Brown Et Al 2015 Abacus

bs_bs_banner
ABACUS, Vol. 51, No. 1, 2015
doi: 10.1111/abac.12041
CHRISTINE BROWN, JOHN HANDLEY, AND JAMES ODAY
The Dividend Substitution Hypothesis:

Australian Evidence
In a perfect capital market firms are indifferent to either dividends or

repurchases as payout mechanisms, suggesting that the two payout
methods should be perfect substitutes. Empirical research at the single
country level, as well as cross country studies, provide evidence that dividends and repurchases act as substitutes (the dividend substitution hypothesis), and that the tax treatment of dividends versus capital gains affects
this relation. Australia, which operates under a full dividend imputation
system, has two types of repurchases: on- and off-market. On-market
repurchases are taxed as capital gains while off-market repurchases comprise a large dividend component carrying valuable tax credits. Australia
thus provides a natural setting to investigate how the tax treatment of
proceeds affects the dividend substitution hypothesis. Dividend substitution is found to exist for on-market repurchases but not for off-market
repurchases, thus providing further support for the idea that the tax treatment of proceeds affects the substitutability of repurchases and dividends.
Key words: Buyback; Dividend imputation; Equal access; Franking;
Off-market; On-market; Payout; Repurchase.
Share repurchases were not an important form of corporate payout when Lintners
(1956) seminal research on dividend policy was conducted. However, in the US,
which has the longest history of share repurchases, share repurchases grew to
become the most important method of cash distribution to shareholders. Grullon
and Michaely (2002) show that share repurchases in the US grew at an average rate
of 26.1% over the period 19802000 and surpass ordinary dividends in the dollar
value of cash distributed to shareholders in the years 1999 and 2000.
The theory of payout policy was established by Miller and Modigliani (1961).They
show that when capital markets are perfect and frictionless, dividends and share
repurchases are perfect substitutes and that the size of the payout has no effect on
Christine Brown (christine.brown@monash.edu) is at the Department of Banking and Finance, Monash

Business School, Monash University, Victoria. John Handley (handleyj@unimelb.edu.au) is at the
Department of Finance, University of Melbourne, Victoria. James ODay was an honours student at The
University of Melbourne when the first version of this paper was written. We thank Anne Ritter for her
excellent research assistance. We are grateful to Kevin Davis, Bill Griffiths, Krishnan Maheswaran, Garry
Twite, Peter Swan, Glenn Boyle, and seminar participants at The University of Melbourne, The Queens
University Belfast, The University of New South Wales and Victoria University Wellington for helpful
comments on earlier versions of the paper. This research was supported under the Australian Research
Councils Discovery Project funding scheme (project number DP0878537).
37
2015 Accounting Foundation, The University of Sydney
A BAC U S
firm value. Whether introduction of market frictions in the form of agency costs,
information asymmetries, and taxation affect the Miller and Modigliani propositions
has been the focus of numerous theoretical and empirical studies. Imperfections
could imply that the size of the distribution matters, the form of distribution matters,
or that they both matter.
Recent studies on payout policy integrate dividend and share repurchases as
alternative forms of cash disbursement, and attempt to answer whether these two
mechanisms are complements or substitutes. Whether repurchases act as a substitute
for dividends is termed the dividend substitution hypothesis. Brav et al. (2005)
conduct surveys and interviews with CFOs and conclude that dividends and share
repurchases are not viewed as pure substitutes; managers value the flexibility inherent in share repurchases in contrast to the relative rigidity of dividends. The reticence to increase dividends found by Lintner (1956) is also reflected in the finding of
Brav et al. that managers would not use a hypothetical reduction in repurchases to
increase dividends. Empirically, a negative relation between changes in dividends
and changes in repurchases would suggest that dividends and repurchases are substitutes. Grullon and Michaely (2002) conduct such an empirical test and find that
companies in the US are buying back shares with funds that would otherwise have
been used to increase dividends. Skinner (2008) reports that US firms increasingly
use repurchases to pay out earnings, while Floyd et al. (2012) confirm this finding
using more recent US data. These findings tend to support the view expressed in
Brav et al. that once free of historical constraints on dividend payouts, managers
would substitute from dividends towards repurchases. Nevertheless the dividend
payout for large, mature industrial firms in the US continues to increase (Floyd et al.,
2012).
Differential taxation treatment of capital gains and dividends is a significant
determinant of the market reaction to share repurchase announcements (Grullon
and Michaely, 2002). But these authors and others (for example Dittmar, 2000) have
argued that taxes alone do not explain the extent of repurchase activity in the US.
Brav et al. (2005) find that in the US managers view tax considerations as of second
order importance in the choice of disbursement mechanism. Nevertheless, evidence
provided by Chetty and Saez (2005) and Julio and Ikenberry (2004) suggests that
dividend increases and initiations are on the rise in the US since the tax changes
introduced in the Jobs & Growth Tax Relief Reconciliation Act of 2003, which
lowered the tax rate on dividend income.1 Chetty and Saez conclude that when key
shareholders are affected by changes in taxation policy the corporate response to the
changes is greater. De Angelo et al. (2004) show that large, mature and profitable
firms dominate the group of US firms paying dividends. Hoberg and Prabhala (2009)
find that risk explains over 40% of the decline in dividends and that, once risk is
controlled for, the disappearing dividend puzzle (Fama and French, 2001) is not
explained by Baker and Wurglers (2004a, b) suggestion that firms cater to transient
fads for dividends.
1
The 2003 tax reform results in dividends being taxed at a rate of 15% instead of facing the regular
income tax schedule with a top rate of 35%.
38
T H E D I V I D E N D S U B S T I T U T I O N H Y P OT H E S I S : A U S T R A L I A N E V I D E N C E
US evidence is mixed as to the extent to which differential taxation treatment of

repurchases and dividends affects the payout decision. Jacob and Jacob (2013) argue
that single country studies provide mixed evidence, whereas cross-country studies
are likely to have sufficient variation in the tax rates across time and countries to
derive robust conclusions. In their cross-country study they find that firms propensity to pay dividends or repurchase shares (and the amounts) are closely associated
with the tax penalty on dividends relative to capital gains.
In Australia, Partington (1989) finds that managers consider the tax position of
shareholders the least important of all variables examined that affect dividend
policy. In more recent Australian evidence, Coleman et al. (2010) show that managers do consider the ability to distribute imputation tax credits important when
setting payout policy. We therefore use the Australian tax environment, where dividends are not as tax-disadvantaged relative to capital gains, to further investigate the
extent to which dividends and repurchases act as substitutes. For certain shareholders in Australia dividends are tax advantaged because they carry imputation tax
credits, which can be used to offset personal income tax liabilities. Consistent with
this, Pattenden and Twite (2008) find that dividend payouts in Australia increased
with the introduction of the imputation tax system.2 Therefore, our primary objective in this paper is to use the unique opportunity provided by the tax environment
in Australia to investigate the relation between repurchase activity and dividend
payout in a non-classical tax environment.
There are a number of inter-related factors that potentially affect firms in their
decision to repurchase. We control for these firm characteristics in investigating the
relation between repurchase yield (defined as dollar volume repurchased divided by
market value of the firm in the previous period) and changes to dividend payout.The
aim of our paper is to test the hypothesis that firms are substituting repurchases for
dividends in Australia. If increases in dividends are associated with lower repurchase
yield we take this as evidence in support of the substitutability of repurchases and
dividends. If, on the other hand, firms that increase dividends do not decrease
repurchase yield we take this as evidence against the substitution hypothesis and in
favour of treating the two forms of disbursement as complements.
Share repurchases have become an important capital management mechanism for
some Australian companies. Our study provides a comprehensive investigation
because it covers virtually all firms that have undertaken repurchases over the
period 1997 to 2009. When on-market repurchases and off-market repurchases are
investigated separately, there is evidence for substitution only in the on-market
repurchase sample. The proceeds from on-market repurchases in Australia are taxed
as capital gains. So they are taxed as if in a classical tax system and the substitution
effect is evident for these repurchases.
On the other hand off-market repurchases generally contain a substantial dividend component with associated tax credits. For these repurchases there is no
2
Currently Australia, Chile, Mexico, and New Zealand have a full imputation system where a tax credit
is given to shareholders for the full corporate tax. Canada, the United Kingdom and South Korea have
partial imputation systems.
39
A BAC U S
evidence of substitution. In fact, increases in repurchase yield are significantly and

positively related to increases in dividends, consistent with the observation that
large, dividend paying companies tend to undertake off-market repurchases. Companies buying back shares off-market are not using funds that would have otherwise
been distributed as dividends. Using the demarcation between the tax treatment of
the proceeds of on- and off-market repurchases facilitates a natural experiment, the
results of which show that taxes are an important determinant of a firms choice of
payout method.
Furthermore, when the whole sample of repurchases is combined to investigate
dividend substitution, the results show a negative relation between repurchase yield
and dividend changes, a finding that supports the notion that in Australia repurchases and dividends in the aggregate act as substitutes. The paper makes an important contribution to the literature by providing direct empirical evidence that the
taxation treatment of the proceeds of a repurchase affects the substitutability of
dividends and repurchases.
SHARE REPURCHASES IN AUSTRALIA
The study of repurchases in Australia3 is driven by two legislative issues: company
law enabling buybacks, and legislation determining the tax treatment of the buyback
price for participating shareholders. In 1989 legislation was introduced that allowed
companies to repurchase shares. But it was not until December 1995, when the rules
were considerably simplified, that repurchase activity surged. Firms can now repurchase their shares through two main vehicles: on-market or off-market repurchases.4
Off-market repurchases can be categorized into equal access, selective, or minimum
holding. The Appendix describes these categories and gives details of the regulations
governing repurchases in Australia.
Australia changed from a classical tax system to a full dividend imputation system
on 1 July 1987. Under the imputation system, resident companies generate imputation or franking credits for the company tax paid. Part 3-6 of the Income Tax
Assessment Act 1997 (ITAA97) deals with the tax law pertaining to the dividend
imputation system in Australia. The major influences on the franking account
balance (FAB) are as follows.According to s. 205-15 ITAA97 Franking Credits, if the
company pays income tax and satisfies the residency requirement, it credits
(increases) the franking account by the amount of tax paid. If the company receives
Early papers include Harris and Ramsay (1995) and Balachandran and Faff (2004) investigating
announcement day effects; Lamba and Ramsay (2005) look at the impact of deregulation on repurchases; Mitchell and Robinson (1999) provide a study of the regulatory environment and the motivations for repurchases prior to the first Corporate Law Simplification Bill of December 1995, which
simplified the process for firm undertaking buybacks. More recent papers are described in the text.
The terms repurchase and buyback are used inter-changeably throughout the paper. In the US
on-market repurchases are generally referred to as open-market and off-market repurchases are
referred to as self-tender offers. In Australia, as in the US, in an off-market repurchase the buyback
price can be fixed or determined through a Dutch auction.
40
a franked distribution and satisfies the residency requirement, it credits the franking
account by the amount of the franking credit attached to the distribution. According
to s. 205-30 ITAA97 Franking Debits, if the company pays a franked distribution, it
debits (reduces) the franking account by the amount of the franking credit attached
to the distribution. This includes franking credits attached to the dividend component of an off-market buyback. Under certain conditions an on-market repurchase
will also result in a reduction in the FAB.5
Resident shareholders declare the dividend (grossed up to equal the precompany-tax profit from which the dividend was paid) as income, and then the tax
credit is used to offset personal income tax obligations.6 If D is the cash dividend
paid, the resulting taxable personal income is D/(1 tc) where tc is the corporate tax
rate. The tax levied on the investor is thus tD/(1 tc), where t is the investors
marginal tax rate, but the investor also receives a tax credit of tcD/(1 tc) such that
the tax payable (or rebateable) is (t tc)D/(1 tc). Overseas investors cannot use the
franking credits.
Australias capital gains tax (CGT) provisions operate to tax capital gains on the
sale of assets acquired after 20 September 1985 as assessable income in the year of
disposal of the asset. Prior to September 1999 the capital gains tax liability was
calculated using the indexation method, whereby the inflation adjusted capital gain
is included in ordinary income. Under the discount method for assets acquired after
September 1999 (and which taxpayers can elect to use for assets acquired prior to
that date), 50% of the nominal capital gain accrued on assets held for longer than
one year is included as income. Capital losses are offset against capital gains in the
year of calculation or carried forward.
The major difference between on- and off-market repurchases for Australian
shareholders is the tax treatment of the proceeds for the shareholder. As far as the
tax liability of the company is concerned, a repurchase is tax neutral. Under the rules
governing off-market repurchases in Australia companies can split the repurchase
price into capital and dividend components. This split must be confirmed by a
ruling from the Commissioner of Taxation subsequent to which companies with
5
ITAA s. 160AQCC, s. 160AQE(5). On-market share repurchases can result in a deduction to the
franking account of an equivalent amount of franking credits, in other words for tax purposes treating
the company as if it had conducted an off-market repurchase rather than an on-market repurchase.
This occurs if the on-market buyback is conducted using funds sourced from profits. Thus it is likely
that only those companies that have low FABs (those that have not generated excess franking credits
or those that have recently conducted an off-market buyback) will undertake on-market buybacks.
Brown and Norman (2010) show that companies conducting on-market buybacks have significantly
lower FABs than those conducting off-market buybacks. See http://www.ato.gov.au/Media-centre/
Speeches/Current-taxation-issues-arising-out-of-capital-management-strategies/ where the ATO
states that companies could [by undertaking off-market and on-market buy-backs in tandem, target
different shareholder profiles. For instance, an off-market buy-back with a substantial dividend component would be attractive to resident shareholders and a subsequent on-market buy-back out of
capital could be used to make distributions to non-resident shareholders.
Australian resident individuals, complying superannuation (pension) funds, registered organizations

and life assurance companies may use distributed franking credits to offset their tax liabilities. If all the
franking credits are distributed, and all recipients are able to fully utilize them, then the imputation
system effectively eliminates the double taxation of dividends (Officer, 1994).
41
A BAC U S
accumulated tax credits can fully frank the dividend component of the repurchase.
The capital amount is calculated as the deemed tax value7 of the offer price minus
the dividend. The participating shareholder benefits from the imputation credits and
may also benefit if the cost base for calculation of CGT implies that the sale results
in a capital loss.8 This structure has resulted in many off-market repurchases occurring at a discount to the market price. Brown and Efthim (2005) find that the size of
the discount of the offer price to the current share price is significantly related to the
proportion of the repurchase price offered as a franked dividend. Brown and Davis
(2012) estimate that equilibrium, market clearing discounts in tenders averaged
around 20% for the buybacks in their sample, but actual discounts were constrained
by the Australian Tax Office to 14%.9 The dividend/capital gains breakdown of the
offer price means that the sum of the tax gains from any capital losses on the sale and
the imputation tax credits attaching to the dividend component may be of greater
value to certain shareholders than the loss involved in selling shares at a price below
the market price.
From the shareholders perspective, on-market repurchases are treated in the
same way as any sale of shares, and the proceeds are subject to CGT. Underpinning
our empirical approach is the stark comparison of the tax treatment applying to onand off-market repurchases. Several changes to the tax treatment of income, capital
gains, and imputation credits over the sample period10 have not affected the fundamental outcome that off-market repurchases carry large tax benefits for some resident shareholders, while the proceeds of on-market repurchases are taxed as capital
gains.
Institutional investors such as Australian superannuation funds, on a marginal tax
rate of 15% compared to the company tax rate of 30%, benefit from selling shares
into off-market repurchases (Brown and Efthim, 2005). Low tax paying institutions
The deemed tax value of shares under TD2004/22 (Tax Determination) is calculated as volume
weighted average price on the ASX over the last five trading days before the first announcement of the
buyback, adjusted for the movement in the S&P/ASX200 Index from the opening of trading on the
announcement date to the close of trading on the day the buy back.
The following example provides an illustration of the breakdown between dividend and capital
components. On 11 April 2003 Woolworths completed an off-market buyback at a price of $11.40, of
which $8.52 was designated a fully franked dividend and $2.88 the capital amount. Participating
shareholders use $2.88 as the sale price for calculation of CGT. The share price on announcement of
the repurchase was $11.04; many participating shareholders could have claimed a capital loss on the
sale.
In late 2007, the ATO released a Practice Statement (PSLA 2007/9) stating that the maximum discount
allowed in an off-market repurchase is 14% calculated by reference to the volume weighted average
price on the five days leading up to and including the closing date of the repurchase. In practice the
ATO had been applying this maximum discount in private rulings for some years prior to its official
announcement.
10
In terms of changes to the value of the imputation credits that investors receive, the corporate tax rate
has been 30% since July 2001, was 34% in the preceding fiscal year and was 36% for the prior five
years. Additionally, since July 2000 Australian resident investors have been able to receive a refund
from the tax office if taxable income is insufficient to use all franking credits received.
42
form a very important part of the shareholder base for most companies. For example
in the UK, Rau and Vermaelen (2002) argue that firms are influenced by large
institutional investors such as pension funds when setting their payout policy.
Coleman et al. (2010) use management surveys and interviews to show that Australian CFOs are impacted in their financial decisions by pressure from expectations of
different stakeholder groups. AuYong et al. (2014) highlight the importance of tax
motivations in explaining abnormal trading volumes around key dates for offmarket repurchases.
PAYOUT POLICY
Are dividends and share repurchases inter-changeable payout methods? The theoretical answer to this question begins with the seminal work of Miller and Modigliani
(1961), who first argued that the value of a firm is entirely determined by its
investment policy and is consequently unaffected by the mix of retained earnings
and payout. This result relies on perfect capital markets, in which dividends and
share repurchases are equivalent. The introduction of market frictions in the form of
taxes, information asymmetries and agency costs may affect the equivalence of
dividends and repurchases in a number of ways. The following subsections review
the theoretical and empirical research on the motivation for repurchases and the
relation between repurchases and dividends.
Taxation
The US operates under a classical company tax system where dividends are paid out
of after company tax income. This results in dividends being taxed twice. Until the
introduction of the Jobs & Growth Tax Relief Reconciliation Act of 2003 dividend
income was taxed at the marginal tax rate of the receiving shareholder.
Repurchases (both open-market and self-tender offers) in the US are taxed on a
capital gains basis. Since the tax rate on capital gains is generally lower than that on
dividend income, most investors would prefer the company to disburse cash via
share repurchases rather than dividends (Black, 1976; Barclay and Smith, 1988). So,
in an otherwise perfect capital market, the effect of taxation in the US is to induce
a preference for payout in the form of repurchases. Grullon and Michaely (2002) find
that differential taxation is also important in market reactions to share repurchases
with the reaction to repurchases being more positive when tax gains on repurchases
relative to dividends are larger. They suggest, as do Jagannathan et al. (2000), that
differential taxation is, however, not sufficient to fully explain the observed increase
in share repurchase activity, since the upsurge in repurchases coincided with legislation that made repurchases less tax effective.11 In support of this argument,
Dittmar (2000) finds that repurchasing firms do not have lower dividend payout
ratios and argues that it is not the tax benefits of repurchases that cause firms to
repurchase stock.
11
The 1986 Tax Reform Act eliminated the preferential tax treatment for realized capital gains.
43
A BAC U S
Australia operates a full dividend imputation system as discussed in the previous

section. Companies accumulate tax credits in the franking account and must decide
when and how these franking credits are to be distributed to shareholders. Prior to
legislation enabling share repurchases, the primary method of distributing the tax
credits to shareholders was through the payment of fully franked ordinary dividends.
Brown and Howard (1992) argue that the Australian imputation tax system is biased
towards high dividend payouts. Monkhouse (1993), in deriving the CAPM under an
imputation system, finds that the optimal dividend policy is for a firm to distribute all
its imputation credits because they lose value as time passes.12 Many resident shareholders would prefer to receive returns in the form of fully franked dividends as the
tax rate on this form of distribution is lower than that on capital gains.13 Thus
dividends in Australia are not tax disadvantaged to the same extent as they were in
the US prior to 2003.
On-market repurchases in Australia are taxed as capital gains. So assuming that
tax treatment of proceeds affects the substitutability of dividends and repurchases
(Jacob and Jacob, 2013), we expect to find the substitution hypothesis supported for
on-market buybacks in Australia.
Australian firms that have accumulated imputation tax credits in excess of needs
under ordinary dividend policy will be influenced in their choice of buyback by
tax-related factors such as the size of the franking account, how the payment is taxed
when received by shareholders and whether the cash distributed via a repurchase
has franking credits attached (Brown and Norman, 2010). Off-market repurchases
offer a mechanism for firms to distribute these excess franking credits. In support of
this, 54% of respondents in the survey conducted by Coleman et al. (2010) indicated
that the level of franking credits available to be paid out to shareholders was
considered important in the decision to undertake a buyback.
Firms may be reluctant to use off-market repurchases as a substitute for ordinary
dividends, because this action will disadvantage a group of shareholders who do not
find it advantageous to participate in the buyback (Brown and Efthim, 2005; Brown
and Davis, 2012). But suppose that firms follow an optimal dividend policy and
distribute franking credits through ordinary dividends to the maximum extent possible, given the current dividend policy. This suggests that while Australian firms are
unlikely to substitute from ordinary dividends towards off-market repurchases they
12
The Income Tax Assessment Act (ITAA s.160AQE) governs the extent to which companies can frank
a dividend distribution. Entities may frank dividends subject to a benchmark rule, which provides that
all dividends paid within a franking period must be franked to the same extent. Breaches may result
in penalties.
13
Compare $1 received as a fully franked dividend versus $1 paid as a capital gain using the discount
method to calculate CGT. Assume personal tax rate p and company tax rate t. After-tax income from
a dividend payment is [1/(1 t)] * (1 p) while after-tax income from a $1 capital gain is 1 0.5p.
Shareholders prefer dividends provided p < 0.46, for the current corporate tax rate t = 0.3. The top
marginal tax rate in Australia during the period of this study was 0.47 (plus the Medicare levy of
0.0125). Thus all shareholders other than those on the top marginal rate prefer dividend income over
capital gains. Thus the imputation system in Australia substantially reduces the tax disadvantage to
dividends that was present in the US during the period of the Grullon and Michaely (2002) study.
44
may use off-market repurchases as a mechanism to distribute franking credits excess

to the requirements of ordinary dividend policy. We therefore do not expect the
substitution hypothesis to hold for off-market buybacks.
Other Motivations for Share Repurchases
Grullon and Michaely (2004) argue that the two predominant theories explaining
firms motivations for undertaking repurchases are the information/signalling
hypothesis and the free cash flow hypothesis. The information/signalling hypothesis
has its root in the information asymmetries that exist between managers and outsiders. It is based on the idea that managers use share repurchases to signal better
prospects for the company. Dann (1981), Vermaelen (1981) and Comment and Jarrell
(1991) find that the market reacts positively to the announcement of a repurchase, a
result which is consistent with the information/signalling hypothesis. However,
Stephens and Weisbach (1998), Nohel and Tarhan (1998), Ikenberry et al. (2000) and
Grullon and Michaely (2004) provide more recent empirical evidence that does not
support the signalling hypothesis. Mitchell and Dharmawan (2007) find strong
support for signalling incentives in their empirical study of Australian buybacks.
The free cash flow hypothesis is based on the work of Easterbrook (1984) and
Jensen (1986). Repurchases and dividends are mechanisms to distribute excess cash
to shareholders and lower the agency costs of free cash flow. Based on this hypothesis one would expect firms with large excess free cash flow to repurchase more
shares. Stephens and Weisbach (1998), Dittmar (2000), Grullon and Michaely (2004)
find support for the free cash flow hypothesis.
There are a number of other explanations for firms repurchasing behaviour.
Firms with high leverage are less likely to repurchase (Bagwell and Shoven, 1988;
Lie, 2002). In addition, cross-sectional analysis shows that dividends are used to pay
out cash flow that is likely to be permanent whereas share repurchases are used for
more volatile cash flows (Jagannathan et al., 2000; Guay and Harford, 2000).
Grullon and Michaely (2002) directly explore the substitutability of dividends and
share repurchases by examining the correlation of share repurchases with deviations
from expected payout policy. Using the Lintner (1956) model of expected dividends
and controlling for firm characteristics they find strong evidence that for the period
from 1972 to 2000 US firms completed repurchases using funds that would otherwise
have been used to increase dividends. Importantly, market participants are aware of
this substitution, as evidenced by the insignificant impact of the announcement of
dividend decreases on the share price of repurchasing firms.
Summary
The tax environment is a key component in the relation between dividends and
share repurchases (Lie and Lie, 1999; Jacob and Jacob, 2013). In a classical tax
environment such as the US where the proceeds of both on- and off-market
buybacks are taxed as capital gains there are clear incentives for companies to
distribute cash via repurchases. If tax treatment of the proceeds matter, one might
expect on- and off-market repurchases in Australia to have different effects on
the relation between dividends and repurchases. While dividends are, for many
45
A BAC U S
shareholders, preferred over capital gains, it remains an empirical issue as to whether

dividend substitution will be present for the whole sample of buybacks. Nevertheless, we have argued that it is unlikely that dividend substitution will be found for
off-market repurchases.
While our central concern is to study the substitutability of repurchases and
dividends in Australia we control for other variables affecting the firms repurchase
decision.
DATA AND DESCRIPTIVE STATISTICS
We collect data for firms that undertook repurchases over the years 1997 to 2009;
prior to 1997 there were few repurchases (Lamba and Ramsay, 2005). A database
consisting of all on- and off-market share repurchases completed between 1997 and
2009 (inclusive)14 is constructed using data sourced from SIRCA (Securities Industry Research Centre of Asia-Pacific), backed up by the Securities Data Company
(SDC). We believe that we have managed to identify all repurchases undertaken
over this period. The total number of repurchases collected is 899. Companies
purchasing shares on-market make daily statements to the Australian Securities
Exchange detailing the previous days repurchase activity, including the number of
shares repurchased and the average price paid. This rigorous disclosure regime has
also been described by Mitchell and Dharmawan (2007) and Holub and Mitchell
(2012). The on-market repurchase data are aggregated and cross-checked with
company final buyback announcements in order to calculate accurately on-market
repurchase activity over each financial year. The off-market sample is constructed
from the SDC database and supplemented with searches of DatAnalysis.15
Initially we filter out only those transactions where the repurchase yield cannot be
calculated. This results in deletion of 12 observations and a dataset consisting of 887
firm-year observations. We provide descriptive statistics on this dataset, as it provides a comprehensive description of buyback activity in the Australian market since
the relaxing of the rules governing buybacks.
The relevant historical financials16 between 1 January 1996 and 31 December 2009
are obtained from FinAnalysis17 supplemented with information from the AGSM
14
The starting date for the sample period was chosen to avoid any confounding effects of the introduction of the First Corporate Law Simplification Bill in December 1995. The introduction of this Bill
reduced the previously stringent regulations governing share buybacks and consequently lowered the
high transactions costs associated with initiating a share buyback. Prior to the introduction of the bill
only 32 repurchases were undertaken over the period 19891995 (Lamba and Ramsay, 2005).
15
DatAnalysis is provided by Aspect Huntley and contains financial data for all companies listed on the
ASX from 1997 onwards. The search is conducted on the keywords off-market or equal-access,
selective and buyback.
16
Financial items and buyback activity are collected on a firm-specific, financial-year-end basis.
17
FinAnalysis is provided by Aspect Huntley and contains financial data for all companies listed on the
ASX from 1997 to present.
46
Table 1
This table provides a description of financial items. Firm statistics collected from FinAnalysis,
supplemented by the AGSM database for firms delisted between 1995 and 1997.
Item
Description
MV
Market Value of Equity:

Market capitalization of the firm.
BV
Book Value of Shareholders Equity
TA
Total Assets:
Book value of total assets of the firm.
Market to Book ratio:
MBt
MVt + (TAt BVt )

TAt
The ratio of market value of firm to book value.
Casht
Balance sheet cash:
(Casht + short term investmentst )

TAt
NPAT
Levt
The ratio of cash plus short term investments to total assets.

NPAT is reported Net Profit after Tax (pre Abnormal items).
Balance sheet debt:
(short term debt t + long term debt t )

TAt
The leverage of the firm defined as the ratio of short term plus long term debt to total assets.
All figures are measured at the firms balance date (commonly 30 June).
database.18 Financial data at the end of the financial year preceding the year of the
repurchase are collected: market capitalization (MV), book value of shareholders
equity (BV), net profit after tax (NPAT) pre abnormal items, book value of total
assets (TA), cash on the balance sheet (Cash), dividends (Div), and a measure of
leverage (Lev). While the FAB has been disclosed in the annual reports since 1999,
it is not available electronically.19 Given the large number of observations in the
sample the FAB has not been collected for this project. Definitions of the financial
metrics are given in Table 1.
Our full sample of repurchases (denoted as Rep = 1) consists of firms that have
undertaken either an on- or off-market buyback over the period 1997 to 2009.
18
The Australian Graduate School of Management (AGSM) Centre for Research in Finance Share
Price and Price Relative Database is now managed by SIRCA.
19
AASB 1034 was introduced in October 1999, with section 5.3(f) mandating the disclosure of franking
account balances.The adoption of AASB 101, to replace AASB 1034 in July 2004 as part of Australias
adoption of International Financial Reporting Standards, maintained this requirement in Aus 126.5.
However, the information is disclosed in the notes to the accounts and so must be hand collected.
47
A BAC U S
Sample firms are initially categorized on the basis of whether they paid at least one
ordinary dividend (denoted as Div = 1). Table 2 Panel A reports the descriptive
statistics for repurchasing firms that did not pay dividends over the sample period
(denoted as Div = 0, Rep = 1), for firms that did pay dividends (denoted as Div = 1,
Rep = 1) and for the complete sample (denoted as Div = 0, 1, Rep = 1).
Table 2 gives a comprehensive representation of the dividend and repurchase
behaviour of the sample of firms that have conducted repurchases over the data
period. It is of interest to note that, for our sample, the combined dollar payout via
repurchases is similar to that disbursed through dividends.20 Repurchases have
clearly become an important mechanism for some Australian companies to return
cash to shareholders. However, repurchases have not overtaken dividends as the
dominant payout method as Grullon and Michaely (2002) find for US firms.
Although we have captured all repurchases, our sample of repurchasing companies
is a small subset of the complete universe of companies listed on the Australian
Securities Exchange (ASX), and dollars paid out through repurchases constitute a
small proportion of the overall payout of Australian firms.
As highlighted by the work of Guay and Harford (2000), Jagannathan et al. (2000)
and Stephens and Weisbach (1998), the financial characteristics and performance of
a firm may influence its payout policy choices, particularly in the choice between
increasing dividends and undertaking a share repurchase. From Table 2 Panel A, we
observe that the dividend-paying firms in our sample are on average larger, more
profitable and hold less cash than non-dividend paying firms. They have higher MB
ratios, which in combination with their larger size, suggests that they are less likely
to be undervalued (Dittmar, 2000; Ikenberry et al., 1995; Vermaelen, 1981, 1984). For
the overall sample the mean market value of equity is $2.6b while the mean total
assets are $7.7b with mean leverage at 19.1%. This compares to an average reported
leverage of 21.7% for all listed companies on the ASX at end 2008. Comparing
means and medians in Panel A, it is clear that the data are skewed. There are a large
number of small companies in the sample.21
Off-market repurchases, although fewer in number, are important in the choice of
payout mechanism. As shown in Table 2 Panel B, the full sample of 83 off-market
repurchases disbursed $25,879.1m to shareholders while the 823 on-market repurchases disbursed $43,865.4m.22 It is also clear that firms undertaking off-market
repurchases are larger on average than those undertaking on-market repurchases.
We investigate companies that make repeat repurchases, defined as companies
undertaking three or more repurchases over the sample period (485 observations
20
Because we do not have all listed firms in our sample this does not imply that repurchases have
replaced dividends in the Australian market as the dominant payout method.
21
Note that in later analysis repurchases buying back more than 10% are removed from the sample, and
the analysis is performed on financial and non-financial companies. This approach removes the
extreme outliers and is a robustness check of the results.
22
A number of companies undertook both on- and off-market repurchases, so the sum of off-market
and on-market repurchases is greater than the total number of repurchases.
48
Table 2
This table provides a summary of financials for all firms completing buybacks between years 19972009.
Financial statistics for the same financial year that the buyback occurred are obtained from the
FinAnalysis database, supplemented by the AGSM database. Share repurchases calculated from
company announcements to the ASX, sourced from SIRCA and backed up by IRESSs Signal G
database. Div represents the dollar value of ordinary dividends paid during the firms respective financial
years. Rep is the dollar value of both on- and off-market repurchases made during the firms financial year.
Div represents the total dollar value of dividends paid by all repurchasing firms in the sample and Rep
is the total dollar value of repurchases paid by all firms over the period. Other variables are defined in
Table 1.
Panel A: There are 368 companies (with necessary financial data) in the sample completing 887
repurchases over the sample period (Div = 0, 1, Rep = 1). Of those, there are 256 companies completing
repurchases, that also paid dividends over the period (Div = 1, Rep = 1). There are 167 companies
completing 269 repurchases that did not pay dividends (Div = 0, Rep = 1).
Panel B: The same statistics are presented for the on- and off-market samples separately.
Panel A
Div = 0, Rep = 1
Div = 1, Rep = 1
Div = 0, 1, Rep = 1
167
269
0
4726.5
256
618
69601.2
65036.3
368
887
69601.2
69744.5
No. of companies
No of observations
(Div)($m)
(Rep)($m)
MV($m)
TA($m)
MB
Cash (%)
Lev (%)
NPAT($m)
Divs($m)
Reps($m)
Mean
Median
Mean
Median
Mean
Median
309.3
318.3
1.34
29.5
14.8
13.9
0
17.5
28.2
37.9
1.01
19.2
5.4
0.7
0
0.8
3521.1
10875.1
1.57
16.0
20.9
241.3
112.6
105.2
213.5
249.5
1.13
7.7
15.8
15.7
9.9
2.9
2559.7
7673.5
1.50
20.1
19.1
172.3
78.5
78.6
117.3
136.7
1.11
10.4
12.9
8.1
3.6
1.7
Panel B
No. of companies
No of observations
(Div)($m)
(Rep)($m)
MV($m)
TA($m)
MB
Cash(%)
Lev (%)
NPAT($m)
Divs($m)
Reps($m)
Off-market
On-market
61
83
20603.0
25879.1
333
820
51986.1
43865.4
Mean
Median
Mean
Median
10415.2
21949.6
1.37
17.5
23.0
695.1
248.2
311.8
914.6
628.7
1.09
6.4
16.1
58.2
20.5
113.1
1830.9
6402.0
1.51
20.1
18.7
122.0
63.4
53.5
109.7
121.3
1.11
10.6
12.2
7.5
3.2
1.42
49
A BAC U S
and 118 companies). Untabulated results show that these companies have lower
average values of Cash (16.5% versus 20.4%), higher NPAT ($272.8m versus
$172.3m), higher MV ($3.9b versus $2.6b) and higher Div ($120m versus $78.5m).
These comparisons are significant at the 1% level. Companies making repeat repurchases also have significantly higher (at 5%) average repurchase amount ($80.8m
versus $78.6m), and leverage (20.8% versus 19.1%).
The dataset represented in Table 2 is not suitable for an investigation of the
substitution hypothesis for a number of reasons. Under the Corporations Act 2001
(see the Appendix) an Australian company is entitled to buy back its shares without
shareholder approval provided that the total number of shares bought back does not
exceed 10% of the smallest number of shares on issue over the previous 12 months.
We therefore restrict the sample to repurchase yields less than or equal to 10%.23
Companies that buy back more than 10% are often subsequently wound up or
delisted; others are very small companies such as CTEs.24 Our sample construction
therefore tends to remove the outliers.
We first construct the on-market repurchase sample, which consists of 773
announced and completed on-market repurchases by 323 companies. This facilitates
a direct comparison of our results with those for open-market repurchases from the
US (Grullon and Michaely, 2002).25 We also restrict the sample to non-financial
companies (722 firm-year observations and 310 companies).26 Given support for the
substitution hypothesis in the US where the proceeds are taxed as capital gains, we
expect substitution to hold for the Australian sample of on-market buybacks.
We next construct the off-market repurchase sample, which consists of 50 firmyear observations27 conducted by 34 companies. When financial companies are
removed, the off-market non-financial firm sample consists of 41 firm-year observations and 30 companies. Given the different tax treatment of off-market buybacks,
and the likelihood of these buybacks being used to distribute excess tax credits, we
do not expect substitution to hold for this sample.
A sample is also constructed that combines the shares bought back each financial
year via each method for each company. This aggregate sample consists of 805
firm-year observations and 341 companies (745 firm-year observations and 324
23
We experiment with different thresholds for repurchase yield. Our results are not changed materially
by the application of different thresholds around 10%.
24
A Commitment Test Entity (CTE) is a small unprofitable company listed on the ASX based on
commitments to spend the funds raised under an Initial Public Offering (IPO).
25
Note that US studies have tended to use open-market repurchases to investigate substitution (e.g.,
Grullon and Michaely (2002) and Skinner (2008)) largely because open-market repurchases dominate
self-tender offers in both volume and value. In Australia, there are many more on-market repurchases
but off-market repurchases are economically important, so we include both in our analysis.
26
Financial firms that are constrained in various ways by regulations may have different payout
behaviour. We follow Dittmar (2000) and Grullon and Michaely (2002) to create a sub-sample that
omits banks and other financial companies.
27
Note that both Minimum Holding and Employee Share Scheme repurchases were excluded from this
analysis.
50
companies for the non-financials), and is used to answer the question as to whether
in aggregate buybacks are being used by Australian companies as a substitute for
distributing the cash via dividends. We have no prior expectations as to whether
substitution will be found for the full sample, because it consists of a mixture of onand off-market repurchases with the resulting mix of tax treatments. Nevertheless
this exploration provides important insights concerning company payout behaviour.
MEASURING DIVIDEND CHANGES
There are few studies on the explanatory power of models to estimate dividends in
an Australian setting. Early studies include Shevlin (1982). The major difficulty in
using the standard estimation approach of the Lintner (1956) model for forecasting
dividends with Australian data is obtaining a long enough time series of firms with
continuous dividend payments to estimate firm-specific variables.28 Instead we construct a measure of dividend changes as follows.29 We measure the dividend per
share for firm i in period t1, di,t1, and use this to estimate the total dividend payout
in period t. The estimated dividend payout for firm i at time t, Ei,t is given by
Ei ,t = di ,t 1 Number of shares outstanding at time t
(1)
We then use this estimate to measure the change in dividends as
DDivi ,t =
(Divi ,t Ei ,t )
MVi ,t 1
(2)
where MVi,t1 is the market capitalization of the firm in the previous period, Ei,t is the
estimated dollar value of dividends as given in equation (1) and Divi,t represents the
actual (total) dollar value of dividends paid in period t. DDivi,t therefore simply
represents the change in the dividend per share from period t 1 to period t.
We combine firm characteristics that have been found to affect repurchase activity
with the change in dividends as measured by equation (2) into the regression model
described in equation (3). The dependent variable, repurchase yield, is defined as the
dollar value of repurchases in the financial year divided by the market value of
equity at the end of the previous financial year. This variable is estimated accurately
from our data, but is observed only if the firm undertakes a repurchase so the
regression is estimated as a truncated regression model of the form,
28
One approach could be to use a panel regression as in da Silva et al. (2004) to estimate the coefficients
of the Lintner model. This approach was used in an earlier version of this paper and gave realistic
payout ratios. It was not used for this dataset because it gave an unrealistic overall payout ratio.
29
This is a simple model but is not unrealistic. In the survey and interviews conducted by Brav et al.
(2005) 88% of managers very strongly or strongly agree that they consider the level of dividends per
share paid in recent quarters when choosing current dividend policy. 94% strongly or very strongly
agree that they try to avoid reducing dividends. Together, these two results provide strong support for
our simple model.
51
A BAC U S
RYieldi ,t = 1 + 2 Cashi ,t 1 + 3 MBi ,t 1 + 4 ln (TA)i ,t 1

(+)
( )
( )
+ 5 Levi ,t 1 + 6 DDivi ,t + 7Taxt + ui ,

( )
(3)
( )
where we have indicated the expected signs on the coefficients of the independent
variables. Our primary focus is the relation between repurchase yield and DDivi,t. A
positive coefficient (or a non-significant negative) on DDivi,t (positive 6) will result
in the rejection of the substitution hypothesis.
Table 3 summarizes how each of the control variables in equation (3) is measured.
Cash is used to proxy for free cash flow and is expected to have a positive coefficient
(Jensen, 1986; Grullon and Michaely, 2004; Nohel and Tarhan, 1998). If the market to
book ratio (MB) is used to represent the markets assessment of growth opportunities for the firm, it is expected to have a negative coefficient (Grullon et al., 2002;
Grullon and Michaely, 2004). The natural logarithm of total assets ln(TA) has been
used as a proxy for information asymmetry; large firms are less likely to be undervalued (Vermaelen, 1981). Therefore, if firms are using buybacks as a signal to the
market that the firm is undervalued the coefficient on this variable should be
negative. Finally, firms with high debt levels are likely to repurchase less (Bagwell
Table 3
This table explains the variables used in the truncated regression model:
RYieldi,t = 1 + 2Cashi,t1 + 3MBi,t1 + 4ln(TA)i,t1 + 5Levi,t1 + 6DDivi,t + 7Taxt + ui
Variable
RYield
Cash
Description and expected sign

The repurchase yield, defined as the dollar value of repurchases in a firms financial year
divided by the market value of equity at the end of the previous financial year. This is the
dependent variable in the regression and is only observed if a firm undertakes a
repurchase, that is, if repurchase yield is positive.
Cash is used to test the free-cash-flow hypothesis. As defined in Table 1 cash is measured
using balance sheet data. Cash should be positively associated with repurchase yield.
MB
As previously defined in Table 1, this is used to test the investment and information
signalling hypotheses, as it represents the markets assessment of growth opportunities for
the firm. MB should be negatively related to repurchase yield.
Ln(TA)
The log of the book value of total assets of the firm, measured at the end of the previous
financial year to that of the repurchase. This measure is commonly used to test for
information asymmetry so that ln(TA) should be negatively related to repurchase yield
As previously defined in Table 1, this is used to test the leverage hypothesis. Lev should be
negatively related to the repurchase yield
DDiv is the change in dividends measured as the difference between the total dividends
paid in period t and period t1, divided by the market value of equity at the end of
period t 1. If firms are substituting share repurchases for dividends DDiv should be
negatively related to share repurchase yield.
Tax is a time dummy used to investigate whether changes to the tax rates had an effect on
repurchase yields.
Tax1 = 0 for years before 1999 and 1 from 1999 onwards (CGT change)
Tax2 = 0 for years before 2000 and 1 from 2000 onwards (Franking credit rebate change)
Lev
DDiv
Tax
52
and Shoven, 1988; Lie, 2002), so the coefficient on the leverage variable is expected
to be negative. We investigate whether tax changes have had an effect on repurchase
yields with the use of the Tax dummy.
Since observations of zero repurchases are not included in our sample we use the
truncated regression approach pioneered by Hausman and Wise (1976, 1977). This
approach is also used by Dittmar (2000). The repurchase yield model is estimated
using maximum likelihood estimation. We use a quadratic hill-climbing optimization
algorithm with the OLS estimates as the initial starting values for our different
samples. Maximizing the log-likelihood function with respect to the parameters
yields coefficients that capture two effects: an effect on the mean of the dependent
variable (modelled using a latent variable approach) given that it is observed, and an
effect on the probability of the dependent variable being observed. We decompose
these two effects and concentrate on the former because we are interested in
examining substitution in firms that actually undertake repurchases. This approach
can be summarized as follows.
We observe the repurchase yield only when the firm has repurchased shares (Ryield
> 0) and so the dependent variable is left truncated at zero. Neglecting the truncation
can lead to biased estimates of the coefficients.The general model is given as yi = 0 +
1xi + i where yi > 0 and i N (0, 2) and in this case we are interested in the
distribution of yi given that yi > 0. (Greene, 2012) shows that,
E [ yi yi > 0 ] = 0 + 1 xi +
( ( 0 + 1 xi ) )
.
1 ( ( 0 + 1 xi ) )
(4)
The last term on the right-hand side of equation (4) is called the Mills ratio, and can
be used to obtain the marginal effects in the population where yi > 0.
RESULTS
Before analyzing the regression results we first calculate the simple correlation
between repurchase yield and changes in dividends for each firm over the sample
period. The correlation across the whole sample is -2.7%. The small negative correlation gives some weak preliminary evidence that the substitution hypothesis may
hold in the Australian market. In order to control for other variables that are
expected to influence repurchase activity we estimate equation (3). Note that interpretation of the marginal effects from the truncated regression model requires a
transformation of the coefficient estimates.
Because we have different expectations regarding the sign of the coefficient on the
dividend change variable DDiv we run the regression given by equation (3) separately for on- and off-market repurchases, restricting the repurchase yield to be less
than or equal to 10%. Heteroscedasticity in the errors can cause significant problems
with truncated regression model estimation (Greene, 2012). We therefore use
Huber-White standard errors and covariances.
Table 4 Panel A documents the results for the on-market repurchase sample (773
observations) and Panel B for the same sample excluding banks and other financial
53
Predicted
Sign
Variable
Intercept
Cash
54
MB
ln(TA)
Lev
DDiv
0.1083
13.6700
0.0178
Likelihood ratio
LR p-value
1.6137
(0.0403)
1.8329***
(3.3983)
0.1073
1.6006
0.0418
LR p-value
Likelihood ratio
(0.5149)
1.8428***
(3.5573)
0.0340
(0.6916)
0.3750
0.0421
0.5028
(1.3075)
0.1463**
3.0095***
(3.2686)
0.0044
(0.1011)
0.0195
0.2140
0.7792
2.3547
Median
Coefficient
(2.1407)
0.2148
0.7864
2.3668
Average
Marginal effects
0.0079
15.6532
0.1071
1.5997
0.0418
0.2125
0.7729
2.3500
Average
0.1071
1.6031
0.0419
0.2124
0.7745
2.3468
Median
Marginal effects
Panel B: On-market repurchases excluding

financial firms (722 observations)
(2.1022)
0.6099*
(1.6515)
0.1418**
2.3715***
(2.8729)
Coefficient
Panel A: On-market repurchases

(773 observations)
This table presents truncated regression results for on-market repurchases with repurchase yield set to be less than or equal to 10%. Panel A includes all
firm-year observations satisfying these restrictions (773 Observations), while Panel B removes financial firms and contains 722 observations. The regression
takes the form detailed in Table 3 and equation (3). Coefficients are estimated using Maximum Likelihood Estimation (Quadratic hill-climbing optimization
algorithm) with convergence achieved in 3 iterations. Predicted signs link to the arguments in Section 3. Z-statistics are reported in parenthesis and are
calculated from Huber/White heteroscedastic consistent errors. Coefficients significant at 10%, 5% and 1% levels are respectively marked with *, ** and ***
in superscripts. Marginal effects are calculated as average and median of the truncated sample.
Table 4
A BAC U S
companies (722 observations). The size of the coefficients and the levels of significance are approximately the same in the two panels. The coefficient on dividend
changes is negative and significant providing support for the dividend substitution
hypothesis.That is,companies engaging in on-market repurchases tend to reduce their
dividends. The coefficient on MB is significant at 5%; firms with greater growth
opportunities repurchase less. The tax dummy (Tax1) used in this regression is found
to be insignificant and is not included in the Table. The results documented in Table 4
are robust to running the regressions on a sample restricted to companies that pay
dividends.
Table 5 documents the results for off-market repurchases. Here the results are
markedly different from those for on-market repurchases and as expected do not
support the dividend substitution hypothesis. The positive and significant coefficient
on DDiv for both samples (Panel A the full sample and Panel B excluding financial
firms) indicates that the repurchase yield for off-market repurchases is positively
and significantly related to changes in dividends. In other words, changes in repurchase yield and dividends move in the same direction. This is consistent with the
previous argument that companies are using off-market repurchases to distribute
excess franking credits, and the observation that large dividend paying companies
tend to undertake off-market repurchases. To the extent that a company has franking credits over and above those needed for ordinary dividend policy, off-market
repurchases and increases in dividend payments are possible mechanisms to distribute the franking credits. Large companies in Australia undertaking off-market repurchases often do so to distribute franking credits to shareholders (Brown and
Norman, 2010; Brown and Davis, 2012). The results also show that off-market
repurchase yield is negatively related to size of the company; larger companies buy
back a smaller proportion. The tax dummy (Tax2) used in this regression is found to
be insignificant and is not included in the Table.
Overall results on control variables in both Tables 4 and 5 are consistent with
other studies. Summing up, we find evidence consistent with previous studies conducted in the US that on-market repurchase yields tend to be negatively related to
changes in dividends. However, we find no evidence of substitution for off-market
repurchases. We argue that this result is driven by the need for these companies to
distribute excess imputation tax credits. In this case changes in dividends and
repurchase yields tend to move in the same direction.
Because there are a number of companies that undertake both types of buyback
in the fiscal year, in order to investigate whether in aggregate a substitution effect
exists for Australian repurchases the total dollars repurchased via the two types need
to be combined. Table 6 Panel A summarizes the results for repurchases (combining
on- and off-market total for each company for each financial year restricting repurchase yield to be less than or equal to 10%), and also gives the marginal effects. Panel
B reports the results when the sample excludes financial firms.
In both panels we find that repurchase yield is negatively and significantly related
to dividend changes, after controlling for firm characteristics that influence
repurchase yield, suggesting that there is an overall substitution effect for Australian
firms. The results are also robust to running the regressions only on the sample of
55
Predicted
Sign
Variable
Intercept
Cash
56
MB
ln(TA)
Lev
DDiv
38.1662
15.5168
0.0084
Likelihood ratio
LR p-value
3.0593
(0.2946)
33.7150***
(2.9621)
38.1921
3.0607
0.6003
LR p-value
Likelihood ratio
(0.0477)
34.2562***
(2.9763)
0.4927**
(2.0833)
0.1765
0.6000
19.4651
(1.3019)
0.4383
17.3879***
(3.5010)
0.5877***
(2.6474)
0.5589
2.1394
13.0517
23.0537
Median
Coefficient
(0.7521)
2.1383
13.0504
23.0444
Average
Marginal effects
0.0989
9.2667
38.1637
3.0589
0.6001
2.1385
13.0523
23.0489
Average
38.1988
3.0609
0.6003
2.1395
13.0589
23.0552
Median
Marginal effects
Panel B: Off-market repurchases

excluding financial firms (41 observations)
(0.6741)
18.2317
(1.2762)
0.3768
19.0943***
(3.9148)
Coefficient
Panel A: Off-market repurchases

(50 observations)
This table presents truncated regression results for off-market repurchases with repurchase yield set to be less than 10%. Panel A includes all firm-year
observations satisfying these restrictions (50 observations), while Panel B removes financial firms and contains 41 observations.The regression takes the form
detailed in Table 3 and equation (3). Coefficients are estimated using Maximum Likelihood Estimation (Quadratic hill-climbing optimization algorithm)
with convergence achieved in four iterations. Predicted signs link to the arguments in Section 3. Z-statistics are reported in parenthesis and are calculated
from Huber/White heteroscedastic consistent errors. Coefficients significant at 10%, 5% and 1% levels are respectively marked with *, ** and *** in
superscripts. Marginal effects are calculated as average and median of the truncated sample.
Table 5
A BAC U S
MB
ln(TA)
Lev
DDiv
Intercept
Cash
Predicted
Sign
Variable
10.8338
0.0539
LR p-value
1.3597
0.5001
1.9550***
(3.6850)
Likelihood ratio
0.0919
0.0409
0.4157
0.9716
4.4893
Average
1.3643
0.5635
0.0924
0.4169
0.9760
4.5098
Median
Marginal effects
(0.9448)
0.0630
(0.1310)
(0.8239)
0.1251***
(3.8521)
1.7111**
(2.1102)
0.3126
Coefficient
Panel A: 805 repurchases

includes financial firms
LR p-value
1.9719***
(3.8458)
Likelihood ratio
(0.5132)
0.4118
(0.5513)
0.0255
(0.5583)
0.1281*
(1.8670)
2.0720**
(2.2200)
0.2321
Coefficient
0.0617
10.5229
1.3600
0.5647
0.0923
0.4176
0.9761
4.5037
Average
1.3663
0.5662
0.0927
0.4187
0.9817
4.5250
Median
Marginal effects
Panel B: 747 repurchases

excludes financial firms
This table presents truncated regression results for the combined repurchase sample restricted to repurchase yield less than or equal to 10% consisting of
805 firm-year observations (Panel A) and for the same sample with financial firms removed in Panel B (747 observations). The regressions take the form
detailed in Table 3 and equation (3). Coefficients are estimated using Maximum Likelihood Estimation (Quadratic hill-climbing optimization algorithm)
with convergence achieved in 3 iterations. Predicted signs link to the arguments in Section 3. Z-statistics are reported in parenthesis and are calculated from
Huber/White heteroscedastic consistent errors. Coefficients significant at 10%, 5% and 1% levels are respectively marked with *, ** and *** in superscripts.
Marginal effects are calculated as average and median of the truncated sample.
Table 6
57
A BAC U S
firms that have paid dividends.30 The results in Table 6 provide support for the
investment hypothesis. Firms with greater growth opportunities (higher MB) repurchase less, for both the full sample (Panel A) (at a 1% level of significance) and the
subsample (Panel B) (at a 10% level of significance). If MB is a proxy for investment
opportunities then one might expect cash to also have an important influence on
repurchase yield. In this regard, contrary to the findings of Nohel and Tarhan (1998)
and Grullon and Michaely (2004) we do not find support for the free-cash-flow
hypothesis. The coefficients on our measure for free cash flow are positive but not
significant and are consistent with the findings of Mitchell and Dharmawan (2007) in
their study of Australian on-market repurchases. Our results are also similar to those
of Dittmar (2000), who finds a positive but insignificant coefficient on the cash
variable for a number of the years in her study of US open-market repurchases. The
management surveys of Brav et al. (2005) and Mitchell et al. (2001) suggest that
management is unlikely to use payout policy to impose self-discipline, a prediction
that is supported by our findings. The signalling hypothesis states that larger firms
should repurchase less, because they have less information asymmetry. We find no
support for this hypothesis with the coefficient on the size variable insignificant in
both panels. Leverage is also insignificant for both samples.
CONCLUSIONS
There is evidence that US firms substitute share repurchases for dividends (Grullon
and Michaely, 2002; Skinner, 2008). This paper examines share buybacks in Australia
and adds evidence on whether dividend substitution is prevalent in a country where
dividends are not as tax disadvantaged as in countries where a classical tax system
prevails. Australia provides an ideal environment to explore the effects of the tax
system on the question of dividend substitution. The proceeds from on-market
repurchases are taxed as capital gains, but the proceeds from selling shares into an
off-market repurchase carry tax credits. The differing tax treatment of on- and
off-market repurchases in Australia offers a natural setting to investigate whether
the taxation treatment of repurchase proceeds affects the extent to which firms will
substitute repurchases for dividends.
We make an important contribution to the literature on whether taxes affect
payout decisions. First, for on-market (open-market) buybacks in Australia, which
are taxed as if in a classical tax system, the relation between share repurchase
yields and dividend changes is negative. This mirrors the results from studies conducted on US open-market repurchases and supports the dividend substitution
hypothesis. Second, yields on Australian off-market repurchases (self-tender
offers), which carry significant tax benefits for participating shareholders, tend to
move in the same direction as changes in dividends. That is, off-market repurchase
yields increase (decrease) when dividends increase (decrease); the dividend
30
The results are quantitatively similar when the model is estimated on the sample of frequent
repurchasers. For non-financials Lev and DDiv are significant at 5%. When financial companies are
included DDiv is significant at 5%.
58
substitution hypothesis does not hold for off-market repurchases in Australia. We

argue that off-market repurchases are used to distribute imputation tax credits
excess to the needs of ordinary dividend policy and therefore do not substitute for
ordinary dividends. Firms have an incentive to distribute the tax credits because
they are valuable to the shareholders who receive them. The findings are consistent with those of Jacob and Jacob (2013) that the differential tax treatment of
dividends and capital gains matters for payout choice.
For the sample of firms that have undertaken a buyback, irrespective of the type,
there is support for the view that in aggregate repurchases and dividends tend to act
as substitutes. This finding suggests that the increase in repurchase activity witnessed
after 1996 in Australia has been financed with funds provided by changes in dividend
policy, consistent with the patterns of dividend payouts observed by Coulton and
Ruddock (2011).
Our results must be tempered with the observation that repurchasing firms in
Australia make up a small proportion of all firms, and the dollar distribution
through repurchases remains a small proportion of the overall cash distribution of
Australian companies. Therefore repurchases have not replaced dividends as the
most important form of payout. Nevertheless, for firms that undertake repurchases
the amount of cash distributed to shareholders is comparable to that distributed
via dividends, and particularly so for large firms undertaking off-market repurchases. Our results are important because they show that, holding all else constant,
the relation between repurchases and dividends depends on the way that the
repurchase proceeds are taxed. Our study thus adds to a growing body of
literature that suggests that taxation plays an important role in understanding firm
payout decisions.
REFERENCES
AuYong, H., C. Brown, and C. Ho (2014), Off-market Buybacks in Australia: Evidence of Abnormal
Trading around Key Dates, International Review of Finance, Vol. 14, No. 4, pp. 551585.
Bagwell, L. and J. Shoven, (1988), Share Repurchases and Acquisitions: An Analysis of which Firms
Participate, in A. J. Auerbach (ed.), Corporate Takeovers: Causes and Consequences, University of
Chicago Press, Chicago, pp. 191220.
Baker, M. and J. Wurgler (2004a), Appearing and Disappearing Dividends: The Link to Catering
Incentives, Journal of Financial Economics, Vol. 73, No. 2, pp. 27188.
(2004b), A Catering Theory of Dividends, Journal of Finance, Vol. 59, No. 3, pp. 1125
65.
Balachandran, B. and R. Faff (2004), Special Dividends, Bonus Issues and On-market Buybacks: Signalling in an Imputation Environment, Finance Letters, Vol. 2, No. 1, pp. 2326.
Barclay, M. and C. Smith Jr (1988), Corporate Payout Policy: Cash Dividends Versus Open-market
Repurchases, Journal of Financial Economics, Vol. 22, pp. 6182.
Black, F. (1976), The Dividend Puzzle, Journal of Portfolio Management, Vol. 2, No. 2, pp. 58.
Brav, A., J. Graham, C. Harvey, and R. Michaely (2005), Payout Policy in the 21st Century, Journal of
Financial Economics, Vol. 77, No. 3, pp. 483527.
Brown, C. and D. Norman (2010), Management Choice of Buyback Method: Australian Evidence,
Accounting and Finance, Vol. 50, No. 4, pp. 76782.
Brown, C. and K. Davis (2012), Taxes, Tenders and the Design of Australian Off-market Share Repurchases Accounting and Finance, Vol. 52, No. 1, pp. 10935.
59
A BAC U S
Brown, C. and K. Efthim (2005), The Effect of Taxation on Equal Access Share Buybacks, International
Review of Finance, Vol. 5, Nos 34, pp. 199218.
Brown, R. and P. Howard (1992), Dividend Policy and Capital Structure under the Imputation Tax
System: Some Clarifying Comments, Accounting and Finance, Vol. 32, pp. 5162.
Chetty, R. and E. Saez (2005), Dividend Taxes and Corporate Behavior: Evidence From The 2003
Dividend Tax Cut, The Quarterly Journal of Economics, Vol. 52, pp. 791833.
Coleman, L., K. Maheswaran, and S. Pinder (2010), Narratives in Managers Corporate Finance Decisions, Accounting and Finance, Vol. 50, No. 3, pp. 605633.
Comment, R. and G. Jarrell (1991), The Relative Signalling Power of Dutch-auction and Fixed-price
Self-tender Offers and Open-market Share Repurchases, Journal of Finance, Vol. 46, No. 4,
pp. 124371.
Coulton, J. and C. Ruddock (2011), Corporate Payout Policy in Australia and a Test of the Life-cycle
Theory, Accounting and Finance, Vol. 51, No. 2, pp. 381407.
Da Silva, L., M. Goergen, and L. Renneboog (2004), Dividend Policy, Earnings, and Cash Flow:
A Dynamic Panel Data Analysis in Dividend Policy and Corporate Governance, Oxford University
Press, Oxford.
Dann, L. (1981), Common Stock Repurchases: An Analysis of Returns to Bondholders and Stockholders, Journal of Financial Economics, Vol. 9, No. 2, pp. 11338.
De Angelo, H., L. De Angelo, and D. Skinner (2004), Are Dividends Disappearing? Dividend Concentration and the Consolidation of Earnings, Journal of Financial Economics, Vol. 72, No. 3,
pp. 425456.
Dittmar, A. (2000), Why Do Firms Repurchase Stock? Journal of Business, Vol. 73, No. 3, pp. 33155.
Easterbrook, F. (1984), Two Agency-cost Explanations of Dividends, American Economic Review, Vol.
74, No. 4, pp. 65059.
Fama, E. and K. French (2001), Disappearing Dividends: Changing Firm Characteristics or Lower
Propensity to Pay?, Journal of Financial Economics, Vol. 60, No. 1, pp. 344.
Floyd, E., N. Li, and D. Skinner (2012), Payout Policy Through the Financial Crisis: The Growth of
Repurchases and the Resilience of Dividends, Working Paper No. 12-01, The University of Chicago
Booth School of Business.
Greene, W. (2012), Econometric Analysis, 7th Ed., Prentice-Hall, New Jersey.
Grullon, G. and R. Michaely (2002), Dividends, Share Repurchases and the Substitution Hypothesis,
Journal of Finance, Vol. 57, No. 4, pp. 164984.
(2004), The Information Content of Share Repurchase Programs, Journal of Finance, Vol. 59, No.
2, pp. 65180.
Grullon, G., R. Michaely, and B. Swaminathan (2002), Are Dividend Changes a Sign of Firm Maturity,
Journal of Business, Vol. 75, No. 3, pp. 387424.
Guay, W. and J. Harford (2000), The Cash-Flow Permanence and Information Content of Dividend
Increases Versus Repurchases, Journal of Financial Economics, Vol. 57, No. 3, pp. 385415.
Harris, T. and I. M. Ramsay (1995), An Empirical Investigation of Australian Share Buybacks,
Australian Journal of Corporate Law, Vol. 4, No. 4, pp. 393416.
Hausman, J. and D. Wise (1976), The Evaluation of Results from Truncated Samples: The New Jersey
Negative Income Tax Experiment, Annals of Economics and Social Measurement, Vol. 5, No. 4,
pp. 421445.
(1977), Social Experimentation, Truncated Distributions and Efficient Estimation, Econometrica,
Vol. 45, No. 4, pp. 31939.
Hoberg, G. and N. Prabhala (2009), Disappearing Dividends, Catering and Risk, The Review of Financial
Studies, Vol. 22, No. 1, pp. 79116.
Holub, M and J. Mitchell (2012), On-Market Share Buy-Backs: ASX Disclosure Requirements and
Compliance, Abacus, Vol. 48, No. 1, pp. 3158.
Ikenberry, D. L., J. Lakonishok, and T. Vermaelen (1995), Market Under-reaction to Open-market Share
Repurchases, Journal of Financial Economics, Vol. 39, Nos. 23, pp. 181208.
60
(2000), Stock Repurchases in Canada: Performance and Strategic Trading, Journal of Finance,
Vol. 55, No. 5, pp. 23732397.
Jacob, M. and M. Jacob (2013), Taxation, Dividends, and Share Repurchases: Taking Evidence Global,
Journal of Financial and Quantitative Analysis, Vol. 48, No. 4, pp. 124169.
Jagannathan, M., C. Stephens and M. Weisbach (2000), Financial Flexibility and the Choice Between
Dividends and Stock Repurchases, Journal of Financial Economics, Vol. 57, No. 3, pp. 35584.
Jensen, M. (1986), Agency Costs of Free Cash Flow, Corporate Finance and Takeovers, American
Economic Review, Vol. 76, No. 2, pp. 32329.
Julio, B. and D. Ikenberry (2004),Reappearing Dividends, Journal of Applied Corporate Finance,Vol. 16,
No. 4, pp. 89100.
Lamba, A. S. and I. Ramsay (2005), Comparing Share Buybacks in Highly Regulated and Less Regulated
Market Environments, Australian Journal of Corporate Law, Vol. 17, pp. 26180.
Lie, E. (2002), Do Firms Undertake Self-Tender Offers to Maximize Capital Structure? The Journal of
Business, Vol. 75, No. 4, pp. 60939.
Lie, E. and H. Lie (1999), The Role of Personal Taxes in Corporate Decisions: An Empirical Analysis of
Share Repurchases and Dividends, Journal of Financial and Quantitative Analysis, Vol, 34, No. 4,
pp. 53352.
Lintner, J. (1956), Distribution of Incomes of Corporations Among Dividends, Retained Earnings and
Taxes, American Economic Review, Vol. 46, pp. 97113.
Miller, M., and F. Modigliani (1961), Dividend Policy, Growth and the Valuation of Shares, Journal of
Business, Vol. 34, No. 4, pp. 411433.
Mitchell, J. and G. Dharmawan (2007), Incentives for On-market Buy-macks: Evidence from a Transparent Buy-back Regime, Journal of Corporate Finance, Vol. 13, pp. 14669.
Mitchell, J. and P. Robinson (1999), Motivations of Australian Listed Companies Effecting Buybacks,
Abacus, Vol. 35, No. 1, pp. 91119.
Mitchell, J., G. Dharmawan, and A. Clarke (2001), Managements Views on Share Buy-backs:
An Australian Survey, Accounting and Finance, Vol. 41, Nos 12, pp. 93129.
Monkhouse, P. (1993), The Cost of Equity Under the Australian Dividend Imputation Tax System,
Accounting and Finance, Vol. 33, pp. 118.
Nohel, T. and V. Tarhan (1998), Share Repurchases and Firm Performance: New Evidence on the Agency
Costs of Free Cash Flow, Journal of Financial Economics, Vol. 19, pp. 187222.
Officer, R. (1994), The Cost of Capital of a Company Under an Imputation Tax System, Accounting and
Finance, Vol. 34, No. 1, pp. 117.
Partington, G. (1989), Variables Influencing Dividend Policy in Australia: Survey Results, Journal of
Business Finance and Accounting, Vol. 16, No. 2, pp. 16582.
Pattenden, K. and G. Twite (2008), Taxes and Dividend Policy under Alternative Tax Regimes, Journal
of Corporate Finance, Vol. 14, No. 1, pp. 116.
Rau, P. and T. Vermaelen (2002), Regulation, Taxes and Share Repurchases in the United Kingdom,
Journal of Business, Vol. 75, No. 2, pp. 24582.
Shevlin, T. (1982), Australian Corporate Dividend Policy: Empirical Evidence, Accounting and Finance,
Vol. 22. No. 1, pp. 122.
Skinner, D. (2008), The Evolving Relation Between Earnings, Dividends, and Stock Repurchases,
Journal of Financial Economics, Vol. 87, No. 3, pp. 582609.
Stephens, C. and M. Weisbach (1998), Actual Share Reacquisitions in Open-Market Repurchase
Programs, Journal of Finance, Vol. 53, No. 1, pp. 31333.
Vermaelen, T. (1981), Common Stock Repurchases and Market Signalling, Journal of Financial
Economics, Vol. 9, No. 2, pp. 13983.
(1984), Repurchase Tender Offers, Signalling and Managerial Incentives, Journal of Financial and
Quantitative Analysis, Vol. 19, No. 2, pp. 16381.
61
A BAC U S
APPENDIX 1
The table describes the types of share repurchases permitted for ASX (Australian Stock Exchange)-listed
companies in Australia under the Corporations Act 2001. To undertake a repurchase, companies must
comply with Chapter 2J, Part 2J.1, Division 2 of the Corporations Act 2001. Each type involves different
legal and disclosure formalities. In general, companies are able to repurchase up to 10% of their ordinary
shares in any 12-month period (commonly referred to as the 10/12 limit) and once the transaction is
completed the shares must be cancelled and the Australian Securities and Investment Commission
(ASIC) notified. Company conduct during the repurchase is constrained by ASX Listing Rules 3.8A, 7.29
and 7.33.
Type of
repurchase
Description
On-market
Repurchases undertaken in the course of ordinary trading on the

Australian Stock Exchange (ASX). Shareholder approval via an
ordinary resolution is only required if the 10/12 limit is exceeded. Legal
requirements of share buybacks in Australia are currently contained in s.
257 of the Corporations Act 2001, and company conduct during the
repurchase is constrained by Listing Rules 3.8A, 7.29 and 7.33.31
Cancellation of the shares immediately after they have been registered
to the company is governed by Sections 257H and 254Y of The
Corporations Act 2001.
Equal access
Repurchases where the company makes an offer to each shareholder to

repurchase some or all of each shareholders ordinary shares. Usually
they are conducted off-market and set at a price specified by the
company. The proposed repurchase must be approved by a
shareholder-approved ordinary resolution if it exceeds the 10/12 limit.
Off-market repurchases made by a company where shares are acquired
from specified shareholders, to the exclusion of others, at a specified
price. Offers may also pertain to holders of shares other than ordinary
shares. A selective buyback must be approved by all 75% of
shareholders (with no votes being cast by those holders whose shares
are to be repurchased).
Selective
Minimum holding
Off-market repurchases of all of a holders shares if the number of shares

held is less than a marketable parcel. No resolution is required.
Employee share scheme
The acquisition of shares in a company by, or on behalf of, employees or

directors who are employed by the company, or a related corporate
body. Shareholder approval via an ordinary resolution is only required if
the 10/12 limit is exceeded.
31
The consequence of this legal framework can be summarized as follows: when companies announce
the buyback they are required to lodge a proforma announcement notice (Appendix 3C). Under
Listing Rule 3.8A a company undertaking an on-market buyback must lodge an Appendix 3E, which
is a daily notification at least half an hour before the start of trading on the business day after which
any shares are bought back. Companies may repurchase shares only if transactions in the companys
shares were recorded on ASX on at least 5 days in the three months before it buys back shares (Listing
Rule 7.29). A company may buy back shares at a price which is not more than 5% above the average
of the market price for securities in that class (Listing Rule 7.33).
62

Brown Et Al 2015 Abacus

Diunggah oleh

Informasi Dokumen

Deskripsi Asli:

Hak Cipta

Format Tersedia

Bagikan dokumen Ini

Bagikan atau Tanam Dokumen

Opsi Berbagi

Apakah menurut Anda dokumen ini bermanfaat?

Apakah konten ini tidak pantas?

Hak Cipta:

Format Tersedia

Brown Et Al 2015 Abacus

Diunggah oleh

Hak Cipta:

Format Tersedia

bs_bs_banner

ABACUS, Vol. 51, No. 1, 2015

CHRISTINE BROWN, JOHN HANDLEY, AND JAMES ODAY

The Dividend Substitution Hypothesis:

In a perfect capital market firms are indifferent to either dividends or

Christine Brown (christine.brown@monash.edu) is at the Department of Banking and Finance, Monash

US evidence is mixed as to the extent to which differential taxation treatment of

evidence of substitution. In fact, increases in repurchase yield are significantly and

Australian resident individuals, complying superannuation (pension) funds, registered organizations

Australia operates a full dividend imputation system as discussed in the previous

may use off-market repurchases as a mechanism to distribute franking credits excess

shareholders, preferred over capital gains, it remains an empirical issue as to whether

Market Value of Equity:

Book Value of Shareholders Equity

MVt + (TAt BVt )

Balance sheet cash:

(Casht + short term investmentst )

The ratio of cash plus short term investments to total assets.

(short term debt t + long term debt t )

Ei ,t = di ,t 1 Number of shares outstanding at time t

We then use this estimate to measure the change in dividends as

RYieldi ,t = 1 + 2 Cashi ,t 1 + 3 MBi ,t 1 + 4 ln (TA)i ,t 1

+ 5 Levi ,t 1 + 6 DDivi ,t + 7Taxt + ui ,

Description and expected sign

2015 Accounting Foundation, The University of Sydney

Panel B: On-market repurchases excluding

Panel A: On-market repurchases

2015 Accounting Foundation, The University of Sydney

Panel B: Off-market repurchases

Panel A: Off-market repurchases

Panel A: 805 repurchases

Panel B: 747 repurchases

2015 Accounting Foundation, The University of Sydney

substitution hypothesis does not hold for off-market repurchases in Australia. We

Repurchases undertaken in the course of ordinary trading on the

Repurchases where the company makes an offer to each shareholder to

Off-market repurchases of all of a holders shares if the number of shares

Employee share scheme

The acquisition of shares in a company by, or on behalf of, employees or

Anda mungkin juga menyukai