Documents
Resources
Learning Center
Upload
Plans & pricing Sign in
Sign Out
Get this document free

The information content of share repurchases

VIEWS: 131 PAGES: 41

									    Why do firms repurchase shares: free cash flow or
                          information signaling?




      Susanne Espenlaub, Stephen Lin, Norman Strong, Chuan-San Wang*




__________________________________
*Susanne Espenlaub and Stephen Lin are Senior Lecturers in Accounting and Finance,
Norman Strong is a Professor of Accounting and Finance, Chuan-San Wang is a PhD
candidate, all in the Manchester Accounting and Finance Group, Manchester Business
School, The University of Manchester.
Address for correspondence: Professor Norman Strong, Manchester Accounting and
Finance Group, Manchester Business School, The University of Manchester, Booth Street
West, Manchester M15 6PB, UK. Email: norman.strong@mbs.ac.uk.
Acknowledgement: We have benefited from the comments of Ian Garrett, Alex Taylor,
Stuart Hyde, Stephen Young and seminar participants at the University of Manchester. Any
remaining errors are our own.



                                                                                       1
 Why do firms repurchase shares: free cash flow or information signaling?


                                         Abstract


       We test the free cash flow and information signaling explanations of share repurchases
by examining the operating performance, capital expenditures, cash reserves, and equity risk
of firms before and after actual share repurchases. Our evidence suggests that operating
performance does not improve after actual share repurchases, and there is no relation between
post-event changes in operating performance and the size of share repurchases. We find that
investment opportunities and cash reserves decrease from pre- to post-repurchase periods, and
that firms repurchase shares to re-balance risk. Overall, our evidence favours the free cash
flow hypothesis over the information-signaling hypothesis.




Key Words: share repurchases, agency costs, information asymmetry.


JEL classification: D82, G35




                                                                                           2
 Why do firms repurchase shares: free cash flow or information signaling?

       We examine UK repurchase execution announcements, which are compulsory and

timely disclosures of actual ordinary share repurchases on the London Stock Exchange (LSE).

More specifically, we compare the operating performance, capital expenditures, cash reserves,

and equity risk of firms before and after actual share repurchases to try to distinguish between

the information-signaling and free cash flow hypotheses. The former hypothesis argues that

deliberate cash payouts signal superior future performance; the latter states that distributing

excess cash curtails value-destroying investment.

       Numerous studies have tested the information-signaling and free cash flow hypotheses

to explain the positive market reaction to share repurchases. Most studies on tender-offer

repurchases, where a firm specifies a price (range) to repurchase a fixed number of shares

during a certain period, favor the information-signaling hypothesis (e.g. Vermaelen 1981,

Dann 1981, Howe et al. 1992, Perfect et al. 1995). Only Nohel and Tarhan (1998) find

evidence supporting the free cash flow hypothesis. In contrast to prevailing support for the

information-signaling hypothesis in studies of tender-offer repurchases, there is little

consensus on the information content of open-market share repurchases, which involve

flexible repurchase volumes and timing. Bartov (1991), Comment and Jarrell (1991), and Lie

(2005) examine announcements of US open-market repurchase intentions and favour the

information-signaling hypothesis; investigating the same events, Jagannathan and Stephens

(2003) and Grullon and Michaely (2004) support the free cash flow hypothesis. These studies

agree that share repurchases convey favourable information to the market but disagree about

what causes the observed value changes.

       The non-committal nature of repurchase intention announcements adds to the

difficulty of attributing the information content of open-market share repurchases. Unlike a

tender offer repurchase announcement, an open-market repurchase intention announcement



                                                                                              3
does not commit the firm to repurchase shares.             Therefore, a repurchase intention

announcement is a misleading indicator of whether a firm actually repurchases shares. A

sample of firms announcing repurchase intentions may include firms that subsequently do not

buy back shares. Moreover, given that repurchase intention announcements cannot lead to a

precise measure of actual cash payouts (Jagannathan et al. 2000, Stephens and Weisbach

1998), evidence showing a statistical relation between post-repurchase operating performance

and the size of repurchase intentions can be invalid. To mitigate data biases in repurchase

intention announcements, Lie (2005) examines the operating performance of US firms with

and without actual repurchases in the quarter of the repurchase intention announcement. In

contrast to Jagannathan and Stephens (2003) and Grullon and Michaely (2004), who find no

improvement in profitability after US repurchase intention announcements, Lie concludes that

post-event operating performance improves only for firms with actual share repurchases.

Lie‘s results highlight the importance of actual cash payout data in repurchase studies.

       While it is difficult for studies of US repurchases to acquire accurate execution data,

UK repurchase disclosure regulations mean that actual cash payout data are available for UK

companies via repurchase execution announcements.           We examine a sample of 5,159

repurchase execution announcements by 301 UK firms over 569 firm–years during September

1997 to July 2003. To facilitate comparison with past studies, we follow the approach of

Grullon and Michaely (2004) in analysing unexpected changes in operating performance,

capital expenditures, cash reserves, and equity risk after repurchase executions.

       Overall, our results support the free cash flow hypothesis in two respects. First, the

profitability of repurchasing firms does not improve after actual share repurchases. Instead,

capital expenditures and cash reserves decrease.       Second, we find no relation between

repurchase size and post-event operating performance. Both results suggest that repurchase




                                                                                            4
execution announcements do not contain privately held information about firms‘ superior

future performance.

        Our paper extends past studies in several respects. First, it is the first study testing the

free-cash flow and information-signaling hypotheses using UK data on share repurchases.

The UK has one of the largest stock markets in the world and between 1980 and 1998, UK

firms made approximately 60% of all share repurchase disclosures in Europe (Rau and

Vermaelen 2002). 1 Second, by examining UK data, we provide out-of-sample evidence

complementing studies using US data. Third, events in our study are announcements of

repurchase executions, not of repurchase intentions. For an individual firm, the former events

can repeat over time at irregular intervals, whereas the latter events are relatively isolated.

Because of this difference, we use accounting periods to separate repurchase execution

announcements and partition our sample into firms with announcements in single and

multiple firm–years. This partition allows us to explore differences between single and

multiple repurchasers, effectively distinguishing firms making small and large payouts—and

mitigates potential problems of defining a base year and associating subsequent performance

changes with a specific repurchasing year. Finally, UK repurchase execution announcement

data include repurchase prices and volumes, eliminating uncertainty in estimating cash

payouts via share repurchases and facilitating tests of the information-signaling prediction of

a positive relation between post-event operating performance and the percentage of

outstanding shares repurchased.



1. Repurchase regulation, past research and testable implications

1.1. Repurchase regulation


1
 The figure of 60% is probably an underestimate as Oswald and Young (2004a) report that the SDC database,
which Rau and Vermaelen use, underestimates UK repurchase intention announcements by over 100 percent and
underestimates the value of UK share repurchases by almost 400 percent.


                                                                                                       5
         UK repurchase regulations are stricter than US counterparts in three dimensions:

repurchase legislation, the authority required for repurchase, and the disclosure of repurchase

activity. It was not until 1981 that the Companies Act made share repurchases legal in the

UK. Moreover, before 1 December 2003 firms had to cancel repurchased shares and could

not hold them as treasury stock for re-sale. Due to this restriction, Oswald and Young

(2004b) suggest that before 1 December 2003, UK share repurchases are likely to be

independent of employee share option schemes because a firm could not transfer its

repurchased shares to such schemes. 2 In contrast, US firms have always been able to

repurchase shares and to retain them for re-issue. Consequently, US firms are more likely to

repurchase shares for motives such as funding employee share option programs or offsetting

stock option dilution (Kahle 2002, Bens et al. 2002, Bens et al. 2003, Hribar et al. 2004).

         Before executing a repurchase transaction, a UK firm must have articles of association

permitting repurchases and a special resolution conferring repurchase authority. A special

resolution requires a firm to send a meeting notice to shareholders and to obtain a 75%

majority of shares voting at the meeting. In contrast, a US firm‘s board of directors can

authorize repurchases. The difference in repurchase authority between the UK and the US

suggests that a UK repurchase intention announcement is not a well-defined event, as the

public announcement may refer to one of three stages: (i) the board‘s intention to seek

shareholder approval; (ii) a repurchase resolution passed by shareholders at a general meeting;

or (iii) a firm‘s formal declaration of its repurchase intention passed by a general meeting.

      After completing a repurchase transaction, the LSE Listing Rules require a firm to

notify the Company Announcement Office (CAO) of the repurchase execution no later than



2
  Before 1 December 2003, UK firms had to establish an employee benefit trust (EBT) to acquire its own shares
to provide employees with benefits including share options. Shares acquired by an EBT do not meet the legal
definition of a repurchase. In a trust relation, the trustee is the nominal owner of trust property, and any
contractual relationship over the trust property is between the trustee and the third party. Thus, shares acquired
by an EBT are not shares repurchased by a firm establishing the EBT.


                                                                                                                6
7.30 a.m. on the next business day following the repurchase transaction.3 This is one of the

‗continuing obligations‘ originating from paragraph 15.9 of the LSE Listing Rules. The

notification must include (1) the repurchase date, (2) the number of shares repurchased, and

(3) the price paid, or the highest and lowest prices paid. After receiving the notification, the

CAO disseminates it as an announcement via information provider services (PIP services)

such as the Regulatory News Service (RNS). A real example of a repurchase execution

announcement in the RNS is,

      Tomkins PLC Purchase of Own Securities.
      TOMKINS PLC 2nd September 1997
      SHARE REPURCHASE

      Tomkins PLC announced that on 2 September 1997 it purchased for cancellation
      500,000 of its ordinary shares at a price of 296.5p per ordinary share.


      Compared to repurchase intention announcements, repurchase execution announcements

are superior data for testing the information content of share repurchases. This is because

neither a UK nor a US firm announcing a repurchase intention has any obligation to

repurchase shares. Stephens and Weisbach (1998), studying 450 US repurchase intention

announcements between 1981 and 1990, find that ‗10 percent of [their sample] firms bought

less than 5 percent of the number of shares announced, and a substantial number of firms

reacquired no shares at all‘ (p.314).         Therefore, both UK and US repurchase intention

announcements are imprecise measures of actual payouts and misleading indicators of

whether firms actually repurchase shares.

      In addition to the non-committal nature of repurchase intention announcements, US

repurchase regulations prevent accurate measurement of the actual number or value of shares

repurchased. US firms do not have to reveal repurchase executions except via standard and


3
  The CAO used to be the Regulatory News Service (RNS) owned and operated by the LSE. In 2001, the
Financial Services Authority (FSA) moved to a competitive model for company announcements. The CAO now
refers to primary information provider services (PIP services) approved by the FSA to distribute regulatory
announcements. The RNS is one of the PIP services.


                                                                                                         7
aggregate disclosures of cash spending on share repurchases in their financial statements,

reported by Compustat as data item #115. This is the data source that Lie (2005) uses.

However, Stephens and Weisbach (1998) and Jagannathan et al. (2000) point out that this

cash spending overestimates open-market repurchases because the data include other capital

transactions, such as conversions of other classes of stock into common stock, stock

retirements and redemptions, privately negotiated repurchases, and self-tender offer

repurchases.     Given the difficulty in acquiring accurate repurchase data, US repurchase

studies can only estimate actual shares repurchased within upper and lower bounds (Stephens

and Weisbach 1998; Jagannathan et al. 2000). 4 Measurement errors in payout sizes may

invalidate tests of the information-signaling and free cash flow hypotheses.5



1.2. Past research and testable implications

        Miller and Modigliani (1961) prove the irrelevance of cash payout to firm value under

perfect market assumptions. Relaxing the perfect market assumptions to let managers be

better informed, they suggest that payout policy can reveal unrecognised firm value. In the

presence of information asymmetry between investors and managers, Easterbrook (1984) and

Jensen (1986) argue that managers are imperfect agents of investors and cash payouts can

mitigate agency conflicts. Consequently, it is possible to reconcile the positive association

between cash payout announcements and abnormal returns found in numerous empirical

studies by asymmetric information and agency costs of free cash flow.

        Vermaelen (1981) is the first study using information signaling to explain why tender-

offer repurchase announcements attract significant positive abnormal returns. Subsequent



4
  Stephens and Weisbach (1998) report that US firms on average complete 74% to 82% of announced
repurchases during 1981 to 1990; Jagannathan et al. (2000) find a ratio of 53% to 72% between 1985 and 1996.
5
  Grullon and Michaely (2004) use the number of shares US firms announce they intend to repurchase to
measure the dissipative cost of share repurchases. This variable induces measurement error as actual
repurchases can differ substantially from announced intentions and repurchase executions can lag announced


                                                                                                          8
studies, including Ofer and Thakor (1987) and Persons (1997), permit managers to choose

between dividends and tender-offer repurchases as signaling tools, to explain stylised facts of

tender-offer repurchases. Bagnoli et al. (1989) model a manager who obstructs a takeover by

repurchasing shares to persuade shareholders not to tender shares to hostile bidders. Despite

differences in design, the information-signaling models of tender-offer repurchases have the

same underlying argument that wealth transfers from non-tendering shareholders to tendering

shareholders if true stock prices are below repurchase prices.                   A firm without private

information about superior prospects would not execute a tender-offer repurchase as tender-

offer repurchase prices are on average 22.5% higher than prevailing market prices (Masulis

1980, Dann 1981, Vermaelen 1981). The information-signaling hypothesis, therefore, has

two testable implications. First, operating performance after repurchases exceeds market

expectations because observable cash payout conveys information about unexpected future

performance prospects to the market (Bhattacharya 1979, Miller and Rock 1985). Second, the

unexpected improvement in operating performance is positively related to the size of share

repurchases. Payout levels imply managerial confidence about operating prospects and a

larger payout increases signaling costs, discouraging bad firms from making similar payouts.

Bartov (1991), Jagannathan and Stephens (2003), Grullon and Michaely (2004), and Lie

(2005) among others, test one or both of these implications using data on US open-market

repurchase intention announcements. In addition, Lie (2005) suggests that (US) firms take

advantage of inside information about future performance when executing repurchases. This

means that post-event operating performance improves only for firms with actual share

repurchases. Using US repurchase execution data from Compustat and quarterly accounting

data, Lie shows that firms repurchasing more than 1% of the market value of equity in the

quarter of the repurchase intention announcement have better operating performance than


intentions by several weeks, months, or years (Cook et al. 2004, Stephens and Weisbach 1998, Jagannathan et al.
2000).


                                                                                                             9
control firms.      This performance improvement persists for at least two years after the

repurchase intention announcement quarter.

        Although Lie (2005) finds that US repurchasing firms outperform their peers, his

findings cannot exclude the free cash flow hypothesis for three reasons. First, as explained

above, his repurchase data are likely to be inaccurate.                 Second, Lie finds that operating

performance relative to the previous year‘s level decreases significantly after repurchase

intention announcements whether firms actually repurchase shares or not: repurchasing firms

only outperform their industry-, performance-, and market-to-book-matched control firms.

Finally, as Lie does not decompose the improvement in adjusted operating performance into

its components, the source of the improvement is unclear. Nohel and Tarhan (1998) find that

firms with limited investment opportunities improve their operating efficiency by re-

deploying assets and returning proceeds of asset sales via share repurchases. Performance

improvement from asset re-deployment is more in the spirit of the free cash flow hypothesis

than the information-signaling hypothesis (Nohel and Tarhan 1998, p189).6

        Shrinking firm assets and distributing the proceeds to increase operating efficiency is

consistent with the solutions suggested by numerous studies to overcome the over-investment

problem caused by free cash flow. In general, if shareholders can restrict the assets under

management control, it is harder for management to over-invest in negative net present value

projects, to consume perquisites, or to be slothful (Grossman and Hart 1980, Easterbrook

1984, Jensen 1986, Stulz 1990). In particular, Jensen (1986) argues that share repurchases

can take surplus cash from the firm, reducing potential over-investment and increasing firm

value. This implies that firms repurchase shares to distribute free cash flow when they face


6
  Nohel and Tarhan (1998) decompose cashflow return on assets into cashflow margin (cashflow divided by
sales) and asset turnover (sales divided by the market value of assets). An increase in cashflow margin indicates
better control over costs or more profitable products, supporting the information-signaling hypothesis. Higher
asset turnover suggests efficient use of assets, favoring the free cash flow hypothesis. Nohel and Tarhan (1998)
find that firms with limited growth opportunities improve their operating performance through increases in asset
turnover associated with asset sales.


                                                                                                             10
limited investment opportunities. Consequently, the free cash flow hypothesis predicts that

capital expenditures and cash reserves are lower after repurchases.



2. Data and sample selection

        We collect data on repurchase executions announced by UK listed firms between

September 1997 and July 2003 from the database Company REFS.7,8 Each announcement

contains an announcement date, a repurchase price, and the number of shares repurchased.

We verify more than 3% of the 9,020 repurchase execution announcements by checking with

original announcements in the RNS of the LSE and Factiva and confirm that the data in

Company REFS is an accurate abstract of the announcements.

        Our final sample comprises 5,159 repurchase execution announcements by 301 firms.

Table 1 summarises our five sample selection criteria. First, we only examine announcements

of ordinary share repurchase executions.              Second, announcements must have complete

information on announcement date, repurchase price, and the number of shares repurchased.

Third, we exclude firms that lack Datastream codes, which are essential to retrieve necessary

data. Fourth, following Oswald and Young (2004b) we exclude closed-end investment trusts.

Finally, to reduce potential errors in identifying firms, we measure the ratio of repurchase

price to unadjusted closing price on the announcement day and delete the top 0.5% and

bottom 1%. 9 In addition to the five sample selection criteria, we combine announcements

released by the same firm on the same day (since some UK firms disclose more than one

repurchase execution on the same day).




7
  Company REFS, published by HS Financial Publishing Ltd (formerly Hemmington Scott LTD), covers all UK
quoted companies.
8
  To check that the abolition of Advanced Corporation Tax (ACT) in the UK in April 1999 drives our results, we
repeat our subsequent analysis on the sub-sample of repurchase execution announcements made after the
abolition of ACT. This analysis confirms the results we report in Tables 3–9 below.
9
  Deleting the bottom 1% (0.5%) results in a ratio ranging between 0.8772 (0.5019) and 1.0871.


                                                                                                           11
        The 5,519 repurchase execution announcements cover 569 firm–years and 301 firms.

This shows that some firms repurchase shares in more than one financial year and raises a

potential problem of defining a base year and associating performance changes with a specific

repurchase year. To mitigate this problem and to explore differences between firms that

repurchase shares in one and multiple years, we define a single repurchaser as a firm with

repurchase execution announcements in one accounting year and a multiple repurchaser as a

firm with repurchase execution announcements in more than one accounting year during our

sample period. 10 Previous research shows that firms consider accounting performance,

particularly earnings per share, when making repurchase decisions (Kahle 2002, Bens et al.

2002, Bens et al. 2003, Hribar et al. 2004). Thus, we assume that a firm‘s accounting period

acts as a boundary separating its announcements.

        Table 2 reports descriptive statistics for 128 single repurchasers and 149 multiple

repurchasers. 11 Panel A shows that 44 (34.38%) single repurchasers have repurchase

execution announcements spread over a period longer than the minimum length (28 trading

days) of multiple repurchasers‘ time spans and 132 (88.59%) multiple repurchasers have

announcement time spans greater than the maximum time span (179 trading days) of single

repurchasers. 12 This means that most single repurchasers have shorter announcement time

spans than multiple repurchasers. Panel B shows that multiple and single repurchasers differ



10
   Defining all firm–years with repurchase execution announcements as year 0, a test may inappropriately align a
firm‘s multiple repurchasing firm–years as base years. To mitigate the potential problem in defining base years,
we perform additional tests using only the first firm–year with announcements over the sample period. These
results confirm the results we report so we do not separately tabulate them. Grullon and Michaely (2004)
perform their empirical analysis using a full sample and a sub-sample comprising only the first repurchase
intention announcement of a firm during their sample period, and also report that results of the two samples are
qualitatively the same.
11
   For single repurchasers, we use Worldscope data item 04751 (purchase of common and preferred stock) to
identify other repurchase activity during 1994 to 2004. As a result, we remove 24 firms (with 143
announcements) who are single repurchasers but have other repurchase activity indicated by this item. We do
not treat the 24 firms as multiple repurchasers, because item 04751, similar to Compustat data item #115,
includes various capital transactions such as repurchase of preferred stock and exchange of common stock for
debentures. However, we do retain the 24 firms in our ―All repurchasers‖ sample in subsequent tables.
12
   The time span of a firm‘s repurchase execution announcements is the number of trading days from the first to
the last announcements of the firm over the sample period.


                                                                                                            12
in their repurchase behavior: besides having longer time spans, multiple repurchasers acquire

more shares with higher execution frequencies than single repurchasers.

        Panel C of Table 2 provides some evidence that multiple repurchasers have lower

market-to-book ratios (MTB) and higher free cash flow. 13,14 To the extent that MTB proxies

for investment opportunities (Lang and Litzenberger 1989), multiple repurchasers have less

investment opportunities than do single repurchasers. Together with a higher free cash flow,

the evidence suggests that multiple repurchasers buy back shares to distribute excess cash.

On the other hand, MTB can also proxy for information asymmetry (Barth and Kasznik

1999), in which case high MTB suggests that single repurchasers have more privately held

information and are more likely to repurchase shares for signaling reasons.



3. Operating performance following repurchase executions

        In this section, we examine changes in operating performance due to repurchases

following Grullon and Michaely (2004). We then test the information-signaling hypothesis,

which predicts that cash payouts reveal privately held information about future performance.



3.1. Univariate statistics

        We use return on assets (ROA) as a primary measure of operating performance. 15

ROA is income scaled by the average of opening and closing book value of total assets,

                                          EBITDAt
                                                               ,                                       (1)
                                    ( ASSETt  ASSETt 1 ) / 2




13
   MTB is market value of equity divided by book value of equity at the year-end before the repurchase
execution announcements year.
14
   Following Lehn and Poulsen (1989), we measure free cash flow as after-tax cash flow not distributed to
security holders as either interest or dividend.
15
   To test the robustness of results using ROA, we use three other measures: return on cash-adjusted assets
(ROCAA), return on sales (ROS), and cash-flow return on assets (CFROA). We do not present these robustness
checks, but they are available on request, and we refer to the results below.


                                                                                                        13
EBITDA t = annual income before interest, taxes, depreciation and amortization = pre-tax

profit (Datastream item 154) + total interest charges (item 153) + depreciation (item 136) +

amortization of intangibles (item 975), and

ASSETt = book value of total assets (item 392). 16

   To detect whether repurchasing firms experience unexpected changes in operating

performance, we use two benchmarks to measure the unexpected change in operating

performance after share repurchases: the change in ROA and the difference in ROA between

the repurchasing firm and a matched non-repurchasing firm. The second benchmark matches

a repurchasing firm with a non-repurchasing firm in the same year based on combinations of

the following four factors, stated in order of precedence,

   1) Change in ROA in year 1 ( ROA1 ), where year 0 is the repurchase year,

   2) Datastream‘s level 4 industrial classification,

   3) ROA in year 1 ( ROA1 ), and

   4) Market-to-book ratio in year 1 ( MTB 1 ), which is market value of equity divided by

       book value of equity (Datastream item 305).

   The first factor recognises that earnings momentum may exist before and persist after

share repurchases. Barber and Lyon (1996) suggest that test statistics using the change in a

firm‘s profitability relative to a control firm are consistently more powerful than test statistics

using the level of a firm‘s profitability relative to its control. The second and fourth factors

are common in matching firms to measure unexpected changes in operating performance.

The third factor recognises that earnings and profitability tend to mean revert.

   Based on these factors, we select control firms according to six progressively weaker

criteria stated below, where percentages are deviations from the repurchasing firms‘ values.

Criteria 4 to 6 repeat criteria 1 to 3 but relax the restriction of the same industry code.




                                                                                                14
       1) Change in ROA 20%, same industrial code, level of ROA 20%, MTB 20%.

       2) Change in ROA 20%, same industrial code, level of ROA 20%.

       3) Change in ROA 20%, same industrial code.

       4) Change in ROA 20%, level of ROA 20%, MTB 20%.

       5) Change in ROA 20%, level of ROA 20%.

       6) Change in ROA 20%.

Apart from the six criteria, the comparison group consists of UK firms, including live and

dead firms after September 1997, included in the LSPD with no share repurchase activity

during 1994–2004. We check for the absence of share repurchases between 1994 and 2004

using Worldscope data item 04751, in addition to our repurchase sampling criterion. The size

of the control firm group is 2,055 firms. If a sample firm has more than one matching firm,

we select a matching firm satisfying the criterion,

                               mini | ROA1, sample firm  ROA1, matching firm i |
                                          | ROA1, sample firm  ROA1, matching firm i |         (2)
                                          | MTB1, sample firm  MTB1, matching firm i | .

Lie (2001) shows that this performance-adjusted benchmark yields more powerful test

statistics than other benchmarks.

           Table 3 reports changes in ROA and cumulative changes in ROA for repurchasing

firms. Results on unadjusted changes in Panel A show that repurchasing firms have reduced

profitability in the repurchasing year, and they experience a further decline in profitability in

the year after: mean unadjusted change in ROA is 0.87% in year 0 and 1.28% in year 1.

Panel B shows that over the full test period 2 to 3, the mean change in ROA for the sample

of all repurchasers is 6.06%, significant at 1%. Mean cumulative unadjusted abnormal

changes in ROA over the pre-repurchase (2 to 0) and post-repurchase (0 to 3) periods are


16
     These and subsequent references to (Datastream) item numbers relate to pre-Worldscope data.


                                                                                                   15
1.45% and 4.94%, both significant at 1%. A paired t-test shows that unadjusted cumulative

abnormal change in ROA deteriorates from the pre- to the post-repurchase periods, with a

difference of 4.88%, significant at 1%.                  This suggests that repurchase execution

announcements do not foreshadow increases in ROA. These results extend to the single and

multiple repurchaser sub-samples. In fact, 29 of the 30 mean and median changes in ROA in

Panel A are negative, 19 significantly so, all mean and median cumulative changes in Panel B

are significantly negative, and deterioration in ROA is generally significantly greater in the

post- than in the pre-repurchase period. The only difference between single and multiple

repurchasers is that, for each cumulative abnormal change, the former underperform the latter,

particularly in the post-repurchase period—while multiple repurchasers have mean post-

repurchase cumulative unadjusted change in ROA of 3.73%, single repurchasers have a

corresponding figure of 12.37%. These results are generally consistent with Grullon and

Michaely (2004) and Lie (2005). The right side of Table 3 reports performance-adjusted

changes in ROA. Panel A provides some evidence that ROAs of repurchasing firms increase

by more than those of their control firms in the repurchasing year 0. However, there is no

evidence of repurchasing firms outperforming their peer firms in any of the three years

following share repurchases, and no apparent difference between single and multiple

repurchasers. 17 Panel B shows that cumulative adjusted changes in ROA over the full test

period 2 to 3 are insignificant.18 Separating pre- and post-repurchase periods, there is some

evidence pre-repurchase of repurchasers‘ ROAs falling by less than those of their matching

firms, but not post-repurchase.            This, again, suggests that share repurchases do not

foreshadow better operating performance.




17
   It seems that single repurchasers have better pre-repurchase operating performance and worse post-repurchase
operating performance than multiple repurchasers.
18
   The median cumulative adjusted change in ROA for single repurchasers is 9.269%, significant at 10%.


                                                                                                            16
        Three conclusions emerge from Table 3. First, repurchasing firms‘ ROAs decline

over time. Moreover, cumulative unadjusted abnormal changes in ROA deteriorate from the

pre- to the post-repurchase periods. The difference in cumulative changes in ROA between

the pre- and post-repurchase periods is significant at 1%. Second, despite declining ROAs,

repurchasing firms generally maintain ROAs similar but not superior to their peer firms in the

post-event period. Finally, there is evidence of repurchasing firms outperforming their peer

firms in the repurchasing year. Cumulative changes in ROA between repurchasers and non-

repurchasers are significant in the pre-repurchase period but not over the whole or post-

repurchase periods.      Similar to Grullon and Michaely (2004), we conclude that UK

repurchase execution announcements may contain information about current changes in ROA

but not about future changes.19



3.2. Regression analysis

        In this subsection we test the implication of the information-signaling hypothesis that

a larger percentage of shares repurchased indicates better future performance. Assuming

shares repurchased scaled by outstanding shares proxies for dissipative costs, we test this

implication using the following cross-sectional regression,

                          POSTROAi   0  1 AG _ REPi   2 PREROAi   i ,                     (3)

where POSTROAi is the median over the post-repurchase period (years 1 to 3) of differences

in (changes or levels of) ROA between repurchasing and control firms, PREROAi is the

corresponding median over the pre-repurchase period covering years 2 through 0, 20 and

AG _ REPi is the aggregate ratio of shares repurchased in each announcement to outstanding

shares one day before over a financial year.


19
  We conduct robustness checks for the results in Table 3 by replacing ROA with ROCAA, ROS, and CFROA,
and find the results lead to similar conclusions.


                                                                                                   17
        Depending on model specification, POSTROA and PREROA i in equation (3) relate
                                                 i



to changes or levels of ROA. As Healy et al. (1992) suggest, pre- and post-repurchase

operating performances may be correlated. Prior evidence also indicates that profitability is

mean reverting (e.g. Brooks and Buckmaster 1976, Fama and French 2000). Therefore, with

equation (3) specified in terms of levels,  2 captures the correlation between pre- and post-

repurchasing years, while with equation (3) specified in terms of changes,  2 measures the

magnitude of mean reversion in ROA.

        Controlling for correlation or mean reversion, the test centres on  0 and  1 . A

positive  0 is consistent with both the free cash flow and information-signaling hypotheses. 21

 1 in equation (3) measures the association between post-repurchase ROA and the magnitude

of dissipative costs (the aggregate percentage of outstanding shares repurchased by a firm

over a financial year). A positive  1 supports the information-signaling hypothesis.

        Equation (3), which excludes dividends, may be sufficient to capture the signaling

effect of share repurchases, as Lintner (1956) suggests that dividends tend to be smooth over

time. Similarly, Fama and French (2001) and DeAngelo et al. (2004) find that dividend

payers use share repurchases to increase the overall fraction of distributed earnings.

However, if share repurchases substitute for dividends, as argued by Grullon and Michaely

(2002), a model including dividends better captures the marginal signaling effect of share

repurchases. Therefore, we provide robustness checks of equation (3) by estimating the

following two cross-sectional regressions,

                   POSTROAi   0  1 V_REPi   2 DIVi  3 PREROAi   i ,                             (4)


20
   We test the robustness of our results by excluding year 0 from the repurchase period. This robustness check
confirms the results we report below.
21
   In the free cash flow hypothesis, post-repurchase operating performance increases if share repurchases
mitigate conflicts of interest over excess cash.


                                                                                                           18
                       POSTROAi   0  1 PAYOUTi   2 PREROAi   i ,                         (5)

where V_ REPi is the value of shares repurchased in a firm–year divided by market value of

equity at the start of the repurchasing year, DIV i is Datastream item 187 (total dividends paid

on ordinary and participating preference shares) divided by market value of equity at the start

of the repurchasing year, and PAYOUT i is the sum of V_ REPi and DIV i .

           Table 4 reports descriptive statistics for the variables in equations (3)–(5), and

suggests there is no serious collinearity between the regressors.                 Although V_ REPi ,

AG _ REPi , and PAYOUT i are highly correlated, they do not cause a collinearity problem

because they appear in different equations. Table 4 also indicates that most of the variation in

PAYOUT i comes from variation in share repurchases. In addition, Table 4 shows that pre-

event changes in ROA are insignificant, because repurchasing and non-repurchasing firms are

matched on similar changes in operating performance in year 1.

           Table 5 report the results of estimating equations (3)–(5). Panel A reports results for

equation (3) using all firm–years with repurchase execution announcements as year 0. The

intercepts suggest that multiple repurchasers‘ ROAs increase, on average, by 5% per year in

the post-repurchase period.                In contrast, there is no significant increase for single

repurchasers. 22 A factor explaining the increase in multiple repurchasers‘ ROAs is their

larger cash payouts. 23 Just as Nohel and Tarhan (1998) suggest that (tender-offer) share

repurchases can be part of a restructuring package, firms that repurchase shares in multiple

years have more opportunity to re-deploy assets or adjust operating strategies. If the 5%

improvement in ROA results from operating efficiencies linked to reduced asset size rather

than from positive NPV projects, these results provide more support for the free cash flow

hypothesis than the information-signaling hypothesis.


22
     The smaller sample size of single repurchasers contributes to this result.


                                                                                                 19
        More direct evidence contradicting the information-signaling hypothesis is that the

percentage of outstanding shares repurchased provides little information in predicting superior

changes in post-repurchase ROA. Panel A shows either a negative or no relation between

post-repurchase changes in ROA and AG_ REPi , contradicting the information-signaling

hypothesis and consistent with the conclusion derived from Table 3.24

        Panel B provides a robustness check on Panel A by replacing AG _ REPi with

V_ REPi and DIV i . Panel C provides a further robustness check by combining V_ REPi and

DIV i into PAYOUT i . Using ROA levels as a performance measure, Panel A shows some

significantly negative coefficients on AG _ REPi with dividends excluded, particularly for

multiple repurchasers, but Panel B, which includes DIV i , shows that V_ REPi is no longer

significant. 25 In Panel C the coefficients on PAYOUT i are all insignificant, suggesting cash

payouts contain no information about future ROA performance. 26



4. Changes in investment opportunities, cash holdings, and risk

        The evidence so far favors the free cash flow hypothesis, that firms repurchase shares

to distribute free cash flow when they face limited investment opportunities. Therefore, we

examine whether firms have fewer investment opportunities and lower cash holdings after

repurchasing shares.       We then test whether share repurchases reveal information about

changes in equity risk.




23
   Table 2 shows that multiple repurchasers buy back a greater percentage of outstanding shares than do single
repurchasers.
24
   Results in Panel A are also consistent with previous studies in that profitability levels persist over time
(estimates of  2 are significantly positive in levels regressions) and changes in profitability mean revert
(estimates of  2 are significantly negative in changes regressions).
25
   From Table 4, the correlation between AG _ REPi and V_ REPi is 0.949, which suggests the two variables
are interchangeable.


                                                                                                           20
4.1. Investment opportunities and cash holdings

        To detect changes in investment opportunities and cash holdings, we examine (1)

capital expenditures (item 1024—cash paid for tangible fixed assets during a year), and (2)

cash reserves (item 375—total cash and equivalent). We scale both measures by the average

of opening and closing book value of total assets. Replacing ROA with capital expenditures

(cash reserves), we use the same procedure as in the previous section to model unexpected

changes in capital expenditures (cash reserves).          Table 6 shows results for changes and

cumulative changes in capital expenditures; Table 7 shows corresponding results for cash

reserves.

        Panel A of Table 6 shows that unadjusted capital expenditures of all repurchasers

decrease over time. Median unadjusted change in capital expenditures is 0.11% in year 1,

the rate of decline accelerating monotonically to 0.30% in year 3. All of these median

changes are significant at 1% or 5%. Cumulative unadjusted changes in capital expenditures

in Panel B are also significantly negative over the whole, pre-, and post-repurchase periods.

Moreover, Panel A of Table 7 shows that unadjusted changes in cash reserves of all

repurchasers increase before share repurchases (year 1) but decrease in the year after share

repurchases (year 1).27 These changes in cash reserves seem to be temporary, as Panel B

shows that cumulative unadjusted changes in cash reserves are statistically zero over the three

periods.    These results suggest that repurchasing firms have a shrinking investment

opportunity set and repurchase shares to reduce temporary cash surpluses, consistent with

Grullon and Michaely (2004) and Oswald and Young (2004b).

        Compared with matching firms, Panels B of Tables 6 and 7 show that all repurchasers

have a marginal increase in capital expenditures before share repurchases, and higher cash


26
   The overall results in Table 5 are similar when we replace ROA with the three alternative measures of
operating performance.
27
   Mean unadjusted changes are significantly positive in years 2 and 3, but median changes are not.


                                                                                                     21
holdings after share repurchases. The result is easier to interpret after partitioning the sample,

because the patterns of changes in capital expenditures and cash reserves differ between

multiple and single repurchasers. Table 6 shows some evidence of single repurchasers‘

capital expenditures falling after share repurchases, while multiple repurchasers have

insignificant changes in capital expenditures. Table 7 shows that multiple repurchasers have

higher cash reserves after share repurchases, while cash reserves of single repurchasers

decline over the post-repurchase period.        These results suggest that with investment

opportunities returning to normal and with substantial free cash flow, multiple repurchasers

buy back shares in multiple years to prevent excess cash piling up, while single repurchasers

experience declines in cash holdings and capital expenditures after share repurchases.

       The overall evidence is consistent with the characteristics of multiple and single

repurchasers, supporting the free cash flow hypothesis.          Table 2 shows that multiple

repurchasers have lower market-to-book ratios and higher free cash flows than single

repurchasers, suggesting they have fewer investment opportunities but more excess cash.

They also acquire a higher percentage of outstanding shares. Thus, it is reasonable to infer

that firms repurchase shares in multiple years to return increasing excess cash to shareholders

in the face of shrinking investment opportunities. Other things equal, multiple repurchasers

would accumulate more free cash if they repurchased fewer shares. On the other hand, share

repurchases reduce investment and cash reserves of single repurchasers, which may explain

why these firms do not repurchase shares in other years. More importantly, the reduction in

capital expenditures and cash reserves lowers the possibility that share repurchases signal

future performance prospects to the market. Given the differences in firm characteristics of

single and multiple repurchasers, it is unlikely that capital expenditures and cash reserves

decline in each year after repurchases.




                                                                                               22
4.2. Equity risk

        To examine changes in equity risk following repurchase execution announcements, we

use the model suggested by Grullon and Michaely (2004) as our primary methodology. To

test the robustness of our results, we use the approach of Bartov (1991). There are two major

differences between the primary methodology and the robustness check. First, the primary

methodology uses monthly data, whereas the robustness check uses daily data. Second, the

primary methodology uses dummy variables for the post-repurchase period to measure the

changes in equity risk due to share repurchases. The robustness check generates the market

model beta for each year and compares the market model beta with benchmark risk to

measure the change in risk.

        We use the CAPM and the Fama and French (1993) three-factor model to calculate

changes in equity risk. Let t* be the month in which a firm announces its first repurchase

execution during the sample period. For each firm with repurchase execution announcements,

we run the following regressions using 73 months (from t*  36 to t* + 36) around the first

repurchase announcement,

                   Rit  R ft    i   i Dt  b i ( Rmt  R ft )  bi Di ( Rmt  R ft )  eit ,   (6)

                   Rit  R ft   i   i Dt  bi ( Rmt  R ft )  bi Dt ( Rmt  R ft )
                                                                                                        (7)
                              si SMBt  si Dt SMBt  hi HMLt  hi Dt HMLt  eit ,

where Rit and Rmt are the monthly returns on stock i and on the FTSE All Share Index in

month t; R ft is the UK three-month Treasury bill rate converted to a monthly rate of return;

SMBt ( HML t ) is the difference in monthly returns on portfolios of small and large (high and

low book-to-market) firms;28 and Dt is a dummy variable equal to 1 if t  t * . In equations (6)

and (7), bi , s i , and hi are the factor loadings (betas) of firm i over the three years before


28
 The FTSE Small Cap and FTSE 100 Indexes proxy portfolios of small and large firms, and the FTSE Global
Value and FTSE Global Growth Indexes proxy portfolios of high and low book-to-market stocks.


                                                                                                        23
the first repurchase execution and bi , s i , and hi are the changes in the factor loadings after

the repurchase execution, the focal point of our tests.  i is the abnormal return of firm i

before the repurchase execution, and  i is the change in the abnormal return after the

repurchase execution. The tests include firms with at least 44 returns available (around 60%

of the 73 observations). 29

           We perform a robustness check using the market model with daily return data. 30

Instead of dummy variables measuring changes in betas after share repurchases, we estimate

the market model beta for each year and compare this with two benchmark risks to measure

changes in equity risk. Because this test uses daily data, estimates of the market model beta

may be biased if thin trading exists (Scholes and Williams 1977). To mitigate this non-

synchronous trading problem, we estimate Scholes–Williams betas for all firms.

           In evaluating unexpected changes in equity risk after share repurchases, we apply two

benchmarks similar to those for operating performance. In tests using monthly data, the first

benchmark takes the coefficients on the dummy variables as unadjusted changes. Using daily

data, unadjusted changes are first differences of market model betas of consecutive periods.

For the second benchmark (adjusted change), for each firm we first find control firms with

market value of equity and book-to-market ratio between 80% and 120% of the values of the

sample firm at the end of year 1. Using monthly return data, we then select the control firm

with the closest buy-and-hold stock return to that of its sample firm during the year before the

first repurchase execution announcement,

                                       12                           12

                               min i    1  R
                                       t 1
                                                t , sample firm      1  R
                                                                     t 1
                                                                             t , matching firm i              (8)


Using daily return data, we set an additional criterion to control for risk changes, that a



29
     The monthly return data end on 2 February 2005.
30
     We include regressions with a minimum of 150 observations (around 60% of total trading days in a year).


                                                                                                               24
control firm also has the closest change in equity risk in year 1 to its sample firm. 31

        Panel A of Table 8 shows that CAPM betas of repurchasing firms remain stable over

the six-year period centred on the month of their first repurchase execution announcement. 32

Results in Panel B indicate that multiple repurchasers experience a marginally significant

increase in systematic risk relative to their peer firms. Given that multiple repurchasers

acquire a greater percentage of outstanding shares than do single repurchasers, these results

are consistent with Hamada (1969) in that share repurchases increase financial leverage,

leading to higher equity risk. Panel C shows significant declines in unadjusted market beta

and SMB beta from the three factor model. This result is similar to Grullon and Michaely

(2004). Again, only multiple repurchasers experience significant changes in risk. However,

Panel D reports inconclusive results for differences in adjusted risk.33 Consequently, while

Grullon and Michaely (2004) show that US firms have significantly smaller market betas and

SMB betas in the three-year period after announcing their repurchase intentions, the overall

results of Table 8 do not fully support their findings.

        Table 9, reporting evidence from the market model, suggests two conclusions about

changes in repurchasing firms‘ risk. First, risk may decrease before actual share repurchases:

repurchasing firms experience significant declines in beta, particularly compared with firms

with similar market values of equity and BE/ME in year –1.                       This implies that firms

repurchase shares when they experience risk declines, as argued by Bartov (1991) and

Grullon and Michaely (2004). Second, beta may rise due to share repurchases. 34                       This is

consistent with Hamada (1969) and Bartov (1991): the former shows that an increasing

financial leverage increases a firm‘s risk, the latter finds that risk increases slightly after share



31
   Change in risk in year 1 is the Scholes–Williams beta in year 1 less the beta of the same firm in year 2.
32
   For unadjusted changes in Table 8, the sample of all repurchasers comprises 320 firms as we are able to
include 19 firms that have returns data but no accounting data.
33
   Market betas are higher for multiple repurchasers at the 10% level. For single repurchasers, SMB beta
increases, while HML beta decreases.


                                                                                                           25
repurchases. In all, the evidence in Tables 8 and 9 implies that firms do not repurchase shares

to signal changes in risk to the market.35 Instead, it suggests that firms repurchase shares to

re-balance risk.

          The findings in Table 9 also help to explain differences between our study and Grullon

and Michaely (2004), who find that both unadjusted and adjusted market betas and SMB

betas decline significantly during a three-year period after US repurchase intention

announcements, and argue that repurchasing firms experience declines in equity risk because

their values rely less on risky growth opportunities. 36 Our evidence does not support their

findings, probably due to different definitions of events used to separate pre- and post-

repurchase periods.       Repurchase execution announcements imply immediate increases in

financial leverage, whereas repurchase intention announcements do not. Equity risk may be

low when firms announce repurchase intentions, but increase after firms actually repurchase

shares.



5. Summary and conclusions

          Several empirical studies document a positive association between share repurchases

and abnormal stock returns. For example, Grullon and Michaely (2004) and Lie (2005) find

that the US market reacts positively to repurchase intention announcements, arguing that

these announcements convey favorable information to investors. However, the former study

concludes that the announcements contain information about past (or present) profitability

and not about the future, whereas the latter argues that firms take advantage of inside

information about future performance when engaging in repurchase transactions.                               As


34
   Restricting attention to the first firm–year of multiple repurchasers shows that adjusted betas decline in the
year after share repurchases. However, this year is likely to be another repurchasing year.
35
   The most consistent result in Tables 8 and 9 is that repurchasing firms‘ post-event risk does not decrease
compared with their control firms.
36
   Grullon and Michaely (2004, 652) argue that ‗repurchases may be associated with a firm‘s transition from a
higher growth phase to a lower growth phase.‘


                                                                                                              26
highlighted by Lie (2005), actual cash payout data are important in clarifying the

disagreement on the information content of open-market repurchases because the information-

signaling and free cash flow hypotheses base their arguments on actual cash payouts rather

than on non-committal announcements of payout intentions.

       Using 569 firm–years with at least one repurchase execution announcement released

by 301 UK firms, we test changes in operating performance, capital expenditures, cash

reserves, and equity risk to distinguish between the information-signaling and free cash flow

hypotheses. Each test uses two benchmarks to measure unexpected changes due to share

repurchases. In addition, we test for a positive relation between post-repurchase operating

performance and the magnitude of share repurchases, as predicted by the information-

signaling hypothesis.   To mitigate the problem in defining a base year and associating

subsequent performance changes with a specific repurchasing year, we partition all

repurchasers into single and multiple repurchasers, based on frequencies of firm–years with

repurchase execution announcements.         This sample partition also contrasts multiple

repurchasers with lower market-to-book ratios, higher free cash flow, and greater percentage

of outstanding shares repurchased than single repurchasers.

       Our findings support the free cash flow hypothesis along several dimensions. First,

multiple repurchasers have a diminishing investment opportunity set and repurchase a

substantial number of shares to avoid increases in cash holdings.          Single repurchasers

experience a reduction in cash reserves immediately after share repurchases, followed by a

reduction in capital expenditures two years after share repurchases.          This means that

repurchasing firms have reduced investment opportunities and reduce financial slack.

Second, repurchasing firms‘ profitability falls from two years before to three years after share

repurchases. They outperform control firms only before and during the repurchasing year but

not after. We find positive cumulative abnormal changes in profitability in the pre-repurchase



                                                                                             27
period but not in the post-repurchase period, indicating no improvement in operating

performance after share repurchases. If firms repurchase shares to signal superior operating

performance in the past, it is unclear why other methods such as financial reports fail to

transmit this information to the market.      Finally, the percentage of outstanding shares

repurchased does not predict unexpected future operating performance, contradicting

signaling models that generally require good firms to repurchase a larger percentage of

outstanding shares to demonstrate optimistic beliefs about the future and to prevent bad firms

from mimicking them.

       Understanding whether operating performance changes following share repurchases is

important not only because of the increasing popularity of share repurchases but also because

of wider implications. For example, we find that share repurchases do not contain valuable

asymmetric information about unexpected future performance prospects.                 Instead,

repurchasing firms may be committed to distributing excess cash (Jagannathan and Stephens

2003, Grullon and Michaely 2004, Oswald and Young 2004b) or may be engaged in re-

deploying assets (Nohel and Tarhan, 1998). Given that information asymmetry is the premise

of misvaluation, results in this study also suggest an explanation of the erosion of stock

performance before repurchase (intention) announcements documented in the literature (e.g.

Stephens and Weisbach 1998, Kahle, 2002).         While prior studies interpret poor return

performance before share repurchases as evidence of misvaluation and firms‘ opportunistic

behaviour (e.g. Stephens and Weisbach 1998, Ikenberry et al. 2000), our results suggest it

may be market discipline for firms‘ reluctance to pay out excess cash, to which managers only

belatedly respond. When they do respond, they do so through share repurchases, which carry

no commitment to future cash disbursements.




                                                                                           28
References
Bagnoli, M., Gordon, R., Lipman, B.L., 1989. Stock repurchase as a takeover defense.

    Review of Financial Studies 2, 423–443.

Barber, B.M., Lyon, J.D., 1996. Detecting abnormal operating performance: the empirical

    power and specification of test statistics. Journal of Financial Economics 41, 359–399.

Barth, M.E., Kasznik, R., 1999. Share repurchases and intangible assets. Journal of

    Accounting and Economics 28, 211–241.

Bartov, E., 1991. Open-market stock repurchases as signals for earnings and risk changes.

    Journal of Accounting and Economics 14, 275–294.

Bens, D.A., Nagar, V., Skinner, D.J., Wong, M.H.F., 2003. Employee stock options, EPS

    dilution, and stock repurchase. Journal of Accounting and Economics 36, 51–90.

Bens, D.A., Nagar, V., Wong, M.H.F., 2002. Real investment implications of employee stock

    option exercises. Journal of Accounting Research 40, 359–393.

Bhattacharya, S., 1979. Imperfect information, dividend policy, and "the bird in the hand"

    fallacy. Bell Journal of Economics 10, 259–270.

Brooks, L.D., Buckmaster, D.A., 1976. Further evidence of the time series properties of

    accounting income. Journal of Finance 31, 1359–1373.

Comment, R. and Jarrell, G.A., 1991. The relative signaling power of Dutch-auction and

    fixed-price self-tender offers and open-market share repurchases. Journal of Finance 46,

    1243–1271.

Cook, D.O., Krigman, L., Leach, J.C., 2004. On the timing and execution of open market

    repurchases. Review of Financial Studies 17, 463–498.

Dann, L.Y., 1981. Common stock repurchases: An analysis of returns to bondholders and

    stockholders. Journal of Financial Economics 9, 113–138.




                                                                                              29
DeAngelo, H., DeAngelo, L., Skinner, D.J., 2004. Are dividends disappearing? Dividend

    concentration and the consolidation of earnings. Journal of Financial Economics 72, 425–

    456.

Easterbrook, F.H., 1984. Two agency-cost explanations of dividends. American Economic

    Review 74, 650–659.

Fama, E.F., French, K.R., 1993. Common risk factors in the returns on stocks and bonds.

    Journal of Financial Economics 33, 3–56.

Fama, E.F., French, K.R., 2000. Forecasting profitability and earnings. Journal of Business 73,

    161–175.

Fama, E.F., French, K.R., 2001. Disappearing dividends: changing firm characteristics or

    lower propensity to pay? Journal of Financial Economics 60, 3–43.

Grossman, S.J., Hart, O.D., 1980. Takeover bids, the free–rider problem, and the theory of the

    corporation. Bell Journal of Economics 11, 42–64.

Grullon, G., Michaely, R., 2002. Dividends, share repurchases, and the substitution

    hypothesis. Journal of Finance 57, 1649–1684.

Grullon, G., Michaely, R., 2004. The Information content of share repurchase programs.

    Journal of Finance 59, 651–680.

Hamada, R.S., 1969. Portfolio analysis, market equilibrium and corporation finance. Journal

    of Finance 24, 13–31.

Healy, P.M., Palepu, K.G., Ruback, R.S., 1992. Does corporate performance improve after

    mergers? Journal of Financial Economics 31, 135–175.

Howe, K.M., He, J., Kao, G.W., 1992. One-time cash flow announcements and free cash-flow

    theory: share repurchases and special. Journal of Finance 47, 1963–1975.

Hribar, P., Jenkins, N.T., Johnson, W.B., 2004. Stock repurchases as an earnings management

    device, Working paper, Cornell University.



                                                                                            30
Jagannathan, M., Stephens, C., 2003. Motives for multiple open-market repurchase programs.

    Financial Management 32, 71–91.

Jagannathan, M., Stephens, C.P., Weisbach, M.S., 2000. Financial flexibility and the choice

    between dividends and stock repurchases. Journal of Financial Economics 57, 355–384.

Jensen, M.C., 1986. Agency costs of free cash flow, corporate finance, and takeovers.

    American Economic Review 76, 323–329.

Kahle, K.M., 2002. When a buyback isn't a buyback: open market repurchases and employee

    options. Journal of Financial Economics 63, 235–261.

Lang, L.H.P., Litzenberger, R.H., 1989. Dividend announcements: cash flow signaling vs.

    free cash flow hypothesis? Journal of Financial Economics 24, 181–191.

Lehn, K., Poulsen, A., 1989. Free cash flow and stockholder gains in going private

    transactions. Journal of Finance 44, 771–787.

Lie, E., 2005. Operating performance following open market share repurchase announcements.

    Journal of Accounting and Economics 39, 411–436.

Lintner, J., 1956. Distribution of incomes of corporations among dividends, retained earnings,

    and taxes. American Economic Review 46, 97–113.

Masulis, R.W., 1980. The effects of capital structure change on security prices: a study of

    exchange offers. Journal of Financial Economics 8, 139–178.

Miller, M.H., Modigliani, F., 1961. Dividend policy, growth, and the valuation of shares.

    Journal of Business 34, 411–433.

Miller, M.H., Rock, K., 1985. Dividend policy under asymmetric information. Journal of

    Finance 40, 1031–1051.

Nohel, T., Tarhan, V., 1998. Share repurchases and firm performance: new evidence on the

    agency costs of free cash flow. Journal of Financial Economics 49, 187–222.




                                                                                           31
Ofer, A.R., Thakor, A.V., 1987. A theory of stock price responses to alternative corporate

    cash disbursement methods: stock repurchases and dividends. Journal of Finance 42,

    365–394.

Oswald, D., Young, S., 2004a. What role taxes and regulation? A second look at open market

    share buyback activity in the UK. Journal of Business Finance and Accounting 31, 257–

    292.

Oswald, D., Young, S., 2004b. Open market share reacquisitions, surplus cash, and agency

    problems. Working paper, London Business School.

Perfect, S.B., Peterson, D.R., Peterson, P.P., 1995. Self–tender offers: the effects of free cash

    flow, cash flow signaling, and the measurement of Tobin's q. Journal of Banking and

    Finance 19, 1005–1023.

Persons, J.C., 1997. Heterogeneous shareholders and signaling with share repurchases.

    Journal of Corporate Finance 3, 221–249.

Rau, P.R., Vermaelen, T., 2002. Regulation, taxes, and share repurchases in the United

    Kingdom. Journal of Business 75, 245–282.

Scholes, M., Williams, J., 1977. Estimating betas from nonsynchronous data. Journal of

    Financial Economics 5, 309–327.

Stephens, C.P., Weisbach, M.S., 1998. Actual share reacquisitions in open-market repurchase

    programs. Journal of Finance 53, 313–333.

Stulz, R.M., 1990. Managerial discretion and optimal financing policies. Journal of Financial

    Economics 26, 3–27.

Vermaelen, T., 1981. Common stock repurchases and market signaling: an empirical study.

    Journal of Financial Economics 9, 139–183.




                                                                                              32
Table 1. Summary of sample selection screening.
The table summarizes the sample selection criteria for the period 1 September 1997 to 31 July 2003.


                                                              Number of                    Number of
                                                            announcements                    firms
Original data from Company REFS                                 9,020                         494
Less:
     Non ordinary share repurchases                               183                          26
     No announcement date                                          1                           1a
     No repurchase price or volume data                            17                         11b
     No Datastream code                                           274                           2
     Investment trusts                                            634                          62
     Implausible reported trading prices                         1,836                         67
Subtotal                                                         6,075                        335
Less:
     Additional announcements on the same day                     337                         112
       Total for the full sample                                 5,738                        335
Less:
     Insufficient accounting data                                 341                          14
       Final sample                                              5,159                        301
a
    One announcement correcting a previous announcement remains in the full sample.
b
    Nine of these announcements correct or supplement previous announcements and remain in the full sample.




                                                                                                              33
Table 2. Comparison between single and multiple repurchasers.
The table compares 128 firms announcing repurchase executions in a single accounting year with 149 firms
announcing repurchase executions in multiple years during the sample period 1 September 1997 to 31 July
2003. Panels B and C report means (medians). The time span of a firm‘s repurchase execution announcements
is the number of trading days from the first to last announcements over the sample period. Market capitalisation
is the market value of equity at the year-end before the announcement, adjusted to 1996 currency. Market-to-
book ratio is the market value of equity divided by the book value of equity at the year-end before the
announcement year. Free cash flow is operating income before depreciation minus total tax, interest expense,
preferred dividends, and ordinary dividends for the announcement year. Outstanding shares are the value one
day before each repurchase execution announcement. The test for difference reports the p-value of a two-tailed
t-test (Wilcoxon rank test).

                                                Single repurchasers   Multiple repurchasers   Test for difference
Number of firms                                          128                  149
Number of firm–years                                     128                  417

Panel A: Time span of a firm's repurchase execution announcements
Maximum number of trading days                       179                     1,368
Minimum number of trading days                         1                       28
More than 28 days (number of firms)                   44                       –
More than 179 days (number of firms)                   –                      132

Panel B: Repurchase execution announcements
Time span of announcements (trading days)               52.50                559.05                <0.0001
                                                        (39.5)               (456.0)              (<0.0001)
Number of announcementsa                                7.438                27.275                <0.0001
                                                        (2.00)               (13.00)              (<0.0001)
Shares repurchased/outstanding sharesa                  0.039                 0.129                <0.0001
                                                       (0.017)               (0.092)              (<0.0001)
Panel C: Firm characteristics
Market capitalisation (in 1996 £m)                     787.08               2,198.45               0.1127b
                                                       (83.020)             (84.324)              (0.2008)b
Free cash flow/total assets                              0.050                0.072                0.0337
                                                        (0.060)              (0.068)              (0.0953)
Market-to-book ratio                                     2.137                1.990                0.6384
                                                        (1.449)              (1.178)              (0.0986)
a
    Aggregated for each firm over the sample period.
b
    Based on the log of market capitalization.




                                                                                                              34
Table 3. Change and cumulative change in return on assets.
The table reports changes and cumulative changes in return on assets (ROA) for firms announcing repurchase
executions during 1 September 1997 to 31 July 2003. ROA is annual income before interest, taxes, depreciation
and amortization scaled by the average of opening and closing book value of total assets. Performance-adjusted
change equals the unadjusted change minus the change in performance of a matching firm selected by the
criteria in Section 3.1. Panel A reports changes in ROA; Panel B reports cumulative changes. Cumulative
change is the sum of changes over the whole (2 to 3), pre- (2 to 0) and post-repurchase (0 to 3) periods.
Difference is the difference in cumulative changes between the post- and pre-repurchase periods. Observations
are winsorized at the 1st and 99th percentiles in calculating means and medians. Significance of means
(medians) are based on a two-tailed t-test (Wilcoxon rank test). a, b, and c denote significance at 1, 5, and 10
percent. All numbers are percentages.

Panel A: Changes in ROA
                          Unadjusted changes                               Performance-adjusted changes
  Year            1     0       1        2            3         1             0       1         2           3
All repurchasers
 Mean         0.573b 0.866a 1.281a 0.512        1.788a     0.037        1.663a       0.201   0.525    2.323
              0.171c 0.226b 0.507a 0.058        0.574     0.002                    0.118 0.356     0.989
                                                           a                       b
 Median                                                                      0.381
 N               519   507      407      289          204        511          444         307     185        113
Multiple repurchasers
 Mean          0.281 0.774b 0.724c 0.752c       0.923c     0.027        1.451b       0.915 0.117      0.452
  Median       0.036 0.353 0.345 0.058
                              b        b
                                                    0.574 b
                                                               0.004        0.311 b
                                                                                         0.118 0.505      0.258
  N             381    372    311     229             159        375          330         239     149         88
Single repurchasers
  Mean         1.097 1.211 3.969a 1.326          7.541c     0.026        3.628c      4.815   3.713   14.480c
  Median      0.538b 0.040 2.439a 0.096         0.267     0.002        0.626c      0.640   0.146    5.144c
  N             114    111      77      51             38        113           94          58       31        21

Panel B: Cumulative changes in ROA
                     Cumulative unadjusted changes                     Cumulative adjusted changes
  Period       2 to 3 2 to 0     0 to 3    Difference        2 to 3   2 to 0       0 to 3     Difference
All repurchasers
  Mean         6.062a 1.451a 4.943a         4.879a         2.356        1.691a         2.309        3.431
               3.140a 0.504a 2.597a         2.299a         0.988                       1.524        2.143
                                                                                     b
  Median                                                                     0.381
  N              197      507       197          197            111           444            111           111
Multiple repurchasers
  Mean         4.323a 1.218a 3.733a         2.801b         0.704         1.453b         0.152        1.051
               2.823a 0.471a 2.383a         1.846a         0.517                       0.187        0.934
                                                                                     b
  Median                                                                     0.250
  N              152      372       152          152             86           330             86            86
Single repurchasers
  Mean         14.95a 1.873c 12.37b         11.24c         14.39        3.657c         13.88        13.76
               6.664a 0.486c 3.763a         4.196          9.269                       7.141        4.391
                                                                       c             c
  Median                                                                     0.866
  N              38       111        38           38             21            94             21            21




                                                                                                                  35
Table 4. Descriptive statistics for regression variables.
The table reports Pearson correlation coefficients, means, medians, and standard deviations (SD) of explanatory
variables in equations (3)–(5). AG_REP is the aggregate ratio of shares repurchased in each repurchase
execution announcement to outstanding shares one day before each announcement over a firm–year. V_REP
and DIV are the value of total shares repurchased and the value of dividends in a firm–year divided by market
value of equity at the start of the repurchasing year. PAYOUT is the sum of V_REP and DIV. Level: ROA uses
the median of paired differences of pre-repurchase period return on assets (ROA), as defined in Table 3.
Change: ROA uses the median of adjusted changes in ROA over the pre-repurchase period, covering years 2
through 0. Significance levels of means (medians) are based on a two-tailed t-test (Wilcoxon rank test). ***,
**, and * denote significance at 1%, 5%, and 10%.

                 AG_REP       V_REP        DIV      PAYOUT         Mean      Median           SD        N
AG_REP            1           0.949***    0.008      0.821***     0.046***   0.025***       0.058      515
V_REP                         1           0.056       0.890***    0.043***    0.023***      0.058      519
                                                           ***         ***           ***
DIV                                       1           0.505       0.043       0.04          0.032      443
                                                                       ***            ***
PAYOUT                                                1           0.088       0.071         0.070      443
                                         0.025
                         *                                             ***            ***
Level: ROA       0.076        0.061                  0.026        0.020       0.005         0.136      511
Change: ROA     0.076       0.063       0.001     0.065        0.000       0.002         0.218      311




                                                                                                             36
Table 5. Regression analysis of changes in return on assets.
The table reports the results of alternative return on assets regressions in Panels A–C. POSTROAi is the
median of paired differences of post-repurchase return on assets (ROA) for repurchasing firm i, AG _ REPi is
the aggregate ratio of shares repurchased in each repurchase execution announcement to outstanding shares one
day before the announcement over a firm–year, V_ REPi and DIVi are the value of shares repurchased in a
fiscal year and the value of dividends, both divided by the start of year market value of equity, PAYOUTi is the
sum of V_ REPi and DIVi , and PREROAi is the median of paired differences of pre-repurchase period ROA.
The pre-repurchase period covers years 2 through 0, and the post-repurchase period covers years 1 through 3.
Level uses levels of ROA; Change uses changes in ROA. ***, **, and * denote significance at 1%, 5%, and
10%. Numbers in parentheses are t-statistics.

Panel A: POSTROAi   0  1 AG _ REP   2 PREROAi   i
             Model          Intercept           AG_REP               PREROA        R2     F-stat.        N
                                     ***
             Level             0.050            0.454**               1.084***   0.255   52.72***      311
All                           (2.89)           (2.07)               (10.14)
repurchasers Change            0.030*           0.360*               0.871***   0.122   21.49***      311
                              (1.98)           (1.87)               (6.40)
             Level             0.053***         0.459*                1.105***   0.239   37.57***      243
Multiple                      (2.80)           (1.88)                (8.61)
Repurchasers Change            0.034**          0.373*               0.727***   0.083   10.86***      243
                              (2.08)           (1.79)               (4.36)
             Level             0.041            0.420                 1.084***   0.308   12.22***       58
Single                        (0.90)           (0.73)                (4.79)
Repurchasers Change            0.014            0.347                1.122***   0.223     7.88***      58
             Change           (0.33)           (0.66)               (3.97)

Panel B: POSTROAi  0  1V _ REPi  2 DIVi  3 PREROAi   i
             Model           Intercept       V_REP          DIV       PREROA        R2     F-stat.       N
             Level                0.047        0.304      0.119      1.089***   0.243   30.83***      293
All                              (1.53)       (1.40)     (0.20)     (9.46)
repurchasers Change               0.018        0.260        0.052    0.767***   0.090    9.51***      293
                                 (0.66)       (1.36)       (0.10)   (5.19)
             Level                0.045        0.350        0.074     1.113***   0.221   21.05***      227
Multiple                         (1.24)       (1.37)       (0.11)    (7.83)
repurchasers Change               0.021        0.332        0.110    0.520***   0.043    3.32**       227
                                 (0.66)       (1.50)       (0.18)   (2.74)
             Level                0.062        0.168      0.874      1.089***   0.309    8.04***       58
Single                           (0.95)       (0.35)     (0.60)     (4.79)
repurchasers Change               0.021        0.130      0.476     1.102***   0.221    5.10***       58
             Change              (0.35)       (0.30)     (0.36)    (3.89)
Panel C: POSTROAi  0  1 PAYOUTi  2 PREROAi   i
             Model          Intercept             PAYOUT             PREROA        R2      F-stat.       N
              Level            0.053**             0.282              1.086***   0.242    46.36***     293
All                             (2.32)           (1.39)                (9.48)
repurchasers Change              0.028             0.223             0.766***   0.089    14.15***     293
                                (1.37)           (1.25)               (5.19)
             Level             0.059**             0.308              1.106***   0.220    31.50***     227
Multiple                        (2.22)           (1.26)                (7.82)
repurchasers Change              0.036             0.288             0.518***   0.041     4.74***     227
                                (1.53)           (1.36)               (2.73)
              Level              0.044             0.273              1.093***   0.307    12.15***      58
Single                          (0.88)           (0.67)                (4.84)
repurchasers Change              0.012             0.182             1.104***   0.220     7.76***      58
             Change             (0.27)           (0.49)               (3.94)




                                                                                                              37
Table 6. Changes and cumulative changes in capital expenditure.
The table reports changes and cumulative changes in capital expenditures for firms announcing repurchase
executions in the period 1 September 1997 to 31 July 2003. Capital expenditure is cash paid for tangible fixed
assets during a year scaled by the average of opening and closing book value of total assets. Adjusted change
equals the unadjusted change minus the change in capital expenditure of a firm matched on equivalent criteria to
ROA (see Section 3.1), except replacing ROA with capital expenditure. Cumulative unadjusted (adjusted)
change is a summary measure of unadjusted (adjusted) changes in capital expenditures over the whole (2 to 3),
pre- (2 to 0) and post-repurchase (0 to 3) periods. Difference is the difference between post- and pre-
repurchase periods. Observations are winsorized at the 1st and 99th percentiles in calculating mean and median
changes. Significance levels of means (medians) are based on a two-tailed t-test (Wilcoxon rank test). a, b, and
c denote significance at 1, 5, and 10 percent. All numbers are percentages.


Panel A: Changes in capital expenditures
                           Unadjusted changes                                        Adjusted changes
  Year            1      0         1       2    3                     1           0        1        2                    3
All repurchasers
 Mean         0.349b 0.496a 0.232 0.582b 0.088                   0.001       0.321       0.129     0.106         0.443
              0.112b 0.179a 0.207a 0.231a 0.304b                                         0.138 0.071
                                                                                          c
 Median                                                               0.001       0.247                                0.234
 N               559     558      457      325 229                     556        475          332    203              125
Multiple repurchasers
 Mean         0.341c 0.572a 0.430b 0.150 0.440                   0.012       0.269       0.283     0.622         0.144
  Median       0.077 0.193a 0.213a         0.165b
                                                          0.373 a
                                                                      0.003c      0.238       0.167     0.157         0.234
  N             413    412     353             260         181         411        353          259       164           101
Single repurchasers
  Mean        0.545c 0.182   0.705          2.023a      1.050     0.038       0.690        0.873 4.814c           0.953
  Median      0.324 0.128 0.287
                     b
                                              0.761a
                                                           0.000     0.004       0.247       0.089 1.517      a
                                                                                                                       0.144
  N             122    122       85              56          41        121        101            59     32               19

Panel B: Cumulative changes in capital expenditures
                        Cumulative unadjusted changes                          Cumulative adjusted changes
 Period       2 to 3       2 to 0    0 to 3      Difference        2 to 3      2 to 0       0 to 3     Difference
 All repurchasers
   Mean      1.682a      0.791a     0.621c           0.473        0.516        0.321        0.190       0.286
  Median     1.444a      0.261a     0.757a           0.094       0.030         0.237c       0.007        0.350
   N           229          558         229              229          125           475          125                 125
 Multiple repurchasers
   Mean      1.900a      0.888a     0.681b           0.508        0.604        0.267        0.235            0.793
  Median     1.280a      0.240a     0.757a           0.253       0.125        0.208        0.007        0.350
   N           181          412         181              181          101           353          101                 101
 Single repurchasers
   Mean       1.603       0.549     0.307            0.722        0.109         0.731        1.904       3.091
  Median     2.389b      0.426b     0.762            1.650c       1.205         0.268        0.278       1.401
   N            41          122          41              41            19           101           19                 19




                                                                                                                               38
Table 7. Changes and cumulative changes in cash reserves.
The table reports changes and cumulative changes in cash reserves (cash and equivalents scaled by the average of
opening and closing book value of total assets) for a sample of firms announcing repurchase executions in the
period 1 September 1997 to 31 July 2003. Adjusted change equals the unadjusted change minus the change in
cash reserves of a control firm matched on equivalent criteria to ROA (see Section 3.1), except replacing ROA
with cash reserves. Cumulative unadjusted (adjusted) change is a summary measure of unadjusted (adjusted)
changes in cash reserves over the whole (2 to 3), pre- (2 to 0) and post-repurchase (0 to 3) periods. Difference
is the difference in cumulative abnormal changes between the post- and pre-repurchase periods. Observations are
winsorized at the 1st and 99th percentiles in calculating the means and median changes. Significance levels of
means (medians) are based on a two-tailed t-test (Wilcoxon rank test). a, b, and c denote significance at 1, 5, and
10 percent. All numbers are percentages.

Panel A: Changes in cash reserves
                          Unadjusted change                                           Adjusted change
  Year            1      0        1       2    3                1            0            1         2                   3
All repurchasers
 Mean           0.435b 0.124 0.414c 0.712b 0.609b             0.008        0.273        0.287      1.233b         0.345
 Median        0.059  c
                          0.000 0.120   a
                                              0.002     0.000    0.000        0.150         0.137      0.355   c
                                                                                                                      0.308
 N               558       558    460          328       233      555          489           351        214            134
Multiple repurchasers
                          0.090 0.121 0.852 0.772             0.026
                                             b      b
 Mean           0.352                                                         0.414        0.696       1.926a         0.211
  Median        0.057      0.000 0.024 0.002 0.007              0.000        0.185        0.368c
                                                                                                       0.454b
                                                                                                                      0.434
  N              413        413   355    262   184                410          363          271         169            108
Single repurchasers
  Mean          0.658     0.336 1.120b 0.322 0.253            0.047         0.022      2.296b       1.129         0.537
 Median         0.308     0.157 0.412 0.007 0.087
                                         b
                                                                0.000         0.048      1.130    b
                                                                                                       1.144         0.299
 N               121       121    86     57     42               121           103         65            37             20

 Panel B: Cumulative changes in cash reserves
                    Cumulative unadjusted changes                        Cumulative adjusted changes
   Period      2 to 3 2 to 0     0 to 3   Difference           2 to 3   2 to 0     0 to 3     Difference
 All repurchasers
  Mean          0.659   0.354      0.916      1.339               1.745            0.259       1.730               1.556
  Median       0.047   0.040      0.000      0.465b              0.167            0.154       0.781               1.007c
  N             233         558       233             233          134             489         134                  134
 Multiple repurchasers
  Mean         1.755b      0.275     1.695b           1.584       3.741b           0.386      3.578b               3.086b
   Median      0.055      0.034     0.109        0.482b          0.762b           0.154      1.723a               2.256b
   N            184         413       184             184          108             363         108                  108
 Single repurchasers
   Mean       3.184c      1.106    2.236c       0.603         6.750b           0.038      5.023b              6.112
   Median      0.255      0.276    0.662         0.183         5.507   b
                                                                                   0.275      3.741   b
                                                                                                                   1.026
   N             42         121       42               42          20              103          20                   20




                                                                                                                              39
Table 8. Changes in risk using the CAPM and the three factor model.
The table reports cross-sectional mean and median estimates of the CAPM and the three factor model,
                            Rit  R ft   i   i Dt  bi ( Rmt  Rft )  bi Di ( Rmt  Rft )  eit ,
     Rit  R ft   i   i Dt  bi ( Rmt  R ft )  bi Dt ( Rmt  R ft )  si SMBt  si Dt SMBt  hi HMLt  hi Dt HMLt  eit ,
where Rit and Rmt are the month t return on stock i and on the FTSE All Share Index, R ft is the month t return
on three-month UK Treasury bills, SMBt is the difference between monthly returns on the FTSE Small Cap and
FTSE 100 Indexes, HMLt is the difference between monthly returns on the FTSE Global Value and FTSE
Global Growth Indexes, and Dt is a dummy variable equal to 1 if t  t * , where t * is the month in which firm i
announces its first repurchase execution during the period 1 September 1997 to 31 July 2003. We use a 73-
month window (36 to +36) to estimate the models. bi , s i , and hi are the factor loadings of firm i during
the 3 years before its first repurchase execution. bi , s i , and hi are the changes in factor loadings after the
repurchase execution. Adjusted regression coefficients equal unadjusted coefficients less the coefficients of
matched firms with market value of equity and book-to-market ratio between 80 and 120 percent of the
corresponding values of the repurchasing firm at the end of year 1 and with the closest buy-and-hold stock
return to a sample firm during the year before the first repurchase execution announcement. We exclude firms
with less than 44 months return data. Significance levels of means (medians) are based on a two-tailed t-test
(Wilcoxon rank test). ***, **, and * denote significance at 1%, 5%, and 10%.

                                                                                          Frequency of share repurchase groups
                                                            All repurchasers                Multiple               Single
Panel A: CAPM unadjusted changes
  bi (  in market beta)         Mean                            0.027                       0.021                        0.099
                                  Median                        0.010                        0.070                        0.099
                                  N                               320                          164                           132
Panel B: CAPM adjusted changes
  bi (  in market beta)         Mean                            0.120*                       0.151*                       0.049
                                  Median                          0.096*                       0.132*                       0.053
                                  N                                279                          141                          115
Panel C: three factor model unadjusted changes
bi (  in market beta)           Mean                           0.149***                  0.173***                   0.117*
                                  Median                         0.202***                  0.244***                   0.143
s i (  in small firm beta)      Mean                           0.083                     0.167**                     0.062
                                  Median                         0.151**                   0.198***                    0.011
hi (  in B/M beta)              Mean                            0.107                      0.118                       0.104
                                  Median                         0.043                      0.065                      0.152
                                  N                                320                         164                         132
Panel D: three factor model adjusted changes
bi (  in market beta)           Mean                             0.100                     0.160                       0.001
                                  Median                           0.079                     0.162*                     0.021
s i (  in small firm beta)      Mean                             0.061                    0.044                       0.200
                                  Median                           0.041                    0.079                       0.194*
hi (  in B/M beta)              Mean                           0.100                     0.015                      0.174
                                  Median                         0.083                      0.042                      0.422*
                                  N                                281                         141                         117




                                                                                                                                          40
Table 9. Changes in risk using the market model.
The table reports changes in market model beta for firms announcing repurchase executions in the period 1
September 1997 to 31 July 2003. We estimate Scholes–Williams betas using daily stock returns and FTSE All
Share Index returns. Adjusted change equals the unadjusted change less the change in beta of a control firm
with market value of equity and book-to-market ratio between 80 and 120 percent of the market value of equity
and market-to-book ratio of the repurchasing firm at the end of year 1 and with the closest change in beta in
year 1 to that of the sample firm. Significance of means (medians) is based on a two-tailed t-test (Wilcoxon
rank test). a, b, and c denote significance at 1, 5, and 10 percent.

                            Unadjusted changes                            Firm-adjusted changes
      Year           1      0       1       2         3        1         0        1        2          3
All repurchasers
Mean             0.039c   0.028c 0.025    0.029    0.031   0.066b     0.037   0.010   0.007     0.042
Median            0.009   0.010 0.018     0.032    0.021   0.036  b
                                                                         0.032   0.007   0.003    0.003
N                   540      558    454      324      228      479        488      396      284       199
Multiple repurchasers
Mean             0.042c   0.012   0.002   0.034    0.034    0.046     0.043   0.047    0.023     0.040
                  0.006   0.002                             0.010     0.030   0.022             0.003
                                                 c
Median                              0.014   0.041    0.012                                  0.013
N                   402     411      350     259      179      356        360      304       226      155
Single repurchasers
Mean              0.005   0.080c 0.101c 0.025 0.020 0.130b           0.016   0.120    0.115     0.013
                  0.010   0.077c 0.074 0.016 0.082 0.158b            0.076            0.116    0.018
                                        b
Median                                                                            0.050
N                   116      125     87    57     43     101              106      75        50        38




                                                                                                            41

								
To top