcement. Announcement type dummy is one when the announcement date is therepurchase intention date, zero otherwise. Year dummy is one when a repurchase is announced between January 2004 andDecember 2004, zero otherwise. CAR (−40, −3) is the difference between share returns of a repurchase over days −40, −3relative to the announcement and share returns of the market model, the FTSE non-financial index over the same period. Thepercentage of shares sought is the percentage of shares that a firm intends to repurchase at the announcement. Numbers inparentheses are White (1980) heteroskedasticity-adjusted p values.

signal the market’s overreaction towards bad news prior to the announcement, it is difficult to justifythe market reaction surrounding the announcement for small firms or for repurchases announced in1999. It is hard to believe that UK repurchases are motivated by share undervaluation based on short-term share performance. These results are very different from US studies. Ikenberry et al. (1995) showthat US repurchases evidently suffer from pre-announcement share undervaluation. For example,there are negative abnormal returns prior to the announcement, larger share repurchase programmesare received more favourably by the market and two smallest size deciles show the highest abnormalreturns than other size deciles. It is clear that UK repurchases behave differently from US repur-chases in the short-term. It is possible that these differences are attributable to the motivations of UKrepurchases which are still debatable in the literature.

The announcement type panel in Table 2 reveals short-term abnormal returns by two kinds ofannouncement dates, namely, repurchase intention dates and AGM/EGM dates. As explained earlierin Section 3, UK repurchases have different announcement dates from US repurchases. It is possiblethat two kinds of announcement dates might have different impacts on share performance in theshort-term and long-term. Short-term abnormal returns show some differences across these twoannouncement dates. Repurchase intentions seem to suffer from significant share undervaluation in1999 and 2003 (−6.48% and −3.20%, respectively) while AGM/EGM dates show highly positive andsignificant abnormal returns between 2000 and 2003. On average, repurchase intentions performslightly better than AGM/EGM dates around the announcement.

To clarify the nature of announcement abnormal returns, we regress announcement returnson various independent variables described in Section 3. We examine the relationship betweenannouncement returns and these variables individually and in aggregate. The regression results arereported in Table 3 and are consistent with the evidence presented in Table 2. As before, the regressionresults provide no indication of share undervaluation. For instance, book-to-market and size rankshave no impact on the initial market reaction to repurchases announcements while the percentageof shares sought and pre-announcement returns are negatively but insignificantly related to the


Table 4Mean monthly calendar-time portfolio abnormal returns (CTARs) of open market share repurchases announced between 1999and 2004.

N Monthly CTARs Cumulative monthly CTARs t-Statistics (p-values)

All firms12-Month 441 0.1037 1.2444 3.57 (0.0006)24-Month 436 0.0777 1.8648 3.51 (0.0007)

Size quartile 1 (small)12-Month 26 0.0602 0.7224 0.68 (0.4972)24-Month 26 0.0781 1.8739 1.5850 (0.1165)

Size quartile 212-Month 72 0.9368 11.2415 1.3696 (0.1747)24-Month 70 0.0139 0.3332 0.3040 (0.7618)

Size quartile 312-Month 102 0.2373 2.8476 2.2517 (0.0271)24-Month 102 0.0826 1.9822 1.7114 (0.0905)

Size quartile 4 (large)12-Month 241 0.0998 1.1977 2.4439 (0.0167)24-Month 238 0.0774 1.8577 2.6523 (0.0094)

BTMV quintile 1 (low)12-Month 57 0.9237 11.0844 1.0573 (0.2939)24-Month 56 0.0625 1.4996 1.3319 (0.1864)

BTMV quintile 212-Month 48 0.1075 1.2898 0.9257 (0.3576)24-Month 46 0.0915 2.1968 2.0329 (0.0451)



BTMV quintile 5 (high)12-Month 161 0.1803 2.1635 3.5927 (0.0006)24-Month 159 0.0784 1.8821 2.9456 (0.0041)

This table shows 1-year and 2-year equally weighted calendar time abnormal returns for the whole sample as well as by sizequartiles and book-to-market quintiles. The monthly CTARs are calculated each month as the difference between the sampleportfolio and the reference portfolio return matched by size and book-to-market ranking, standardized by the sample portfolioresidual standard deviation. Each month, the equally weighted sample portfolios are formed containing sample companiesthat announced repurchases between 1999 and 2004. The sample portfolio is rebalanced monthly to drop sample companiesthat reach the end of their 1- or 2-year investment horizons. Size quartiles and book-to-market quintiles of sample firms aredetermined again all UK listed firms at the announcement. All UK listed companies are sorted into 4 size quartiles at the end ofeach June. Size 1 contains 25% of all UK listed companies with the smallest market capitalisations. Each size quartile is furthersegmented into equally sized quintiles on the basis of book-to-market ratios, with the lowest ratios assigned to BTMV quintile 1.Reference portfolios are size and book-to-market portfolios matched to the repurchasing companies by size and price-to-bookat the announcement. Cumulative CTARs reported in the Column 4 are calculated by multiplying monthly CTARs (column 3)by 12- or 24-month. Residuals are calculated from monthly return difference between the sample portfolios and referenceportfolios matched by size and book-to-market ranking. Mean monthly CTARs and standard errors are calculated from the timeseries of monthly CTARs.

initial market returns. Finally, the initial market returns are not statistically significantly influencedby different announcement dates.

4.2. Long-term abnormal returns

Equally weighted mean calendar-time 1-year and 2-year abnormal returns and equally weightedbuy-and-hold 1-year and 2-year returns and abnormal returns are reported for the whole sample andby size and book-to-market ranking in Tables 4 and 5.

The equally weighted 1-year and 2-year cumulative calendar-time abnormal returns (CCTARs)for the whole sample are, respectively, 1.24% and 1.86%, both statistically significant at a 1% level.The highest 1-year equally weighted CCTAR, 11.24%, is earned by sample firms in size quartile 2,


Table 5Long run buy-and-hold returns.

N Sample companies Reference portfolio Return differences p-Value

All firms12-Month 441 15.9102 13.2012 2.7090 0.098024-Month 436 39.1036 28.6627 10.4410 0.0030

Size quartile 1 (small)12-Month 26 40.2484 26.3046 13.9438 0.186024-Month 26 104.2952 45.6334 58.6618 0.0400

Size quartile 212-Month 72 13.0987 22.3692 −9.2705 0.878024-Month 70 36.2967 43.6322 −7.3355 0.6870

Size quartile 312-Month 102 22.7683 12.1529 10.6154 0.008024-Month 102 52.1688 27.3374 24.8314 0.0000

Size quartile 4 (large)12-Month 241 11.2219 9.4923 1.7296 0.160024-Month 238 27.2080 22.9739 4.2342 0.0680

BTMV quintile 1 (low)12-Month 57 14.7185 8.4940 6.2246 0.133024-Month 56 22.2187 15.4791 6.7396 0.1540

BTMV quintile 212-Month 48 −0.8195 7.7624 −8.5819 0.800024-Month 46 24.4864 21.8704 2.6160 0.3420

BTMV quintile 312-Month 74 10.6825 10.4319 0.2505 0.418024-Month 74 23.8571 16.8644 6.9927 0.1580

BTMV quintile 412-Month 101 15.0086 12.8270 2.1816 0.303024-Month 101 40.5258 30.4662 10.0595 0.0970

BTMV quintile 5 (high)12-Month 161 24.2883 17.9969 6.2914 0.023024-Month 159 55.4718 39.6163 15.8555 0.0040

This table reports the 1-year and 2-year compounded buy-and-hold returns (in percent) for equally weighted portfolios of UKcompanies announcing open market share repurchases between 1999 and 2004. Returns are reported by size quartile and BTMVquintiles. Size quartile and BTMV quintile of a repurchasing firm is based on the market capitalisation and book-to-market ratioat the announcement relative to 20 size and BTMV portfolios formed at the end of June each year. All UK listed companies inDataStream are ranked into one of four size quartiles. Size 1 contains 25% of all UK listed companies with the smallest marketcapitalisations. Each of size quartiles is further segmented by book-to-market ratios into quintiles. Each of five book-to-marketquintiles contains 20% of all stocks in a given size quartile. Quintile 1 includes stocks with the lowest book-to-market ratioswhile quintile 5 contains stocks with the highest book-to-market ratios. Compounded buy-and-hold returns assume annualrebalancing. Reference portfolios are size and book-to-market portfolios matched to the repurchasing companies by size andprice-to-book at the announcement. P values are derived from bootstrapping. Repurchasing firms that cannot be ranked at theannouncement are excluded.

though statistically insignificant at any conventional level. The highest and significant 2-year equallyweighted CCTAR, 1.98% (t-statistic = 1.71), is generated by book-to-market quintile 3. Sample firmsin size quartiles 3 and 4 (large companies) both attract statically significant abnormal returns for 1-year and 2-year investment horizons while small companies (size quartile 1) generate statisticallyinsignificant 1-year and 2-year abnormal returns. We then focus on sample firms ranked in either thebottom or top quintile on BTMV after controlling size. The 1-year and 2-year equally weighted CCTARsin value firms (BTMV quintile 5) are both positive and significant (2.16%, t-statistic = 3.59; 1.88%, t-statistic = 2.94, respectively). For glamour firms (BTMV quintile 1), the 1-year equally weighted CCTARis dramatic (11.08%) but nevertheless is statically insignificant (t = 1.06), while the 2-year equallyweighted CCTAR of glamour firms is 1.50% with a t-statistic of 1.33.

The 1-year and 2-year equally weighted buy-and-hold abnormal returns of the whole sample are2.71% and 10.44%. Both abnormal returns are statistically significant as the associated p-values are0.098 and 0.003, meaning that 98 and 3 of 1000 corresponding pseudo portfolios specifically formedfor these two investment periods and matched by size and book-to-market ratios performed better.When abnormal returns are reported by size quartiles, it is clear that firms in size quartile 1 (smallest


firms) perform much better than firms in other quartiles in 1-year and 2-years. When abnormal returnsare shown by book-to-market quintiles, it is evidence that value firms (quintile 5) perform better thanglamour firms (quintile 1) 2 years following the announcement while glamour firms perform just aswell as value firms in year one. Value firms attract the highest 2-year buy-and-hold abnormal returnson average (15.86%, p = 0.004) across the book-to-market spectrum.

Abnormal returns in Tables 4 and 5 both confirm that UK repurchases show a positive and significantreturn drift following the announcement. However, the impact of size ranks on abnormal returns isnot consistent under two measurement methods. For the calendar time approach, largest firms seemto perform better than smallest firms in year one and year two. Smallest firms appear to perform betterthan largest firms under the buy-and-hold strategy. However, the buy-and-hold abnormal returns ofsmallest firms might be affected by a small number of cases since small firms (26) consist of less than6% of the sample (468) measured on the number of announcements basis. On the other hand, book-to-market ranks are related to abnormal returns in a similar way using either of these measurementmethods. That is, value firms do not outperform glamour firms during the authorisation but clearlyoutperform glamour firms 2 years following the announcement. The remaining question is whetherabnormal returns are the market delayed reaction towards repurchase news.

To answer this question, we regress 1-year and 2-year buy-and-hold abnormal returns on var-ious factors. The regression results are reported in Table 6. The results show that book-to-markethas explanatory power for 1-year abnormal returns but the power weakens when other indepen-dent variables are included into the regression analysis. Book-to-market is statistically related to2-year abnormal returns in all regressions. For example, a move from the lowest to the highestbook-to-market quintile changes the 2-year buy-and-hold abnormal return by 14.8% (0.0370 × 4).The regression results are somewhat in line with the evidence reported in Tables 4 and 5. Size doesnot influence long run share performance in a significant way while the impact of book-to-marketbecomes statistically significant 2 years following the announcement. In addition, the percentage ofshares sought is unrelated to 1-year and 2-year buy-and-hold long run abnormal returns. The resultssuggest that long run abnormal returns are not explained by the market overreaction hypothesis or themarket underreaction hypothesis. Particularly, 1-year buy-and-hold abnormal returns of value firmsshow no indication of a delayed market reaction to repurchase news or a correction to the marketoverreaction to bad news. Since the market underreaction and overreaction hypotheses are not plau-sible for UK repurchases, we need to identify the reasons for extraordinary share performance amongvalue firms 2 years following the announcement. It appears that value firms have made some financialdecisions during the authorisation period and these decisions are received favourably by the marketlater on. Thus, we focus on actual repurchases which firms can undertake during the authorisationperiod.

The regression results show that long run abnormal returns are not related to actual repurchasesalthough actual repurchases have a higher explanatory power for 2-year abnormal returns. Severalpapers reveal that actual repurchases tend to follow share declines (Ikenberry et al., 2000; Stephensand Weisbach, 1998; Dittmar, 2000). Our results suggest that actual repurchases are not motivatedby share price movement but possibly by other reasons, such as distribution of free cash flows, alter-nation of capital structure or improvement of earnings per share, etc. In particular, the free cash flowhypothesis suggests that share repurchases can be used to disburse free cash flows and value firmsbenefited greatly from actual repurchases since these firms do not have plenty of profitable oppor-tunities but do have free cash. By examining UK actual repurchases, Oswald and Young (2008) findthat actual repurchases are likely to be motivated by the free cash flow hypothesis. It is reasonable toinvestigate actual repurchase activities among samples firms and Table 7 provides a detailed analysisof actual repurchases.

Completion rates change slightly across years, decreasing from 39.64% in 1999 to 24.32% in 2003.Nearly 28% of sample firms completed 50% or over as promised. Only 15 sample firms abandonshare repurchase programmes completely between 1999 and 2004. The majority of sample firmsrepurchase some shares during the authorisation period. We then look at completion rates amongsmall firms and large firms. It is clear that large firms repurchase on average 4% more shares thansmall firms during the authorisation period. Small firms in fact suffer from share undervaluationprior to the announcement and the announcement abnormal return of small firms is lower than

42I.

Crawford,

Z. W

ang /

Research

in International

Business and

Finance 26 (2012) 26– 46

Table 6Cross-sectional regressions of long run abnormal returns.

Y1BHRs Y2BHRs

Model 1 Model 2 Model 3 Model 4 Model 5 Model 6

Intercept −0.1091 (0.4071) −0.1287 (0.3206) −0.1116 (0.4199) 0.1343 (0.6409) −0.0882 (0.6938) −0.1316 (0.5791)Size ranking 0.0120 (0.7036) 0.0342 (0.1650) 0.0283 (0.2707) −0.0514 (0.5182) 0.0257 (0.5441) 0.0223 (0.6102)BTMV ranking 0.027 (0.0612) 0.0255 (0.1046) 0.0229 (0.1635) 0.0384 (0.0792) 0.0370 (0.0745) 0.0363 (0.0979)Announcement type −0.0583 (0.1696) −0.0443 (0.3442) −0.0658 (0.2469) −0.0622 (0.3049)Year dummy 0.0013 (0.9738) 0.0062 (0.8748) 0.0272 (0.7069) 0.0285 (0.7046)Percent of shares sought −0.3126 (0.6120) −0.2419 (0.7077) −0.3120 (0.6858) −0.2224 (0.7844)Log(1 + % completion rate) 0.0052 (0.9513) 0.1531 (0.2220)Adjusted R-squared 0.0031 0.0037 −0.0052 0.0047 −0.0032 −0.0042Number of observations 441 413 379 436 412 379

F-Test (p-values) 0.1872 0.2588 0.6688 0.1325 0.5969 0.6220

This table presents cross-sectional regression results of long run abnormal returns of UK repurchasing firms on various explanatory variables. The dependent variable is the 1-year or2-year compounded buy-and-hold abnormal return defined as the difference in holding a portfolio of sample firms and its corresponding portfolio matched on size and book-to-marketranking. Size ranking (one being the smallest) is based on the market value of equity at the announcement while BTMV ranking (one being the lowest book-to-market ratios) is based onthe book-to-market ratio at the announcement. Announcement type dummy is one when the announcement date is the repurchase intention date, zero otherwise. Year dummy is onewhen a repurchase is announced between January 2004 and December 2004, zero otherwise. The percentage of shares sought is the percentage of shares that a firm intends to repurchaseat the announcement. Numbers in parentheses are White (1980) heteroskedasticity-adjusted p values.

I. Craw

ford, Z.

Wang

/ R

esearch in

International Business

and Finance

26 (2012) 26– 4643

Table 7Completion rates of open market share repurchases between 1999 and 2004.

Year N Mean completion rate Zero completion rate Completion rate

0–5% 5–10% 10–25% 25–50% 50–75% 75–100% Not Infor

1999 65 39.64% 4 2 2 5 3 13 15 212000 73 37.13% 3 5 5 10 3 8 17 222001 69 33.49% 1 11 4 14 11 9 10 92002 93 39.52% 3 9 4 17 24 7 17 122003 80 24.32% 3 12 11 19 12 9 5 92004 88 34.28% 1 8 11 20 16 7 14 11

All years 468 34.61% 15 47 37 85 69 53 78 84

Small firms 26 33.94% 2 4 4 4 4 1 5 2Large firms 241 38.23% 10 27 21 44 36 33 38 32

Extreme growth 57 36.05% 0 6 5 21 9 6 9 1Extreme value 161 42.88% 6 17 12 27 20 22 33 24

This table reports completion rates (in percent) by year, by small firms, large firms, value firms and glamour firms for open market share repurchases announced between 1999 and 2004.Completion rate of an open market share repurchase is calculated as a percentage, representing the amount or the number of shares repurchased in the authorisation period (12 months)relative to the size of the repurchase announced. This table reports completion rates for all sample firms including those that departed from the sample during the authorisation period.Small firms are ranked in size quartile 1 while large firms are ranked in size quartile 5. Extreme value firms are ranked in book-to-market quintile 5 while extreme glamour firms areranked in book-to-market quintile 1. Size quartile and BTMV quintile of a repurchasing firm is based on the market capitalisation and book-to-market ratio at the announcement relativeto 20 size and BTMV portfolios formed at the end of June each year.


pre-announcement share undervaluation. Such a return pattern surrounding the announcement stillfails to stimulate high completion rates among small firms. It looks as if UK actual repurchases are notmotivated by pre-announcement share undervaluation. Turning to value firms and glamour firms,we observe an over 6% completion rate gap. Value firms repurchase more shares than glamour firmsduring the authorisation period. It is reasonable to assume that the market begins to notice sizeableactual repurchases conducted by value firms later on during the authorisation period. This explainsnot only the weak link between 1-year abnormal returns and book-to-market but also the strong linkbetween 2-year abnormal returns and book-to-market. Therefore, long run abnormal returns are themarket’s delayed reaction towards actual repurchases.

5. Conclusion

The unique characteristics of UK open market repurchases have provided us an opportunity toexamine whether extraordinary share performance surrounding and following the announcement isonly a US phenomenon. This paper aims to determine whether share performance is a correction tothe market overreaction to bad news prior to the announcement or is the market delay’s reaction torepurchase news. Thus, the analysis of this paper is centred on share performance surrounding andfollowing the announcement.

We first consider short-term returns surrounding the announcement. UK repurchases on aver-age do not suffer from pre-announcement share undervaluation. In two occasions where smallfirms and repurchases announced in 1999 suffer from significant share undervaluation, the averageannouncement returns are hardly large enough to encourage firms to use repurchases as a signal forpre-announcement share declines. It appears that UK repurchases are not preceded by the marketoverreaction to bad news. Looking at the announcement abnormal returns, there is some evidence torefute share undervaluation. Large repurchase programmes are received less favourably by the mar-ket while glamour firms have a higher announcement abnormal return than value firms. The evidencesuggests that the announcement abnormal returns are not a response to the market overreaction tobad news before the announcement.

We then consider long run share performance following the announcement. Two measurementmethods show that UK repurchases on average attract a positive and significant return during theauthorisation period and 2 years following the announcement, conditional on size and book-to-market. When analysing the causes of long run abnormal returns, we find that the size impact isnot consistent under the two measurement methods. The small number of small firms might haveinfluenced the buy-and-hold long run abnormal returns while the calendar time approach is moresuitable to calculate long run abnormal returns of small firms. The calendar time approach revealsthat size is not the dominant factor in determining abnormal returns. Small firms underperform largefirms 1 year and 2 years following the announcement. The impact of book-to-market under thesetwo measurement methods suggests that value firms outperform glamour firms 2 years following theannouncement. On the other hand, value firms do not outperform glamour firms during the autho-risation period (1 year following the announcement). The regression results show that size does notinfluence long run abnormal returns while book-to-market is a dominant factor in 2-year buy-and-hold abnormal returns. Further analysis reveals that such a return pattern can be explained by actualrepurchase activity. Small firms which suffer from pre-announcement share undervaluation repur-chase fewer shares than large firms which enjoy positive and significant returns before and around theannouncement. In contrast, value firms repurchase considerably more shares than glamour firms dur-ing the authorisation period. Thus, we conclude that the UK market underreacts to actual repurchasesmade by value firms.

Finally, this paper contributes to the literature by examining share performance of UK repurchasesgoverning by regulations and law which are different from US repurchases. We find some similaritiesbetween US and UK repurchases. Like US repurchases, UK repurchases attract positive and significantshort-term abnormal returns surrounding the announcement and long run abnormal returns follow-ing the announcement. Two-year buy-and-hold abnormal returns are particularly high among valuefirms. The differences between US and UK repurchases are striking too. Unlike US repurchases, UKrepurchases on average do not suffer from share undervaluation prior to the announcement. There


is no evidence to indicate that the market reacts overly to bad news prior to the announcement. Inthe long run, abnormal returns among repurchases are only related to book-to-market. It seems thatactual repurchases conducted by value firms promote significant high abnormal returns 2 years fol-lowing the announcement. This conclusion is different from the market underreaction hypothesisor the market overreaction hypothesis. We also investigate the impact of different announcementdates on short-term and long-term share performance. The results suggest that share performance ofUK share repurchases are not statistically influenced by different kinds of announcement dates. Theabnormal long run share performance following open market share repurchases seems to be a uni-versal phenomenon and has not died away after it is first discovered by US studies. It is interesting forfuture research to investigate whether a portfolio of repurchasing firms would outperform benchmarkmarket, size, value and momentum portfolios in the long run.

Acknowledgements

We would like to thank the anonymous referees, Brian Lucey (the editor), Carol Padgett, AlfonsoDufour, Chris Brooks, Alan Gregory for their helpful comments and suggestions, and seminar partic-ipants at the 2007 Midwest Finance Association meetings, the 2009 Australian Banking and Financemeetings, the University of Reading and Birkbeck, University of London. All remaining errors are ourown.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, in the online version, atdoi:10.1016/j.ribaf.2011.05.002.

References

Bagwell, Laurie Simon, 1991. Share repurchase and takeover deterrence. RAND Journal of Economics 22, 72–88.Bagwell, Laurie Simon, Shoven, John B., 1989. Cash distribution to shareholders. Journal of Economic Perspectives 3, 129–140.Barber, Brad M., Lyon, John D., 1997. Detecting long-run abnormal stock returns: the empirical power and specification of test

statistics. Journal of Financial Economics 43, 341–372.Billett, Mattew T., Xue, Hui F., 2007. The takeover deterrent effect of open market share repurchases. Journal of Finance 62,

1827–1850.Brav, Alon, Graham, John R., Harvey, Campbell R., Michaely, Roni, 2005. Payout policy in the 21st century. Journal of Financial

Economics 77, 483–527.Chan, Konan, Ikenberry, David, Lee, Inmoo, 2004. Economic sources of gain in stock repurchases. Journal of Financial and

Quantitative Analysis 39, 461–479.Comment, Robert, Jarrell, Gregg A., 1991. The relative signalling power of Dutch auction and fixed price self-tender offers and

open-market share repurchases. Journal of Finance 46, 1243–1271.Dann, Larry Y., 1981. Common stock repurchases: an analysis of returns to bondholders and stockholders. Journal of Financial

Economics 9, 113–138.Dittmar, Amy K., 2000. Why do firms repurchase stock? Journal of Business 73, 321–356.Dixon, Rob, Palmer, Graham, Stradling, Bob, Woodhead, Anne, 2008. An empirical survey of the motivation for share repurchases

in the UK. Managerial Finance 34, 886–906.Fama, Eugene F., French, Kenneth R., 2001. Disappearing dividends: changing firm characteristics or lower propensity to pay?

Journal of Financial Economics 60, 3–43.Farrugia, Damien, Graham, Michael, Yawson, Alfred, 2011. Economic conditions and the motives for multiple open-market

share buybacks. Research in International Business and Finance 25, 156–168.Grullon, Gustavo, Michaely, Roni, 2004. The information content of share repurchase programs. Journal of Finance 59, 651–680.Grullon, Gustavo, Michaely, Roni, 2002. Dividends, share repurchases and the substitution hypothesis. Journal of Finance 57,

1649–1684.Hertzel, Michael, Jain, Prem C., 1991. Earnings and risk changes around stock repurchase tender offers. Journal of Accounting

and Economics 14, 253–274.Hovakimian, Armen, 2004. The role of target leverage in security issues and repurchases. Journal of Business 77, 1041–1072.Hovakimian, Armen, Opler, Tim, Titman, Sheridan, 2001. The debt-equity choice. Journal of Financial and Quantitative Analysis

36, 1–25.Ikenberry, David, Lakonishok, Josef, Vermaelen, Theo, 1995. Market underreaction to open market share repurchases. Journal

of Financial Economics 39, 181–208.Ikenberry, David, Lakonishok, Josef, Vermaelen, Theo, 2000. Stock repurchases in Canada: performance and strategic trading.

Journal of Finance 55, 2372–2397.Jagannathan, Murali, Stephens, Clifford, 2003. Motives for multiple open-market repurchase programs. Financial Management

(Blackwell Publishing Limited) 32, 71.

http://dx.doi.org/10.1016/j.ribaf.2011.05.002


Jagannathan, Murali, Stephens, Clifford P., Weisbach, Michael S., 2000. Financial flexibility and the choice between dividendsand stock repurchases. Journal of Financial Economics 57, 355–384.

Jensen, Michael C., 1986. Agency costs of free cash flow, corporate finance, and takeovers. The American Economic Review 76,323–329.

Kothari, S.P., Jerold, Warner B., 1997. Measuring long-horizon security price performance. Journal of Financial Economics 43,301–339.

Lasfer, M. Ameziane, 2001. The Market Valuation of Share Repurchases in Europe. City University Business School.Lee, Chun I., Ejara, Demissew Diro., Gleason, Kimberly C., 2010. An empirical analysis of European stock repurchases. Journal of

Multinational Financial Management 20, 114–125.Lie, Erik, 2005. Operating performance following open market share repurchase announcements. Journal of Accounting and

Economics 39, 436–441.Lie, Erik, Lie, Heidi J., 1999. The role of personal taxes in corporate decisions: an empirical analysis of share repurchases and

dividends. Journal of Financial and Quantitative Analysis 34, 533–552.Lyon, John D., Barber, Brad M., Tsai, Chih-Ling, 1999. Improved methods for tests of long-run abnormal stock returns. Journal

of Finance 54, 165–201.Mitchell, Mark L., Stafford, Erik, 2000. Managerial decisions and long-term stock price performance. Journal of Business 73,

287–320.Oswald, Dennis, Young, Steven, 2004. What role taxes and regulation? A second look at open market buyback activity in the

UK. Journal of Business Finance and Accounting 31, 257–292.Oswald, Dennis, Young, Steven, 2008. Share reacquisitions, surplus cash and agency problems. Journal of Banking and Finance

32, 795–806.Peyer, Urs, Vermaelen, Theo, 2009. The nature and persistence of buyback anomalies. Review of Financial Studies 22, 1694–1745.Rau, P. Raghavendra, Vermaelen, Theo, 2002. Regulation, taxes, and share repurchases in the United Kingdom. Journal of Business

75, 245–282.Rees, William, 1996. The impact of open market equity repurchases on UK equity prices. European Journal of Finance 2, 353–370.Skinner, Douglas J., 2008. The evolving relation between earnings, dividends, and stock repurchases. Journal of Financial

Economics 87, 582–609.Stephens, Clifford P., Weisbach, Michael S., 1998. Actual share reacquisitions in open-market repurchase programs. Journal of

Finance 53, 313–333.Stulz, René M., 1988. Managerial control of voting rights: financing policies and the market for corporate control. Journal of

Financial Economics 20, 25–54.Vermaelen, Theo, 1981. Common stock repurchases and market signaling: an empirical study. Journal of Financial Economics

9, 139–183.Von Eije, Henk., Megginson, William L., 2008. Dividends and share repurchases in the European Union. Journal of Financial

Economics 89, 347–374.Wang, Chuan-San, Strong, Norman, Tung, Samuel, Lin, Steve, 2009. Share repurchases, the clustering problem, and the free cash

flow hypothesis. Financial Management 38, 487–505.Zhang, Hua, 2005. Share price performance following actual share repurchases. Journal of Banking and Finance 29.

Documents

Is the market underreacting or overreacting to open market share repurchases? A UK perspective