Download The information content of share repurchases

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts

Individual Savings Account wikipedia , lookup

Internal rate of return wikipedia , lookup

Private equity wikipedia , lookup

Financialization wikipedia , lookup

Beta (finance) wikipedia , lookup

Private equity secondary market wikipedia , lookup

Private equity in the 2000s wikipedia , lookup

Financial economics wikipedia , lookup

Investment fund wikipedia , lookup

Systemic risk wikipedia , lookup

Mark-to-market accounting wikipedia , lookup

Short (finance) wikipedia , lookup

Modified Dietz method wikipedia , lookup

Public finance wikipedia , lookup

Investment management wikipedia , lookup

Business valuation wikipedia , lookup

Stock selection criterion wikipedia , lookup

Private equity in the 1980s wikipedia , lookup

Early history of private equity wikipedia , lookup

Mergers and acquisitions wikipedia , lookup

Global saving glut wikipedia , lookup

Corporate finance wikipedia , lookup

Transcript
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: [email protected].
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 tenderoffer 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 tenderoffer 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 redeploying 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
,
( ASSETt  ASSETt 1 ) / 2
(1)
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
EBITDAt = 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 postrepurchase 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 nonrepurchasers are significant in the pre-repurchase period but not over the whole or postrepurchase 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, POSTROAi and PREROAi in equation (3) relate
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 postrepurchasing 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 PAYOUTi 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 PAYOUTi 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
PAYOUTi 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 PAYOUTi . 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 PAYOUTi 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 ,
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 ,
(6)
(7)
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 nonsynchronous 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
min i
 1  R
t , sample firm
t 1
12
   1  R
t , matching firm i
t 1

(8)
Using daily return data, we set an additional criterion to control for risk changes, that a
29
30
The monthly return data end on 2 February 2005.
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
Change in risk in year 1 is the Scholes–Williams beta in year 1 less the beta of the same firm in year 2.
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.
31
32
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 postrepurchase 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 informationsignaling 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 informationsignaling 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 redeploying 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.
Original data from Company REFS
Less:
Non ordinary share repurchases
No announcement date
No repurchase price or volume data
No Datastream code
Investment trusts
Implausible reported trading prices
Subtotal
Less:
Additional announcements on the same day
Total for the full sample
Less:
Insufficient accounting data
Final sample
a
Number of
announcements
9,020
Number of
firms
494
183
1
26
1a
17
274
634
1,836
6,075
11b
2
62
67
335
337
5,738
112
335
341
5,159
14
301
One announcement correcting a previous announcement remains in the full sample.
Nine of these announcements correct or supplement previous announcements and remain in the full sample.
b
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-tobook 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
128
128
149
417
Number of firms
Number of firm–years
Panel A: Time span of a firm's repurchase execution announcements
Maximum number of trading days
179
Minimum number of trading days
1
More than 28 days (number of firms)
44
More than 179 days (number of firms)
–
Panel B: Repurchase execution announcements
Time span of announcements (trading days)
Number of announcementsa
Shares repurchased/outstanding sharesa
Panel C: Firm characteristics
Market capitalisation (in 1996 £m)
Free cash flow/total assets
Market-to-book ratio
a
Test for difference
1,368
28
–
132
52.50
(39.5)
7.438
(2.00)
0.039
(0.017)
559.05
(456.0)
27.275
(13.00)
0.129
(0.092)
<0.0001
(<0.0001)
<0.0001
(<0.0001)
<0.0001
(<0.0001)
787.08
2,198.45
0.1127b
(83.020)
0.050
(0.060)
2.137
(1.449)
(84.324)
0.072
(0.068)
1.990
(1.178)
(0.2008)b
0.0337
(0.0953)
0.6384
(0.0986)
Aggregated for each firm over the sample period.
Based on the log of market capitalization.
b
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
Year
0
1
2
1
All repurchasers
Mean
0.573b 0.866a 1.281a 0.512
Performance-adjusted changes
0
1
2
3
1
1.788a
0.037
1.663a
0.525
2.323
Median
0.171 0.226 0.507 0.058
N
519
507
407
289
Multiple repurchasers
Mean
0.281 0.774b 0.724c 0.752c
0.574
204
a
0.002
511
0.381
444
b
0.118 0.356
307
185
0.989
113
0.923c
0.027
1.451b
0.915 0.117
0.452
Median
0.036 0.353 0.345 0.058
N
381
372
311
229
Single repurchasers
Mean
1.097 1.211 3.969a 1.326
0.574
159
b
0.004
375
0.311
330
b
0.118 0.505
239
149
0.258
88
7.541c
0.026
3.628c
4.815
3.713
14.480c
0.538b 0.040 2.439a 0.096
114
111
77
51
0.267
38
0.002
113
c
0.640
58
0.146
31
5.144c
21
c
b
a
b
Median
N
b
Panel B: Cumulative changes in ROA
Cumulative unadjusted changes
Period
0 to 3
Difference
2 to 3 2 to 0
All repurchasers
Mean
6.062a 1.451a 4.943a
4.879a
2.356
1.691a
2.309
3.431
0.988
111
0.381
444
b
1.524
111
2.143
111
1.453b
0.152
1.051
b
0.187
86
0.934
86
0.504
507
1.218a
3.733a
2.801b
0.704
Median
2.823
N
152
Single repurchasers
Mean
14.95a
0.471
372
2.383
152
1.846
152
a
0.517
86
0.250
330
1.873c
12.37b
11.24c
14.39
3.657c
13.88
13.76
6.664
38
0.486
111
3.763
38
4.196
38
9.269
21
c
7.141
21
4.391
21
a
Median
N
a
a
a
c
a
a
a
2.299
197
Cumulative adjusted changes
0 to 3
Difference
2 to 0
Median
3.140
N
197
Multiple repurchasers
Mean
4.323a
a
2.597
197
2 to 3
0.626
94
0.201
3
a
c
0.866
94
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
AG_REP
1
V_REP
V_REP
0.949***
DIV
0.008
PAYOUT
0.821***
Mean
0.046***
Median
0.025***
SD
0.058
N
515
1
0.056
0.890***
0.043***
0.023***
0.058
519
***
0.043
***
0.088
***
0.020
0.000
***
DIV
1
PAYOUT
Level: ROA
Change: ROA
0.505
1
0.076
0.076
*
0.061
0.063
0.025
0.001
0.026
0.065
0.04
***
0.032
443
0.071
***
0.070
443
0.005
0.002
***
0.136
0.218
511
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 DIV i 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, PAYOUT i is the
sum of V_ REPi and DIV i , 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: POSTROA i   0  1 AG _ REP   2 PREROAi   i
Model
Intercept
AG_REP
***
Level
0.050
0.454**
(2.89)
All
(2.07)
0.030*
repurchasers Change
0.360*
(1.98)
(1.87)
Level
0.053***
0.459*
Multiple
(2.80)
(1.88)
Repurchasers Change
0.034**
0.373*
(2.08)
(1.79)
Level
0.041
0.420
Single
(0.90)
(0.73)
Repurchasers Change
0.014
0.347
(0.33)
Change
(0.66)
Panel B: POSTROAi  0  1V _ REPi  2 DIVi  3 PREROAi   i
Model
Intercept
V_REP
DIV
Level
0.047
0.304
0.119
(1.53)
All
(1.40)
(0.20)
0.018
0.052
repurchasers Change
0.260
(0.66)
(0.10)
(1.36)
Level
0.045
0.074
0.350
Multiple
(1.24)
(0.11)
(1.37)
repurchasers Change
0.021
0.110
0.332
(0.66)
(0.18)
(1.50)
Level
0.062
0.168
0.874
Single
(0.95)
(0.35)
(0.60)
repurchasers Change
0.021
0.130
0.476
(0.35)
Change
(0.30)
(0.36)
Panel C: POSTROAi  0  1 PAYOUTi  2 PREROAi   i
Model
Intercept
PAYOUT
Level
0.053**
0.282
(2.32)
All
(1.39)
0.028
repurchasers Change
0.223
(1.37)
(1.25)
Level
0.059**
0.308
Multiple
(2.22)
(1.26)
repurchasers Change
0.036
0.288
(1.53)
(1.36)
Level
0.044
0.273
Single
(0.88)
(0.67)
repurchasers Change
0.012
0.182
(0.27)
Change
(0.49)
PREROA
1.084***
(10.14)
0.871***
(6.40)
1.105***
(8.61)
0.727***
(4.36)
1.084***
(4.79)
1.122***
(3.97)
R2
0.255
F-stat.
52.72***
N
311
0.122
21.49***
311
0.239
37.57***
243
0.083
10.86***
243
0.308
12.22***
58
0.223
7.88***
58
PREROA
1.089***
(9.46)
0.767***
(5.19)
1.113***
(7.83)
0.520***
(2.74)
1.089***
(4.79)
1.102***
(3.89)
R2
0.243
F-stat.
30.83***
N
293
0.090
9.51***
293
0.221
21.05***
227
0.043
3.32**
227
0.309
8.04***
58
0.221
5.10***
58
R2
0.242
F-stat.
46.36***
N
293
0.089
14.15***
293
0.220
31.50***
227
0.041
4.74***
227
0.307
12.15***
58
0.220
7.76***
58
PREROA
1.086***
(9.48)
0.766***
(5.19)
1.106***
(7.82)
0.518***
(2.73)
1.093***
(4.84)
1.104***
(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 prerepurchase 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
Year
0
1
2
3
1
All repurchasers
Mean
0.349b 0.496a 0.232 0.582b 0.088
1
Median
0.112b 0.179a 0.207a 0.231a 0.304b
N
559
558
457
325
229
Multiple repurchasers
Mean
0.341c 0.572a 0.430b 0.150 0.440
Median
0.077 0.193
N
413
412
Single repurchasers
Mean
0.545c 0.182
Median
N
0.324
122
b
a
0.128
122
0.213
353
3
0.001
0.321 0.129
0.106
0.443
0.001
556
0.247 0.138
475
332
0.071
203
0.234
125
0.269 0.283
0.622
0.144
0.238 0.167
353
259
0.157
164
0.234
101
c
0.012
b
0.373
181
0.705
2.023a
1.050
0.038
0.690
0.873
4.814c
0.953
0.287
85
0.761
56
0.000
41
0.004
121
0.247 0.089
101
59
1.517
32
0.144
19
a
0.165
260
Adjusted changes
0
1
2
a
a
0.003
411
c
a
Panel B: Cumulative changes in capital expenditures
Cumulative unadjusted changes
Period
2 to 3
2 to 0
Cumulative adjusted changes
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
229
558
229
229
125
475
125
125
N
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
181
412
181
181
101
353
101
101
N
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
41
122
41
41
19
101
19
19
N
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
Year
0
1
2
3
1
All repurchasers
Mean
0.435b 0.124 0.414c 0.712b 0.609b
c
Median
0.059
N
558
Multiple repurchasers
Mean
0.352
Median
0.057
N
413
Single repurchasers
Mean
0.658
Median
N
0.308
121
0.000 0.120
558
460
a
0
0.008
0.273
0.287
3
1.233b
c
0.000
233
0.000
555
0.150
489
0.137
351
0.355
214
0.090 0.121 0.852b 0.772b
0.026
0.414
0.696
1.926a
c
b
0.000
413
0.024
355
0.002
328
Adjusted change
1
2
1
0.002
262
0.308
134
0.211
0.434
108
0.007
184
0.000
410
0.185
363
0.368
271
0.336 1.120b 0.322 0.253
0.047
0.022
2.296b
1.129
0.537
0.048
103
1.130
65
1.144
37
0.299
20
0.157 0.412 0.007 0.087
121
86
57
42
b
Panel B: Cumulative changes in cash reserves
Cumulative unadjusted changes
Period
0 to 3
Difference
2 to 3 2 to 0
All repurchasers
Mean
0.659
0.354
0.916
1.339
Median
0.040
0.000
0.465b
0.047
0.000
121
2 to 3
b
0.454
169
0.345
Cumulative adjusted changes
0 to 3
Difference
2 to 0
1.745
0.167
0.259
0.154
1.730
0.781
1.556
1.007c
N
233
Multiple repurchasers
Mean
1.755b
558
233
233
134
489
134
134
0.275
1.695b
1.584
3.741b
0.386
3.578b
3.086b
0.055
0.034
0.109
0.482b
0.762b
0.154
1.723a
2.256b
413
184
184
108
363
108
108
1.106
2.236c
6.750b
0.038
5.023b
6.112
0.255
0.276
0.662
0.603
0.183
0.275
3.741
1.026
42
121
42
42
20
103
20
Median
N
184
Single repurchasers
Mean
3.184c
Median
N
5.507
b
b
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  R ft )  bi Di ( Rmt  R ft )  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 73month 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
Panel A: CAPM unadjusted changes
Mean
bi (  in market beta)
Median
N
Panel B: CAPM adjusted changes
Mean
bi (  in market beta)
Median
N
Panel C: three factor model unadjusted changes
Mean
bi (  in market beta)
Median
Mean
si (  in small firm beta)
Median
Mean
hi (  in B/M beta)
Median
N
Panel D: three factor model adjusted changes
Mean
bi (  in market beta)
Median
Mean
si (  in small firm beta)
Median
Mean
hi (  in B/M beta)
Median
N
All repurchasers
Multiple
Single
0.027
0.021
0.070
164
0.099
0.099
132
0.151*
0.132*
141
0.049
0.053
115
0.010
320
0.120*
0.096*
279
0.149***
0.202***
0.083
0.151**
0.107
0.043
320
0.100
0.079
0.061
0.041
0.100
0.083
281
0.173***
0.244***
0.167**
0.198***
0.118
0.065
164
0.160
0.162*
0.044
0.079
0.015
0.042
141
0.117*
0.143
0.062
0.011
0.104
0.152
132
0.001
0.021
0.200
0.194*
0.174
0.422*
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
Year
0
1
2
1
All repurchasers
Mean
0.039c 0.028c 0.025 0.029
Median
0.009 0.010
N
540
558
Multiple repurchasers
Mean
0.042c 0.012
Firm-adjusted changes
0
1
2
3
1
0.031
0.066b
0.037
0.010
0.007
0.042
b
0.032
488
0.007
396
0.003
284
0.003
199
3
0.018
454
0.032
324
0.021
228
0.036
479
0.002
0.034
0.034
0.046
0.043
0.047
0.023
0.040
Median
0.006 0.002 0.014 0.041 0.012
N
402
411
350
259
179
Single repurchasers
Mean
0.005 0.080c 0.101c 0.025 0.020
0.010
356
0.030
360
0.022
304
0.013
226
0.003
155
0.130b
0.016
0.120
0.115
0.013
0.158
101
0.076
106
0.050
75
0.116
50
0.018
38
c
Median
N
0.010 0.077
116
125
c
0.074b 0.016 0.082
87
57
43
b
41