Journal of Business Finance & Accounting, 25(9) & (10), November/December 1998, 0306-686X

DOES THE PECKING ORDER HYPOTHESIS EXPLAIN

THE DIVIDEND PAYOUT RATIOS OF FIRMS IN
THE UK?

Abimbola Adedeji*

INTRODUCTION

This study tests the prediction of the pecking order hypothesis (hereafter, the
hypothesis) that there is a negative interaction between the long term values of
dividend payout ratio and investment. The study is undertaken because there
is little information in the literature about the empirical content of this predic-
tion. Previous tests of the hypothesis by Baskin (1989) and Allen (1993) are not
sufficiently comprehensive because while they examine the effects of dividend
yield and investment on financial leverage, and the effects of dividend yield
and financial leverage on rate of investment growth, they do not examine the
effects of financial leverage and investment on either dividend payout ratio or
dividend yield. Similarly, a test of the hypothesis by Ang and Jung (1993) only
examines the preferences of managers of firms in South Korea for the use of
internal equity, debt and external equity to finance investment, but it does
not assess the interrelationship between dividend payout ratio and invest-
ment. Therefore, the results of these studies do not give sufficient indication
as to whether there is a negative interaction between the long term values of
dividend payout ratio and investment.
Another motivation for the study is the observation of Baskin and Allen that
there is a positive interrelationship between financial leverage and investment
in the US and Australia respectively. Ang and Jung also find that firms in
South Korea prefer to use debt than internal equity to finance investment.
These observations are interesting because the pecking order hypothesis does
not make a definitive prediction about the relationship between investment
and financial leverage. However, because of differences in institutional con-

* The author is from the University of the West of England. He is grateful to Norman Strong and
Martin Walker in the School of Accounting and Finance at the University of Manchester, for use-
ful comments. He also thanks the participants at the British Accounting Association Annual Con-
ference held in Cardiff in 1996, the internal seminar of the Institute for Studies in Finance (ISF) at
the University of Essex held in 1997 and the ICAEW-ACCA-JBFA Capital Markets Conference
held at the Lake District in 1998, for their comments on a previous version of this paper, and
thanks Jerry Reid, Bret Giddings and Peter Allott in the Computing Service at the University of
Essex, and Randy Banks and Mark Taylor in the BHPS Research Centre at the University of Es-
sex, for their kind help.
Address for correspondence: Abimbola Adedeji, Bristol Business School, University of the
West of England, Coldharbour Lane, Bristol BS16 1QY, UK.
e-mail: [email protected]

ß Blackwell Publishers Ltd. 1998, 108 Cowley Road, Oxford OX4 1JF, UK
and 350 Main Street, Malden, MA 02148, USA. 1127
1128 ADEDEJI

texts, the empirical relationship between financial leverage and investment in

the UK may be different from what obtains in the US or South Korea. There-
fore, it will be useful to examine the interrelationship between financial lever-
age and investment in the UK.
The prediction of the hypothesis referred to above and the indication of the
previous evidence that there is a positive interrelationship between financial
leverage and investment are tested on a sample of 224 firms in the UK over
the period 1993-96. The test is comprehensive in the sense that it covers the
interaction of dividend payout ratio, financial leverage and investment with
one another. Results obtained show that dividend payout ratio has the pre-
dicted negative interaction with investment and the expected positive interac-
tion with financial leverage. The results also show that there is no significant
interaction between financial leverage and investment. While investment has
a positive influence on financial leverage, financial leverage does not have a
significant influence on investment. It is also observed that irrecoverable ad-
vance corporation tax (ACT) has a positive but weak influence, and overseas
profit has a negative influence, on dividend payout ratio.
The rest of the paper is organised as follows: The pecking order hypothesis
and the UK institutional context are discussed in the next section. This is fol-
lowed by the description of the methodology and the sample. The penultimate
section contains the analyses and results, followed by the conclusions of the
study in the final section.

THE PECKING ORDER HYPOTHESIS

The pecking order hypothesis suggests that firms prefer to use internal equity
to pay dividends and implement growth opportunities. The hypothesis also
suggests that if external finance is needed, firms prefer to raise debt before ex-
ternal equity (Donaldson, 1961; Myers, 1984; and Myers and Majluf, 1984).
There are two divergent views in the literature about why firms prefer internal
equity to external finance. Donaldson (1961) suggests that internal equity is
preferred because firms want to avoid flotation costs which usually accom-
pany external finance. He also suggests that firms prefer debt to external equi-
ty, if external finance is needed, because the flotation cost of debt is usually less
than that of external equity.
Myers (1984) and Myers and Majluf (MM) (1984) disagree with the view
that firms prefer internal equity to debt because of flotation costs. They argue
that the net benefits of debt financing, in terms of tax shield and risk of finan-
cial distress, are likely to outweigh flotation costs. MM are of the view that
firms rely on internal funds because they want to maximise existing share-
holders wealth. They suggest that sale of new shares is not in the interest of
existing shareholders because it usually leads to a decrease in the market price
of the existing shares. Evidence observed by Marsh (1979 and 1982), Masulis

ß Blackwell Publishers Ltd 1998

 DIVIDEND PAYOUT RATIOS OF UK FIRMS 1129

and Korwar (1986) and Mikkelson and Partch (1986) is consistent with this
suggestion.
MM also argue that firms prefer debt to external equity, if they require ex-
ternal finance, because an issue of a risk free debt will not have any impact on
the value of existing shares. It is further argued that even if debt is risky, the
impact of its issue on the value of existing shares will be less than that of an issue
of new shares because the priority of claims usually accorded debt will make
the value of a new risky debt to be less sensitive to the release of new informa-
tion than the value of new shares.
Despite their differences, the explanations of both Donaldson and MM lead
to the same conclusion that firms relate their profitability and growth oppor-
tunities to their long term target dividend payout ratios in order to minimise
the need for external finance. A number of testable predictions follow from this
conclusion. One of them is that profitability has a negative influence on finan-
cial leverage, since a firm which can generate more earnings will borrow less,
all things being equal. Another one is that there is a negative interaction be-
tween long term dividend payout ratio and investment, since a high dividend
payout ratio will lead to a low level of retained earnings and a greater ten-
dency to raise money from external sources to finance growth opportunities.
However, the explanations above do not lead to clear-cut predictions about
the relationship of financial leverage to either dividend payout ratio or invest-
ment because the nature of their relationship depends on how firms respond to
earnings shortages. If firms respond to earnings shortages by borrowing to pay
dividends and finance growth opportunities on a cumulative basis, then, the
long term value of dividend payout ratio and investment should have a posi-
tive influence on financial leverage. Conversely, if firms respond to earnings
shortages by reducing or postponing investment, while borrowing to pay divi-
dends in the short term because of reluctance to cut dividends, financial lever-
age may have a positive relationship with dividend payout ratio, and a
negative relationship with investment, in the long term. But, if, over time,
earnings shortage persists, or there is a change in growth opportunities, firms
will be expected to adjust their dividend payout ratios to their new levels of
earnings and growth opportunities. This is the logic behind the prediction that
there is a negative interaction between the long term values of dividend pay-
out ratio and investment.

Institutional Context
The suggestion of the pecking order hypothesis that only internal funding
pressure makes firms use external finance is questionable because it contra-
dicts some other theories and ignores the effects of various institutional factors
that encourage or discourage the use of debt. For instance, it is not only flota-
tion costs, information asymmetry or concern about undervaluation of new
shares that may make firms reluctant to issue external equity, concern about

ß Blackwell Publishers Ltd 1998

1130 ADEDEJI

control and earnings per share may also bias firms against external equity.
There are also a number of good reasons why firms may not follow the peck-
ing order suggested by the hypothesis. For instance, Viswanath (1993) shows
that long term strategic considerations may make a firm prefer external fi-
nance to internally generated funds in the short term in a multiperiod context.
Myers' (1977) suggestion that firms should not use debt to finance growth op-
portunities if they want to be in a position to accept all positive-NPV projects
and maximise their market values is also relevant. The suggestion implies that
firms which need external finance to create growth opportunities will not fol-
low the pecking order described earlier. Such firms will prefer external equity
to debt in order to avoid the underinvestment problem that may be created by
the use of debt to finance growth opportunities. Similarly, DeAngelo and Ma-
sulis' (1980) proposition that there is an inverse relationship between financial
leverage and the amount of tax-shield substitutes for debt that a firm has1 im-
plies that firms with little or no tax-shield substitutes for debt may prefer to use
debt before internal equity in order to obtain the tax advantage of debt and
optimise their cost of capital and market values.
Therefore, the degree of encouragement given to firms to use debt by var-
ious institutional arrangements is an important factor because it can change
the pecking order and affect the empirical relationship between investment
and financial leverage. Some of the relevant institutional factors that can en-
courage firms to use debt before internal equity are the level of interest rate,
the relationship between creditors (or banks) and firms and the likelihood of
Government intervention if there is a financial crisis (Madura, 1995, p. 549).
The tax system is a particularly important institutional factor that may en-
courage firms to use more or less debt before internal equity. Ang and Jung
(1993) find that firms in South Korea prefer to use money from banks before
internally generated funds (Ang and Jung, 1993, Table 3, p. 40). Although no
specific explanation is given for this order of preference, it can be argued that
there is nothing wrong with it if the firms observed have earnings that can be
shielded from taxation by using debt.
Since the imputation tax system in the UK does not encourage firms to use
debt as much as the classicial tax system does in the US, the relationship be-
tween investment and financial leverage in the UK may be different from that
reported for the US. Evidence reported by Madura (1995, p. 557) indicates
that the debt ratios of firms in the UK are less than those of their counterparts
in the US. It is possible that the difference between the tax systems in the US
and the UK is one of the factors responsible for this. If it is, then the imputation
tax system may weaken the interaction between financial leverage and invest-
ment, and strengthen the interaction between dividend payout ratio and in-
vestment, in the UK. Whether this is the case or not is an empirical issue
which this study tries to shed some light on.

ß Blackwell Publishers Ltd 1998

 DIVIDEND PAYOUT RATIOS OF UK FIRMS 1131

TEST METHODOLOGY

If dividend payout ratio, financial leverage and investment are interrelated,

then, dividend payout ratio should be a function of financial leverage and in-
vestment, after controlling for other factors. Similarly, financial leverage
should be a function of dividend payout ratio and investment, and investment
should be a function of dividend payout ratio and financial leverage, after con-
trolling for other factors. In order to determine whether this is the case, the
following system of equations is formulated:
DIV = f (FL, INV, OTHERDIV) (1)
FL = f (DIV, INV, OTHERFL) (2)
INV = f (DIV, FL, OTHERINV) (3)
where DIV, FL and INV represent dividend payout ratio, financial leverage
and investment respectively. OTHERDIV, OTHERFL and OTHERINV
are vectors of other variables that may have an influence on dividend payout
ratio, financial leverage and investment respectively. The variables that are
included in each vector are chosen on the basis of the results of previous stu-
dies.
The variables in the vector OTHERDIV are irrecoverable Advance Cor-
poration Tax i.e. ACT (IRRTX), industry average dividend yield (INDDY),
liquidity (LIQ), overseas profit (OVPRO), profitability (PRO), size (SZ),
variability of earnings (VAR) and industry factor (IND). The variable
IRRTX is included in the vector because of the view often expressed that the
imputation tax system in the UK encourages firms to pay dividends. An exam-
ple of such views is the recent statement by the Chancellor of the Exchequer
that:
The present system of tax credits encourages companies to pay out dividends rather
than reinvest their profits. . . . This cannot be the best way of encouraging invest-
ment for the long term. . . .(Financial Times, 3 July, 1997, p. 6).
The second sentence in the quotation above is correct only if there is a nega-
tive interaction between dividend payout ratio and investment in the UK. But
there was no direct evidence that could be used to assess its empirical validity
at the time it was made, or at any time before the present study. Certainly, the
sentence will not be correct if there is no significant interaction between divi-
dend payout ratio and investment, or if, as Miller and Modigliani (1961) sug-
gest, dividends and investment are separated by firms,2 or if, as Miller and
Rock (1985) point out, dividends are used as a signal of expected earnings.
Jensen (1986) suggests that it is better for firms to pay more, rather than less,
dividends if they do not have profitable growth opportunities. The suggestion
implies that firms may not reduce dividends if they have free cash flows, de-
spite Government's effort to encourage reduction in dividends. Although the
aspect of the system which enabled advance corporation tax to be refunded to
pension funds and some other tax-exempt institutions has been abolished, it

ß Blackwell Publishers Ltd 1998

1132 ADEDEJI

was there during the period covered by this study.3 This is why variable
IRRTX is in the vector. The variable should have a positive influence on di-
vidend payout ratio (DIV) if firms see irrecoverable ACT as a tax allowance
that can enhance their distributable earnings. Conversely, the variable should
have a negative influence on dividend payout ratio, if firms pay dividends
mainly to create opportunities for tax credits or tax refunds for their share-
holders.
The purpose of the variable representing prevous year's industry average divi-
dend yield (INDDY) is to test whether firms target their industry average divi-
dend yield when setting their dividend payout ratios. If they do, the variable will
have a positive influence on dividend payout ratio. The variable is included in the
vector because a number of previous studies observe that dividend yield is a re-
turn determinant (e.g. Litzenberger and Ramaswamy,1982; Levis, 1989; and
Chui, Strong and Cadle, 1992). This observation provides a good reason for firms
to target their industry average dividend yield.
The liquidity variable (LIQ) is included in the vector because Gill and
Green (1993) suggest that liquidity has a positive influence on dividend pay-
out in the UK. The overseas profit variable (OVPRO) is in the vector because
the imputation tax system in the UK at the time of this study does not grant
tax credit, or allow tax refund, for dividends paid from overseas profit (Lasfer,
1996). The variable is expected to have a negative sign because of this reason
and the fact that overseas profit is susceptible to additional risks, such as for-
eign exchange risk and political risk. However, the variable representing prof-
itability (PRO) is in the vector because Theobald (1978) finds that current
earnings have a positive influence on dividends in the UK. Size (SZ) is in the
vector because Chang and Rhee (1990) observe that the variable has a posi-
tive influence on dividend payout ratio in the US. Little research evidence is
available on the influence of size on dividend payout ratio in the UK before
this study.
Variability of earnings (VAR) is another variable in the vector OTHER-
DIV. The variable is included in the vector because the existing theory sug-
gests that it is a negative determinant of dividend payout ratio (Lintner,
1956; and Weston and Copeland, 1992, p.660). The dummy variables (IND)
are included in the vector to control for industry-specific factors that may in-
fluence dividend decisions.
Evidence reported by Bennett and Donnelly (1993) indicates that deferred
tax (DEFTX), profitability (PRO), size (SZ), structure of assets (STR),
variability of earnings (VAR) and industry dummy variables (IND) have an
influence on financial leverage (FL). That is why the variables are included in
the vector OTHERFL. In accordance with the prediction of pecking order
hypothesis and the previous evidence, defered tax (DEFTX) is expected to
have a negative sign because it represents non-debt tax shield (Bennett and
Donnelly, 1993). Industry average debt ratio (INDFL) is expected to have a
positive sign because firms which have below average debt ratios are likely to

ß Blackwell Publishers Ltd 1998

 DIVIDEND PAYOUT RATIOS OF UK FIRMS 1133

find it relatively easier to raise more debt. Profitability (PRO) is expected to

have a negative sign, given the suggestion of pecking order hypothesis and the
previous evidence.
Size (SZ) is expected to have a positive influence on financial leverage (FL)
because large firms have better access to debt capital market (Marsh, 1982;
Baskin, 1989; Chang and Rhee, 1990; and Bennett and Donnelly, 1993). As-
sets structure (STR) is expected to have a positive influence on financial lever-
age (FL) because firms which have assets with good collateral value (such as
tangible assets) will be able to raise more debt (Scott, 1977; Myers, 1977; and
Marsh, 1982). While the existing theory predicts that earnings variation
(VAR) has a negative influence on financial leverage (Copeland and Weston,
1992, p.580), previous evidence indicates that the influence of the variable is
positive in the UK (Bennett and Donnelly, 1993, Table 4, p. 54 and Table 5, p.
56). Therefore, the variable is expected to have a positive influence.
The research and development variable (RD) is used as a proxy for agency
problems associated with debt financing (Myers, 1977). The variable is ex-
pected to have a negative sign. Again, the effect of an industry dummy vari-
able (IND) could be positive or negative, depending on the nature of the
influence of the industry on financial leverage (FL).
Finally, the variables in the vector OTHERINV are sales growth (GRO),
profitability (PRO), q ratio (q), size (SZ) and industry dummy variables
(IND). Four of these variables, namely, profitability (PRO), q ratio (q), size
(SZ) and industry factor (IND), are among the five control variables in the
investment equation estimated by Baskin (1989). There are only two control
variables in the investment equation estimated by Allen (1993), namely, prof-
itability (PRO) and size (SZ). Both Baskin and Allen find that size (SZ) has a
significant negative effect on investment, but their results differ on the signifi-
cance of profitability (PRO). While Baskin finds that profitability is signifi-
cant in the US, Allen finds that it is not significant in Australia. Baskin also
finds that q ratio (q) has a positive influence, and that industry factor (IND)
has a generally significant influence, on investment in the US.
The fifth control variable used by Baskin is called `specialisation ratio' and
measured by the percentage of a firm's sales in its largest 4-digit SIC industry.
The variable was used to `control for diversifying growth which may have
been accompanied by new equity issues' (Baskin, 1989, p. 32). The variable
is not used in this study because of data problems. Instead, it is substituted with
annual sales growth (GRO) which is used to control the effect of sales growth
on investment. Fama (1974) observes that sales growth (GRO) has a positive
influence on investment in the US. Therefore, the variable is expected to have
a positive sign.
The following model is obtained after substituting the variables referred to
above for the vectors OTHERDIV, OTHERFL and OTHERINV in equa-
tions (1) ÿ (3) and restating each of the equations in a standard regression
form:

ß Blackwell Publishers Ltd 1998

1134 ADEDEJI

DIV ˆ 0 ‡ 1 FL ‡ 2 INDDY ‡ 3 INV ‡ 4 IRRTX ‡ 5 LIQ ‡

X
9
6 OVPRO ‡ 7 PRO ‡ 8 SZ ‡ 9 VAR ‡ i‡9 INDi ‡ " …4†
iˆ1

FL ˆ 0 ‡ 1 DEFTX ‡ 2 DIV ‡ 3 INDFL ‡ 4 INV ‡ 5 PRO‡

X
9
6 RD ‡ 7 STR ‡ 8 SZ ‡ 9 VAR ‡ i‡9 INDi ‡ u …5†
iˆ1

INV ˆ 0 ‡ 1 DIV ‡ 2 FL ‡ 3 GRO ‡ 4 PRO ‡ 5 q ‡ 6 SZ

X
9
‡ i‡6 INDi ‡ e …6†
iˆ1

where:

DEFTX = proxy for tax, represented by (potential deferred tax liability

(Datastream item 308)/total assets);
DIV = dividend payout ratio, measured by dividends/distributable
earnings;
FL = financial leverage, measured by total debt/market value of the
firm.
Total debt is defined as long term debt + current liabilities.
Market value of the firm is defined as the sum of total debt and
market value of equity;
GRO = sales growth, measured by (salest ÿ salestÿ1)/salestÿ1;
INDDY = industry average dividend yield in the previous year;
INDFL = industry average total debt ratio in the previous year;
INV = investment, measured by (total assets in 1996 ÿ total assets in
1993)/total assets in 1993;
IRRTX = irrecoverable advance corporation tax, measured by the ratio
of irrecoverable advance corporation tax (Datastream item
164) to total assets;
LIQ = liquidity, measured by (current assets ÿ stocks)/current liabil-
ities;
OVPRO = overseas profit, represented by a dummy variable which is as-
signed a value of 1 if a firm has overseas sales and positive dis-
tributable earnings (Datastream item 1087). Otherwise, the
variable is assigned a value of zero. Data of overseas sales are
obtained from Datastream by creating an expression for the
variable;
PRO = profitability, measured by profit before interest and tax/total
assets;
q = q ratio, used as a proxy for expected growth. It is represented

ß Blackwell Publishers Ltd 1998

 DIVIDEND PAYOUT RATIOS OF UK FIRMS 1135

by price-to-book-value ratio;
RD = another proxy for expected growth. It is represented by R&D
expenditures/sales;
STR = structure of assets, measured by total net fixed assets (Data-
stream item 339)/Market value of the firm;
SZ = size, represented by the natural logarithm of total assets in De-
cember 1993, measured in » millions;
VAR = variability of earnings, represented by the standard deviation
of (annual change in profit before interest, tax and deprecia-
tion/total assets), calculated over 1986^96;
IND = industry dummy variables;
0 , . . . , 18 , 0 , . . . , 18 and 0 , . . . , 15 are regression parameters, and ", u
and e are error terms.

Given the predictions of pecking order hypothesis and the results of previous
studies, 1 , 2 , 5 , 7 , 8 in equation (4), 3 , 4 , 7 , 8 , 9 in equation (5), 2 ,
3 , 4 and 5 in equation (6) are expected to be positive. 3 , 6 , 9 in equation
(4), 1 , 5 , 6 in equation (5), 1 and 6 in equation (6) are expected to be
negative. 4 in equation (4) could be positive or negative, depending on how
ACT influences corporate dividend decision. Similarly, given the contradic-
tions in the relationship between dividend payout ratio and financial leverage
in the previous evidence, 2 in equation (5) could be positive or negative. Also,
the coefficients of the industry dummy variables (IND) could be positive or
negative, depending on the nature of the influence of the industries that they
represent.
Theobald (1978) and McCabe (1979) argue that the cross-sectional test
method is more appropriate than the intertemporal test method for assessing
long term relationships among variables because the intertemporal test meth-
od reveals only short term relationships. In accordance with this argument,
this study uses the cross-sectional test method. In the test, each of the exogen-
ous variables in equations (4)^(6) is represented by its four year average value,
calculated over 1993^96, except variability of earnings (VAR), which is cal-
culated over 1986^96. The main justification for using average values is that
some previous studies argue that average values are better than a single point
estimate for testing theories which relate to long term behaviour of firms if one
wants to avoid distortions that may be caused by short term variations from
the target (e.g. Titman and Wessels, 1988, for the US, and Bennett and Don-
nelly, 1993, for the UK).
However, the endogenous variables DIV and FL are represented by their
values in 1993 (DIV93 and FL93), 1994 (DIV94 and FL94), 1995 (DIV95
and FL95) and their average values for the four years 1993^96 (AVDIV and
AVFL). The investment variable is represented by a single value calculated
over 1993^96. The use of annual values of dividend payout ratio and financial
leverage side-by-side with the value of investment calculated over a four year

ß Blackwell Publishers Ltd 1998

1136 ADEDEJI

period is consistent with the approach used by Baskin and Allen.

The purpose of the test is to assess the validity of two suggestions of the peck-
ing order hypothesis. One of the suggestions is that large dividend payout in
one period will reduce the funds available for investment in subsequent peri-
od(s) and increase the tendency to raise debt in the period(s). The other is that
large investment outlay over a period of time will impose a downward pressure
on dividend payout and increase the need for external debt financing during
the period. In order to test these suggestions, DIV96 is regressed on FL93,
INV and the control variables in equation (4). Similarly, FL96 is regressed
on DIV93, INV and the control variables in equation (5) and INV is regressed
on DIV93, FL93 and the control variables in equation (6).
If the suggestions are supported by the behaviour of the firms observed,
then, the investment variable (INV) should have a negative influence on divi-
dend payout ratios in 1996 (DIV96) in equation (4), and dividend payout ra-
tios in 1993 (DIV93) should have a negative influence on investment (INV) in
equation (6). Also, if the behaviour of the firms is consistent with the previous
evidence, dividend payout ratios (DIV93) and investment (INV) should have
a positive impact on financial leverage (FL96) in equation (5).
After using DIV93 and FL93 to obtain one estimate of equations (4)-(6),
the variables are replaced with DIV94 and FL94 and the equations are re-es-
timated. Subsequently, DIV95 and FL95 are substituted for DIV94 and
FL94. Finally, the four year average values of DIV and FL (i.e. AVDIV and
AVFL) are used to estimate the equations, giving four estimates altogether.
This is done to minimise the effect of multicollinearity on the test of the influ-
ence of previous years' values of dividend payout ratios, financial leverage and
investment on DIV96 and FL96.

SAMPLE

The data used are obtained from Datastream. Firms are selected for the study
if they have the relevant data for estimating variability of earnings (VAR) in
Datastream for 11 years up to 1996, and they have the relevant data for the
other variables from 1993^96. Application of this criterion resulted in a sam-
ple of 224 firms. Firms in the sample are classified into industries using the
London Stock Exchange industry classification codes in Datastream. This
procedure results in identification of nine industries having at least nine firms
each, and several other industries with fewer than nine firms each. If every in-
dustry in the sample is represented by a dummy variable, estimates of the coef-
ficients of the dummy variables representing industries with few firms will be
unreliable. Therefore, all industries having less than nine firms each are com-
bined into one large group which is called the control group. Each of the other
nine industries is represented by a dummy variable. That is why there are nine
dummy variables in each equation in the model.4 The industries in the sample

ß Blackwell Publishers Ltd 1998

 DIVIDEND PAYOUT RATIOS OF UK FIRMS 1137

and the number of firms in them are building (34), chemicals (14), distributors
(16), electronics and electricals (22), engineering (50), food producers (13),
general retailers (10), leisure and hotels (9), media (13) and a number of other
industries that are used collectively as the control group (43).

ANALYSES AND RESULTS

The summary statistics of the sample are in Table 1. The table indicates that,
on average, about 60% of the distributable earnings of the firms in the sample
are paid out as dividends, and about 42% of their total assets are financed by
debt. The average value of the investment variable (INV) is 1.659. This im-
plies that, on average, the firms grew by about 166% over the study period.5
The average value of the size variable (SZ) is 3.716. This converts to about
»41 million. The average rate of profitability of the firms (PRO) is 13% and
their average annual sales growth rate (GRO) is 9.5%.
The correlations of the variables in Table 1 are summarised in Tables 2a
and 2b. The correlations of the exogenous variables are generally low, but, as
expected, the endogenous variables DIV96 and FL96 have high correlation
with their previous values and the industry average. The financial leverage
variables (FL93, FL94, FL95 and FL96) also have high correlation with the
liquidity variable (LIQ). This is not surprising, given that both variables have
current liabilities in them.
The interaction of dividend payout ratio, financial leverage and investment
is tested by using the three stage least squares regression technique, hereafter
3SLS, in LIMDEP (version 7.0) to obtain the four estimates of equations
(4) ÿ (6) described above. Greene (1993) suggests that 3SLS is more appro-
priate than the OLS regression technique for examining interactions among
endogenous variables in simultaenous equations (see Greene, 1993, p. 615^
6). However, since Baskin and Allen use only OLS regression technique, the
technique is also used here to obtain the four estimates of equations (4)^(6)
referred to earlier in order to have some results that can be compared with
the previous evidence without reference to 3SLS. Another reason for using
OLS technique is that the results that are obtained by it is useful for assessing
the extent to which the results obtained by using 3SLS is influenced by 3SLS.
Details of the 3SLS and OLS results obtained are reported in Appendices
A^H. The summary of the results is in Table 3. The summary indicates that
in the 3SLS results, dividend payout ratio and investment have a generally
negative influence on each other.6 This indication supports the prediction that
the two variables are negatively interrelated. The summary also indicates that
there is a generally positive interrelationship between dividend payout ratio
and financial leverage but that there is no significant interaction between fi-
nancial leverage and investment. The investment variable (INV) has a posi-
tive and significant influence on the financial leverage variables (FL96 and

ß Blackwell Publishers Ltd 1998

1138 ADEDEJI

Table 1
Sample Summary Statistics

Variables Mean Std. Dev. Minimum Maximum

DEFTX 0.024 0.050 ÿ0.262 0.307

DIV93 0.625 0.306 0.000 1.000
DIV94 0.577 0.281 0.000 1.000
DIV95 0.593 0.284 0.000 1.000
DIV96 0.568 0.291 0.000 1.000
AVDIV 0.591 0.214 0.000 1.000
FL93 0.407 0.210 0.059 0.989
FL94 0.419 0.203 0.044 0.995
FL95 0.430 0.209 0.061 0.997
FL96 0.411 0.195 0.050 0.994
AVFL 0.417 0.188 0.065 0.994
GRO 0.095 0.154 ÿ0.258 1.007
INDDY 0.595 0.057 0.000 0.693
INDFL 0.417 0.070 0.000 0.520
INV 1.659 3.823 ÿ0.597 42.961
IRRTX 0.001 0.005 ÿ0.049 0.038
LIQ 0.991 0.523 0.129 4.650
OVPRO 0.652 0.478 0.000 1.000
PRO 0.130 0.057 ÿ0.071 0.284
q 12.743 33.499 0.228 127.193
RD 0.008 0.018 0.000 0.126
STR 0.258 0.159 0.005 0.988
SZ 3.716 1.888 0.229 10.237
VAR 0.067 0.080 0.010 0.670
Notes:
The estimates reported here are obtained by using the summary statistics procedure in LIMDEP
(version 7.0). DEFTX = proxy for tax, represented by (potential deferred tax liability (Datastream
item 308)/total assets); DIV = dividend payout ratio, measured by dividends/distributable earnings;
DIV9x = dividend payout ratio in 199x; AVDIV = average dividend payout ratio over 1993^96;
FL = financial leverage, measured by total debt/market value of the firm. Total debt is defined as long
term debt + current liabilities. Market value of the firm is defined as the sum of total debt and market
value of equity; FL9x = financial leverage value in 199x; AVFL = average financial leverage over
1993^96; GRO = sales growth, measured by (salestÿsalestÿ1)/salestÿ1; INDDY = industry average
dividend yield for the previous year; INDFL = industry average total debt ratio for the previous year;
INV = investment, measured by (total assets in 1996ÿtotal assets in 1993)/total assets in 1993;
IRRTX = irrecoverable advance corporation tax, measured by the ratio of irrecoverable advance
corporation tax (Datastream item 164) to total assets; LIQ = liquidity, measured by (current
assets ÿ stocks)/current liabilities; OVPRO = overseas profit, represented by a dummy variable
which is assigned a value of 1 if a firm has overseas sales and its distributable earnings (Datastream
item 1087) are positive. Data of overseas sales are obtained from Datastream by creating an expres-
sion for the variable; PRO = profitability, measured by profit before interest and tax/total assets; q = a
proxy for expected growth. It is represented by price-to-book-value ratio; RD = another proxy for
expected growth. It is represented by R&D expenditures/sales; STR = structure of assets, measured
by total net fixed assets (Datastream item 339)/Market value of the firm; SZ = size, represented by the
natural logarithm of total assets in December 1993, measured in » millions; VAR = variability of
earnings, represented by the standard deviation of (annual change in profit before interest, tax and
depreciation/total assets), calculated over 1986^96. Each variable (except DIV93^DIV96, FL93^
FL96, INV and VAR) is represented by its four year average calculated over 1993^96.

ß Blackwell Publishers Ltd 1998

ß Blackwell Publishers Ltd 1998

Table 2a
Correlations

Variables DEFTX DIV93 DIV94 DIV95 DIV96 AVDIV FL93 FL94 FL95 FL96 AVFL GRO INDDY INDFL

DIV93 ÿ0.201*

DIVIDEND PAYOUT RATIOS OF UK FIRMS

DIV94 ÿ0.067 0.502*
DIV95 0.012 0.343* 0.442*
DIV96 ÿ0.139 0.340* 0.347* 0.381*
AVDIV ÿ0.137 0.751* 0.771* 0.728* 0.701*
FL93 ÿ0.305* 0.295* 0.225* 0.193* 0.084 0.272*
FL94 ÿ0.237* 0.218* 0.269* 0.232* 0.110 0.280* 0.880*
FL95 ÿ0.271* 0.212* 0.230* 0.318* 0.1580 0.310* 0.786* 0.862*
FL96 ÿ0.215* 0.175* 0.166 0.247* 0.247* 0.282* 0.674* 0.743* 0.865*
AVFL ÿ0.279* 0.245* 0.242* 0.268* 0.161 0.310* 0.908* 0.945* 0.952* 0.886*
GRO 0.062 ÿ0.169 ÿ0.140 ÿ0.265* ÿ0.161 ÿ0.249* ÿ0.275* ÿ0.255* ÿ0.111 ÿ0.081 ÿ0.197*
INDDY 0.049 0.131 0.176* 0.115 0.101 0.177* ÿ0.005 ÿ0.024 ÿ0.045 ÿ0.075 ÿ0.040 ÿ0.091
INDFL ÿ0.068 0.008 0.053 ÿ0.095 0.045 0.004 0.283* 0.332* 0.311* 0.300* 0.332* 0.088 ÿ0.079
INV 0.055 0.143 0.037 0.079 ÿ0.056 0.071 0.202* 0.195* 0.135 0.116 0.176* ÿ0.114 ÿ0.029 ÿ0.001
IRRTX ÿ0.202* 0.086 0.199* 0.055 0.104 0.149 0.107 0.138 0.159 0.178* 0.157 ÿ0.015 ÿ0.016 0.077
LIQ 0.030 ÿ0.075 ÿ0.119 ÿ0.079 ÿ0.082 ÿ0.120 ÿ0.368* ÿ0.432* ÿ0.456* ÿ0.455* ÿ0.463* 0.0141 ÿ0.063 ÿ0.298*
OVPRO 0.128 ÿ0.079 ÿ0.230* ÿ0.182* ÿ0.304* ÿ0.267* ÿ0.190* ÿ0.255* ÿ0.317* ÿ0.296* ÿ0.286* 0.159 0.077 ÿ0.221*
PRO 0.111 ÿ0.063 ÿ0.102 ÿ0.062 ÿ0.028 ÿ0.086 ÿ0.353* ÿ0.320* ÿ0.284* ÿ0.221* ÿ0.320* 0.049 0.030 ÿ0.142
q ÿ0.046 ÿ0.010 0.085 0.047 ÿ0.074 0.015 0.230* 0.219* 0.307* 0.242* 0.271* 0.106 ÿ0.183* ÿ0.006
RD ÿ0.054 ÿ0.253* ÿ0.134 0.009 ÿ0.059 ÿ0.151 ÿ0.228* ÿ0.231* ÿ0.203* ÿ0.217* ÿ0.238* 0.038 ÿ0.289* ÿ0.365*
STR 0.189* 0.113 0.020 0.062 0.136 0.114 0.190* 0.194* 0.145 0.173* 0.190* ÿ0.234* 0.062 0.040
SZ 0.116 0.141 0.195* 0.209* 0.175* 0.243* ÿ0.094 ÿ0.063 ÿ0.103 ÿ0.083 ÿ0.093 ÿ0.099 0.204* ÿ0.060
VAR ÿ0.099 0.001 ÿ0.058 ÿ0.014 0.044 ÿ0.009 0.057 0.026 0.017 0.017 0.032 ÿ0.040 ÿ0.027 0.079

1139
 1140
Table 2b
Correlations

Variables INV IRRTX LIQ OVPRO PRO q RD STR SZ

IRRTX ÿ0.363*
LIQ ÿ0.130 ÿ0.001
OVPRO ÿ0.020 ÿ0.111 0.238*
PRO ÿ0.030 0.007 0.101 0.124
q 0.348* ÿ0.108 ÿ0.146 ÿ0.022 ÿ0.040
RD ÿ0.073 ÿ0.009 0.200* 0.039 ÿ0.013 0.011
STR ÿ0.069 0.042 ÿ0.099 ÿ0.154 ÿ0.142 ÿ0.112 ÿ0.188*
SZ ÿ0.169 0.079 0.017 0.125 0.077 0.050 ÿ0.037 0.024

ADEDEJI
VAR ÿ0.009 ÿ0.050 0.089 ÿ0.030 0.009 ÿ0.052 0.030 ÿ0.160 ÿ0.064

Notes:
* Indicates values that are significant at the 5% or higher levels. The estimates reported in Table 2(a) and this table are obtained by using the Spearman's
moment correlation procedure in SPSS. DEFTX = proxy for tax, represented by (potential deferred tax liability (Datastream item 308)/total assets); DIV = -
dividend payout ratio, measured by dividends/distributable earnings; DIV9x = dividend payout ratio in 199x; AVDIV = average dividend payout ratio over
1993^96; FL = financial leverage, measured by total debt/market value of the firm. Total debt is defined as long term debt + current liabilities. Market value of
the firm is defined as the sum of total debt and market value of equity; FL9x = financial leverage value in 199x; AVFL = average financial leverage over 1993^96;
GRO = sales growth, measured by (salest ÿ salestÿ1)/salestÿ1; INDDY = industry average dividend yield for the previous year; INDFL = industry average total
ß Blackwell Publishers Ltd 1998

debt ratio for the previous year; INV = investment, measured by (total assets in 1996ÿtotal assets in 1993)/total assets in 1993; IRRTX = irrecoverable advance
corporation tax, measured by the ratio of irrecoverable advance corporation tax (Datastream item 164) to total assets; LIQ = liquidity, measured by (current
assets ÿ stocks)/current liabilities; OVPRO = overseas profit, represented by a dummy variable which is assigned a value of 1 if a firm has overseas sales and its
distributable earnings (Datastream item 1087) are positive. Data of overseas sales are obtained from Datastream by creating an expression for the variable;
PRO = profitability, measured by profit before interest and tax/total assets; q = a proxy for expected growth. It is represented by price-to-book-value ratio;
RD = another proxy for expected growth. It is represented by R&D expenditures/sales; STR = structure of assets, measured by total net fixed assets (Data-
stream item 339)/Market value of the firm; SZ = size, represented by the natural logarithm of total assets in December 1993, measured in » millions; VAR = -
variability of earnings, represented by the standard deviation of (annual change in profit before interest, tax and depreciation/total assets), calculated over
1986^96. Each variable (except DIV93 ÿ DIV96, FL93 ÿ FL96, INV and VAR) is represented by its four year average calculated over 1993^96.
 DIVIDEND PAYOUT RATIOS OF UK FIRMS 1141

Table 3
Summary of the Regression Results
Dependent Variables in the Dependent Variables in the
3SLS Regression OLS Regression
Independent DIV96/ FL96/ DIV96/ FL96/ Location of
Variables AVDIV AVFL INV AVDIV AVFL INV the Results
DIV93 Zero Zero Zero Positive Apps. A and E
FL93 Positive Zero Zero Zero ``
INV Negative Positive Zero Zero ``
DIV94 Positive Negative Positive Zero Apps. B and F
FL94 Positive Zero Zero Zero ``
INV Negative Positive Zero Zero ``
DIV95 Positive Zero Positive Positive Apps. E and G
FL95 Positive Zero Zero Zero ``
INV Negative Positive Zero Zero ``
AVDIV Positive Negative Positive Zero Apps. F and H
AVFL Positive Zero Positive Zero ``
INV Zero Positive Positive Zero ``

Control Variables:
DEFTX Zero Negative Apps. A-H
GRO Negative Zero ``
INDDY Zero Positive ``
INDFL Positive Positive ``
IRRTX Zero Positive ``
LIQ Zero Zero ``
OVPRO Negative Negative ``
PRO Zero Zero Zero Zero Negative Positive ``
q Positive Positive ``
RD Positive Negative ``
STR Positive Positive ``
SZ Positive Zero Negative Positive Zero Negative ``
VAR Zero Zero Zero Zero ``
Notes:
DEFTX = proxy for tax, represented by (potential deferred tax liability (Datastream item 308)/total assets);
DIV = dividend payout ratio, measured by dividends/distributable earnings; DIV9x = dividend payout ratio in
199x; AVDIV = average dividend payout ratio over 1993^96; FL = financial leverage, measured by total debt/
market value of the firm. Total debt is defined as long term debt + current liabilities. Market value of the firm is
defined as the sum of total debt and market value of equity; FL9x = financial leverage value in 199x; AVFL = aver-
age financial leverage over 1993^96; GRO = sales growth, measured by (salestÿsalestÿ1)/salestÿ1; INDDY = in-
dustry average dividend yield for the previous year; INDFL = industry average total debt ratio for the previous
year; INV = investment, measured by (total assets in 1996ÿtotal assets in 1993)/total assets in 1993; IRRTX = ir-
recoverable advance corporation tax, measured by the ratio of irrecoverable advance corporation tax (Data-
stream item 164) to total assets; LIQ = liquidity, measured by (current assetsÿstocks)/current liabilities;
OVPRO = overseas profit, represented by a dummy variable which is assigned a value of 1 if a firm has overseas
sales and its distributable earnings (Datastream item 1087) are positive. Data of overseas sales are obtained from
Datastream by creating an expression for the variable; PRO = profitability, measured by profit before interest and
tax/total assets; q = a proxy for expected growth. It is represented by price-to-book-value ratio; RD = another
proxy for expected growth. It is represented by R&D expenditures/sales; STR = structure of assets, measured by
total net fixed assets (Datastream item 339)/Market value of the firm; SZ = size, represented by the natural loga-
rithm of total assets in December 1993, measured in » millions; VAR = variability of earnings, represented by the
standard deviation of (annual change in profit before interest, tax and depreciation/total assets), calculated over
1986^96. Each variable (except DIV93ÿDIV96, FL93ÿFL96, INV and VAR) is represented by its four year
average calculated over 1993^96.

ß Blackwell Publishers Ltd 1998

1142 ADEDEJI

AVFL) in all the estimates, but the financial leverage variables (FL93^FL95
and AVFL) do not have a significant influence on the investment variable
(INV) in any of the estimates. The OLS regression results differ from the
3SLS results in two respects, namely, they indicate that (1) there is generally
no significant interaction between the dividend payout ratios and the invest-
ment variable (INV) and (2) the investment variable (INV) does not have a
significant influence on the financial leverage variables (FL96 and AVFL).
As for the control variables, the 3SLS results indicate that size (SZ) has a
positive influence, and overseas profit (OVPRO) has a negative influence, on
dividend payout ratios (DIV96 and AVDIV). These are as expected. The re-
sults also indicate that previous year's industry average dividend yield (IN-
DDY), irrecoverable advance corporation tax (IRRTX), liquidity7 (LIQ),
profitability (PRO) and variability of earnings (VAR) do not have a signifi-
cant influence on the dividend payout ratios. The influence of previous year's
industry average debt ratios (INDFL) and structure of assets (STR) on finan-
cial leverage (FL96 and AVFL) is positive, as expected, and that of R&D ex-
penditures is also positive, contrary to the prediction that it is negative. In
general, deferred tax (DEFTX), profitability (PRO), size (SZ) and variabil-
ity of earnings (VAR) do not have a significant influence on financial leverage
in the results. Sales growth (GRO) has unexpected negative influence, but q
ratio (q) and size (SZ) have the expected influence, on the investment variable
(INV). The influence of profitability on investment is not significant.
The indications of the OLS results are similar to those of the 3SLS, except
that previous year's industry average dividend yield (INDDY) and irrecover-
able advance corporation tax (IRRTX) have the expected positive influence
on dividend payout ratios (DIV96 and AVDIV). Also, deferred tax
(DEFTX), profitability (PRO) and R&D expenditures (RD) have the ex-
pected negative influence on financial leverage (FL96 and AVFL), and profit-
ability (PRO) has the expected positive influence on investment (INV), in the
OLS results.
Industry classification (IND) has more influence on the investment vari-
able (INV) than on dividend payout ratios (DIV96 and AVDIV) or financial
leverage (FL96 and AVFL) in the 3SLS results. However, in the OLS results,
industry classification has greater influence on financial leverage (FL96 and
AVFL) than on dividend payout ratios (DIV96 and AVDIV) or investment
(INV).
The likelihood ratio technique is used to test the model's specification in the
way suggested by Greene (1993, pp. 616^8). The result of the test is not signif-
icant. This indicates that the model described in equations (4)^(6) is specified
correctly. The results of the F-ratio tests conducted for the OLS regressions
support this indication.8
The results referred to above are compared with the previous evidence re-
ported by Baskin and Allen which are the only two rigorous direct tests of
pecking order hypothesis that have been reported before this study. The result

ß Blackwell Publishers Ltd 1998

ß Blackwell Publishers Ltd 1998

Table 4
Comparison of the Results with the Previous Evidence

DIVIDEND PAYOUT RATIOS OF UK FIRMS

Effect on DIV Effect on FL Effect on INV
Variables Baskin Allen Adedeji Baskin Allen Adedeji Baskin Allen Adedeji
(1989) (1993) (1998) (1989) (1993) (1998) (1989) (1993) (1998)

Positive Positive
DIV ÿ ÿ na Positive Negative Positive Negative Negative Negative

Positive Zero
FL ÿ ÿ Positive Positive ÿ na Positive Positive Zero

Zero Zero
INV ÿ ÿ Negative Positive Zero Positive na na na

Zero Negative Positive

PROFIT ÿ ÿ Zero Negative Negative Zero Positive Zero Zero

Notes:
Indications of the OLS results are in italics, those of the 3SLS are in bold characters. na = not applicable.

1143
1144 ADEDEJI

of the comparison, summarised in Table 4, indicates that the main difference

between their evidence and the results of this study is that financial leverage
does not have a significant influence on investment. The rest of the results are
generally in agreement with the previous evidence as far as the effects of divi-
dend payout ratio on financial leverage and investment, the effect of invest-
ment on financial leverage, and the effects of profitability on financial
leverage and investment, are concerned.

CONCLUSIONS

The study concludes that there is a negative interaction between dividend

payout ratio and investment in the UK over the period 1993^96. It also con-
cludes that there is a positive interaction between dividend payout ratio and
financial leverage over the same period. However, there is no significant inter-
action between investment and financial leverage. Although investment has a
positive influence on financial leverage, financial leverage does not have a sig-
nificant influence on investment.
Another conclusion of the study is that irrecoverable ACT has a positive,
but weak, influence on dividend payout ratio and that overseas profit has a
negative influence on the ratio. It is also observed that liquidity and variabil-
ity of earnings do not have a significant influence on dividend payout ratio,
but previous year's industry average dividend yield and size have a positive
influence. There are also indications that previous year's industry average
debt ratio, profitability, R&D expenditures and structure of assets have a sig-
nificant influence on financial leverage, and that q ratio and size have a signif-
icant influence on investment.
The only disparity between the results of this study and the previous evi-
dence is that financial leverage does not have a significant influence on invest-
ment. The rest of the results are similar to the previous evidence. However,
this study adds to the previous evidence by using simultaneous equations
method to carry out a comprehensive test of the interaction between dividend
payout ratio, financial leverage and investment and by showing that dividend
payout ratio has a positive interaction with financial leverage and a negative
interaction with investment.

NOTES
1 Two examples of tax shield substitutes for debt are R&D expenditures and capital allowances.
2 Although a number of previous studies have examined the empirical relationship between
dividends and investment, their results are mixed and inconclusive (Dhrymes and Kurz,
1967; McCabe, 1979; Fama, 1974; McDonald, Jacquillat and Nussenbaum, 1975; Smirlock
and Marshall, 1983; Gill and Green, 1993; among others). Therefore, their evidence is of
little use in assessing the empirical content of the hypothesis. Another limitation of the evi-

ß Blackwell Publishers Ltd 1998

ß Blackwell Publishers Ltd 1998
APPENDIX A
3SLS Regression Estimates Obtained when Dividend Payout Ratio and Financial Leverage are Represented by DIV93 and FL93 Respectively

3SLS Estimates Obtained Before Adding Industry Dummy Variables 3SLS Estimates Obtained After Adding Industry Dummy Variables

DIV96 FL96 INV DIV96 FL96 INV

Predicted

DIVIDEND PAYOUT RATIOS OF UK FIRMS

Variables Influence Coefs. Coef/se. Coefs. Coef/se. Coefs. Coef/se. Coefs. Coef/se. Coefs. Coef/se. Coefs. Coef/se.

Constant na 0.505 0.557 ÿ1.710 ÿ2.661* 2.290 2.473* 0.433 0.157 8.177 1.564 7.272 3.530*
DEFTX na/ÿ/na ÿ0.095 ÿ0.092 ÿ0.139 ÿ0.081
DIV93 na/+, ÿ/ÿ 0.163 0.265 0.252 0.066 1.453 1.444 1.249 0.318
FL93 +/na/+ 0.758 2.174* ÿ2.400 ÿ0.572 0.709 1.823* ÿ6.927 ÿ1.299
GRO na/na/+ ÿ5.030 ÿ2.770* ÿ7.386 ÿ3.247*
INDDY +/na/na ÿ0.548 ÿ0.321 ÿ0.281 ÿ0.059
INDFL na/+/na 4.392 3.644* ÿ21.837 ÿ1.595
INV ÿ/+/na ÿ0.044 ÿ2.086* 0.025 2.277* ÿ0.029 ÿ1.678* 0.040 1.613
IRRTX +, ÿ/na/na ÿ6.769 ÿ0.855 ÿ5.251 ÿ0.743
LIQ +/na/na 0.075 1.433 0.066 0.846
OVPRO ÿ/na/na ÿ0.187 ÿ4.304* ÿ0.220 ÿ4.704*
PRO +/ÿ/+ 0.835 1.525 0.159 0.369 ÿ2.097 ÿ0.364 0.783 1.327 0.302 0.326 ÿ4.126 ÿ0.686
q na/na/+ 0.048 4.973* 0.062 4.844*
RD na/ÿ/na 4.221 1.384 5.827 0.851
STR na/+/na 0.220 1.150 ÿ0.007 ÿ0.017
SZ +/+/ÿ 0.026 2.207* 0.013 0.620 ÿ0.452 ÿ2.530* 0.027 1.646* ÿ0.058 ÿ1.279 ÿ0.601 3.126*
VAR ÿ/+/na ÿ0.055 ÿ0.192 ÿ0.101 ÿ0.384 0.031 0.098 ÿ0.410 ÿ0.682
Likelihood
ratios 0.748 0.089 0.079 0.458 3.130 0.080
Degrees of
freedom/prob. 7/0.995 7/0.995 10/0.995 7/0.995 7/0.750 10/0.995

Notes:
* Indicates values that are significant at the 5% or higher levels. The estimates reported here are obtained by using the three stage least squares (3SLS) regression

1145
procedure in LIMDEP (version 7.0). In the predicted influence column, x/y/z indicates that the corresponding variable is predicted to have x, y and z influence on
DIV96, FL96 and INV respectively. na means that there is no prediction. Coef/se. ˆ Coefficient/standard error.
 1146
APPENDIX B
3SLS Regression Estimates Obtained when Dividend Payout Ratio and Financial Leverage are Represented by DIV94 and FL94 Respectively

3SLS Estimates Obtained Before Adding Industry Dummy Variables 3SLS Estimates Obtained After Adding Industry Dummy Variables

DIV96 FL96 INV DIV96 FL96 INV

Predicted
Variables Influence Coefs. Coef/se. Coefs. Coef/se. Coefs. Coef/se. Coefs. Coef/se. Coefs. Coef/se. Coefs. Coef/se.

Constant na 0.862 0.951 ÿ1.530 ÿ2.884* 5.476 2.777* ÿ0.507 ÿ0.204 6.092 1.852 7.986 3.654*
DEFTX na/ÿ/na ÿ0.223 ÿ0.588 ÿ1.544 ÿ1.851*
DIV94 na/+, ÿ/ÿ 0.216 0.834 ÿ7.257 ÿ2.003* 0.650 2.261* ÿ6.116 ÿ1.945*
FL94 +/na/+ 1.299 2.711* 2.442 0.730 0.906 1.998* ÿ1.078 ÿ0.288
GRO na/na/+ ÿ4.701 ÿ2.539* ÿ6.861 ÿ3.184*
INDDY +/na/na ÿ1.717 ÿ0.952 1.083 0.251

ADEDEJI
INDFL na/+/na 3.963 3.052* ÿ15.446 ÿ1.852*
INV ÿ/+/na ÿ0.063 ÿ2.663* 0.025 2.684* ÿ0.038 ÿ2.030* 0.036 2.179*
IRRTX +, ÿ/na/na ÿ14.973 ÿ1.580 ÿ6.558 ÿ0.858
LIQ +/na/na 0.144 2.153* 0.089 1.056
OVPRO ÿ/na/na ÿ0.120 ÿ2.394* ÿ0.177 ÿ3.685*
PRO +/ÿ/+ 1.200 1.991* 0.159 0.441 ÿ0.656 ÿ0.121 0.789 1.365 0.044 0.076 ÿ1.260 ÿ0.234
q na/na/+ 0.047 5.395* 0.057 5.690*
RD na/ÿ/na 3.218 1.885* ÿ1.577 ÿ0.741
STR na/+/na 0.220 1.876* 0.297 1.487
SZ +/+/ÿ 0.021 1.722* 0.009 0.598 ÿ0.193 ÿ1.052 0.025 1.488 ÿ0.030 ÿ1.388 ÿ0.354 ÿ2.061*
ß Blackwell Publishers Ltd 1998

VAR ÿ/+/na ÿ0.087 ÿ0.308 ÿ0.080 ÿ0.344 0.121 0.415 ÿ0.104 ÿ0.282
Likelihood
ratios 1.884 0.515 0.014 0.238 2.018 0.036
Degrees of
freedom/prob. 7/0.950 7/0.995 10/0.995 7/0.995 7/0.950 10/0.995

Notes:
* Indicates values that are significant at the 5% or higher levels. The estimates reported here are obtained by using the three stage least squares (3SLS) regression
procedure in LIMDEP (version 7.0). In the predicted influence column, x/y/z indicates that the corresponding variable is predicted to have x, y and z influence on
DIV96, FL96 and INV respectively. na means that there is no prediction. Coef/se. ˆ Coefficient/standard error.
ß Blackwell Publishers Ltd 1998
APPENDIX C
3SLS Regression Estimates Obtained when Dividend Payout Ratio and Financial Leverage are Represented by DIV95 and FL95 Respectively

3SLS Estimates Obtained Before Adding Industry Dummy Variables 3SLS Estimates Obtained After Adding Industry Dummy Variables

DIV96 FL96 INV DIV96 FL96 INV

Predicted

DIVIDEND PAYOUT RATIOS OF UK FIRMS

Variables Influence Coefs. Coef/se. Coefs. Coef/se. Coefs. Coef/se. Coefs. Coef/se. Coefs. Coef/se. Coefs. Coef/se.

Constant na ÿ0.574 ÿ0.740 ÿ1.638 ÿ2.941* 5.564 1.990* ÿ2.600 ÿ1.235 7.674 1.581 7.283 2.786*
DEFTX na/ÿ/na ÿ0.343 ÿ0.798 ÿ1.733 ÿ1.533
DIV95 na/+, ÿ/ÿ 0.062 0.233 ÿ2.359 ÿ0.445 0.956 1.781* ÿ1.897 ÿ0.383
FL95 +/na/+ 0.634 1.670* ÿ2.213 ÿ0.891 0.512 1.253 ÿ4.362 ÿ1.431
GRO na/na/+ ÿ5.375 ÿ1.964* ÿ6.171 ÿ2.562*
INDDY +/na/na 1.327 0.983 4.789 1.343
INDFL na/+/na 4.391 3.477* ÿ19.798 ÿ1.577
INV ÿ/+/na ÿ0.036 ÿ1.764* 0.022 1.962* ÿ0.029 ÿ1.497 0.030 1.332
IRRTX +, ÿ/na/na ÿ5.549 ÿ0.676 ÿ1.746 ÿ0.226
LIQ +/na/na 0.087 1.419 0.029 0.413
OVPRO ÿ/na/na ÿ0.153 ÿ2.958* ÿ0.165 ÿ3.072*
PRO +/ÿ/+ 0.451 0.953 0.144 0.364 ÿ2.205 ÿ0.462 0.329 0.631 0.143 0.176 ÿ1.574 ÿ0.331
q na/na/+ 0.050 5.918* 0.060 6.253*
RD na/ÿ/na 3.539 1.722* ÿ2.503 ÿ0.946
STR na/+/na 0.245 1.768* 0.376 1.483
SZ +/+/ÿ 0.024 2.030* 0.014 0.853 ÿ0.375 ÿ1.807* 0.037 2.366* ÿ0.052 ÿ1.351 ÿ0.487 ÿ2.340*
VAR ÿ/+/na 0.111 0.424 ÿ0.102 ÿ0.408 0.258 0.883 ÿ0.195 ÿ0.372
Likelihood
ratios 0.258 0.055 0.058 0.903 2.597 0.082
Degrees of
freedom/prob. 7/0.995 7/0.995 10/0.995 7/0.995 7/0.900 10/0.995

Notes:
* Indicates values that are significant at the 5% or higher levels. The estimates reported here are obtained by using the three stage least squares (3SLS) regression

1147
procedure in LIMDEP (version 7.0). In the predicted influence column, x/y/z indicates that the corresponding variable is predicted to have x, y and z influence on
DIV96, FL96 and INV respectively. na means that there is no prediction. Coef/se. ˆ Coefficient/standard error.
 1148
APPENDIX D
3SLS Regression Estimates Obtained when Dividend Payout Ratio and Financial Leverage are Represented by AVDIV and AVFL Respectively

3SLS Estimates Obtained Before Adding Industry Dummy Variables 3SLS Estimates Obtained After Adding Industry Dummy Variables

AVDIV AVFL INV AVDIV AVFL INV

Predicted
Variables Influence Coefs. Coef/se. Coefs. Coef/se. Coefs. Coef/se. Coefs. Coef/se. Coefs. Coef/se. Coefs. Coef/se.

Constant na ÿ0.674 ÿ1.218 ÿ1.213 ÿ2.819* 6.543 2.962* ÿ1.574 ÿ1.155 3.253 1.203 7.922 3.867*
DEFTX na/ÿ/na ÿ0.522 ÿ1.510 ÿ0.630 ÿ1.036
AVDIV na/+, ÿ/ÿ 0.233 0.853 ÿ9.521 ÿ1.756* 0.908 2.822* ÿ6.993 ÿ1.659*
AVFL +/na/+ 0.448 1.570 3.091 0.810 0.586 2.265* ÿ0.044 ÿ0.011
GRO na/na/+ ÿ6.437 ÿ3.313* ÿ6.613 ÿ3.535*
INDDY +/na/na 1.617 1.549 3.023 1.295

ADEDEJI
INDFL na/+/na 3.316 3.158* ÿ8.453 ÿ1.214
INV ÿ/+/na ÿ0.001 ÿ0.073 0.030 3.854* ÿ0.010 ÿ0.893 0.034 2.531*
IRRTX +, ÿ/na/na 4.687 0.812 4.527 0.999
LIQ +/na/na 0.053 1.239 0.021 0.499
OVPRO ÿ/na/na ÿ0.092 ÿ2.767* ÿ0.076 ÿ2.583*
PRO +/ÿ/+ 0.096 0.271 ÿ0.254 ÿ0.834 0.404 0.078 0.156 0.460 ÿ0.428 ÿ0.876 0.089 0.018
q na/na/+ 0.042 4.516* 0.053 5.294*
RD na/ÿ/na 2.294 1.622 ÿ1.891 ÿ1.066
STR na/+/na 0.220 2.034* 0.120 0.776
SZ +/+/ÿ 0.029 3.622* 0.009 0.668 ÿ0.141 ÿ0.643 0.035 3.504* ÿ0.029 ÿ1.365 ÿ0.322 ÿ1.687*
ß Blackwell Publishers Ltd 1998

VAR ÿ/+/na 0.025 0.131 ÿ0.038 ÿ0.205 0.069 0.390 ÿ0.064 ÿ0.204
Likelihood
ratios 1.252 6.912 0.027 0.148 0.370 0.042
Degrees of
freedom/prob. 7/0.975 7/0.250 10/0.995 7/0.995 7/0.995 10/0.995

Notes:
* Indicates values that are significant at the 5% or higher levels. The estimates reported here are obtained by using the three stage least squares (3SLS) regression
procedure in LIMDEP (version 7.0). In the predicted influence column, x/y/z indicates that the corresponding variable is predicted to have x, y and z influence on
DIV96, FL96 and INV respectively. na means that there is no prediction. Coef/se. ˆ Coefficient/standard error.
ß Blackwell Publishers Ltd 1998
APPENDIX E
OLS Regression Estimates Obtained when Dividend Payout Ratio and Financial Leverage are Represented by DIV93 and FL93 Respectively

OLS Estimates Obtained Before Adding Industry Dummy Variables OLS Estimates Obtained After Adding Industry Dummy Variables

DIV96 FL96 INV DIV96 FL96 INV

Predicted

DIVIDEND PAYOUT RATIOS OF UK FIRMS

Variables Influence Coefs. t-values Coefs. t-values Coefs. t-values Coefs. t-values Coefs. t-values Coefs. t-values

Constant na 0.261 1.573 0.220 1.971* 1.598 0.945 0.344 2.701* 0.498 7.390* 2.551 1.000
DEFTX na/ÿ/na ÿ0.802 ÿ3.083* ÿ0.856 ÿ3.421*
DIV93 na/+, ÿ/ÿ 0.037 0.772 1.864 1.940* 0.030 0.622 1.841 1.941*
FL93 +/na/+ 0.054 0.482 0.365 0.224 0.035 0.330 0.055 0.027
GRO na/na/+ ÿ3.620 ÿ1.173 ÿ3.889 ÿ1.120
INDDY +/na/na 0.465 2.142* 0.443 3.948*
INDFL na/+/na 0.495 2.254* ÿ0.269 ÿ2.648*
INV ÿ/+/na ÿ0.001 ÿ0.261 0.006 0.955 0.001 0.081 0.007 1.104
IRRTX +, ÿ/na/na 2.510 0.746 3.544 1.192
LIQ +/na/na ÿ0.001 ÿ0.029 ÿ0.001 ÿ0.011
OVPRO ÿ/na/na ÿ0.194 ÿ4.660* ÿ0.231 ÿ5.437*
PRO +/ÿ/+ 0.037 0.086 ÿ0.493 ÿ2.196* 1.754 0.323 0.036 0.084 ÿ0.532 ÿ2.388* 2.985 0.616
q na/na/+ 0.043 2.038* 0.049 1.930*
RD na/ÿ/na ÿ1.278 ÿ1.891* ÿ2.700 ÿ3.583*
STR na/+/na 0.205 2.076* 0.235 2.502*
SZ +/+/ÿ 0.032 3.181* ÿ0.003 ÿ0.454* ÿ0.451 ÿ2.217* 0.033 3.181* 0.001 0.013 ÿ0.516 ÿ2.087*
VAR ÿ/+/na 0.191 0.823 0.044 0.396 0.226 1.043 0.030 0.275
Adj. R2 0.118 0.172 0.187 0.145 0.198 0.185
F-ratio/prob. 4.32*/0.000 6.13*/0.000 9.54*/0.000 3.10*/0.000 4.06*/0.000 4.37*/0.000

Notes:
* Indicates values that are significant at the 5% or higher levels. The estimates reported here are obtained by using the OLS regression procedure in LIMDEP (version
7.0). The t-values are corrected for heteroscedasticity. In the predicted influence column, x/y/z indicates that the corresponding variable is predicted to have x, y and z

1149
influence on DIV96, FL96 and INV respectively. na means that there is no prediction.
 1150
APPENDIX F
OLS Regression Estimates Obtained when Dividend Payout Ratio and Financial Leverage are Represented by DIV94 and FL94 Respectively

OLS Estimates Obtained Before Adding Industry Dummy Variables OLS Estimates Obtained After Adding Industry Dummy Variables

DIV96 FL96 INV DIV96 FL96 INV

Predicted
Variables Influence Coefs. t-values Coefs. t-values Coefs. t-values Coefs. t-values Coefs. t-values Coefs. t-values

Constant na 0.258 1.500 0.205 1.879 2.102 0.899 0.343 2.678* 0.485 7.453* 3.135 0.955
DEFTX na/ÿ/na ÿ0.825 ÿ3.571* ÿ0.865 ÿ3.913*
DIV94 na/+, ÿ/ÿ 0.078 1.645* 0.109 0.152 0.086 1.816* ÿ0.164 ÿ0.201
FL94 +/na/+ 0.056 0.448 1.363 0.743 0.041 0.335 1.323 0.603
GRO na/na/+ ÿ3.816 ÿ1.111 ÿ4.006 ÿ1.041
INDDY +/na/na 0.468 2.145* 0.440 3.981*

ADEDEJI
INDFL na/+/na 0.472 2.176* ÿ0.321 ÿ3.433*
INV ÿ/+/na ÿ0.001 ÿ0.260 0.006 1.032 0.001 0.067 0.007 1.171
IRRTX +, ÿ/na/na 2.467 0.715 3.493 1.144
LIQ +/na/na 0.001 0.005 0.001 0.027
OVPRO ÿ/na/na ÿ0.193 ÿ4.590* ÿ0.230 ÿ5.398*
PRO +/ÿ/+ 0.028 0.067 ÿ0.460 ÿ2.095* 2.143 0.384 0.034 0.081 ÿ0.494 ÿ2.265* 3.241 0.663
q na/na/+ 0.041 1.888* 0.047 1.780*
RD na/ÿ/na ÿ1.294 ÿ1.917* ÿ2.693 ÿ3.535*
STR na/+/na 0.216 2.222* 0.243 2.640*
SZ +/+/ÿ 0.032 3.160* ÿ0.004 ÿ0.703 ÿ0.408 ÿ2.221* 0.033 3.163* ÿ0.002 ÿ0.263 ÿ0.467 ÿ2.081*
ß Blackwell Publishers Ltd 1998

VAR ÿ/+/na 0.195 0.845 0.060 0.532 0.228 1.059 0.046 0.417
Adj. R2 0.118 0.181 0.168 0.145 0.211 0.167
F-ratio/prob. 4.32*/0.000 6.47*/0.000 8.49*/0.000 3.10*/0.000 4.32*/0.000 3.97*/0.000

Notes:
* Indicates values that are significant at the 5% or higher levels. The estimates reported here are obtained by using the OLS regression procedure in LIMDEP (version
7.0). The t-values are corrected for heteroscedasticity. In the predicted influence column, x/y/z indicates that the corresponding variable is predicted to have x, y and z
influence on DIV96, FL96 and INV respectively. na means that there is no prediction.
ß Blackwell Publishers Ltd 1998
APPENDIX G
OLS Regression Estimates Obtained when Dividend Payout Ratio and Financial Leverage are Represented by DIV95 and FL95 Respectively

OLS Estimates Obtained Before Adding Industry Dummy Variables OLS Estimates Obtained After Adding Industry Dummy Variables

DIV96 FL96 INV DIV96 FL96 INV

Predicted

DIVIDEND PAYOUT RATIOS OF UK FIRMS

Variables Influence Coefs. t-values Coefs. t-values Coefs. t-values Coefs. t-values Coefs. t-values Coefs. t-values

Constant na 0.204 1.188 0.151 1.316 2.913 1.193 0.292 2.377* 0.493 7.953* 3.738 1.191
DEFTX na/ÿ/na ÿ0.834 ÿ3.389* ÿ0.870 ÿ3.665*
DIV95 na/+, ÿ/ÿ 0.173 3.740* 1.100 1.481 0.185 4.137* 1.445 1.800*
FL95 +/na/+ 0.131 1.037 ÿ0.932 ÿ0.383 0.143 1.208 ÿ1.465 ÿ0.540
GRO na/na/+ ÿ4.008 ÿ1.289 ÿ4.259 ÿ1.274
INDDY +/na/na 0.462 2.068* 0.404 3.659*
INDFL na/+/na 0.515 2.253* ÿ0.446 ÿ4.587*
INV ÿ/+/na ÿ0.002 ÿ0.323 0.005 0.867 ÿ0.001 ÿ0.071 0.005 0.992
IRRTX +, ÿ/na/na 1.865 0.536 2.714 0.896
LIQ +/na/na 0.012 0.350 0.016 0.479
OVPRO ÿ/na/na ÿ0.186 ÿ4.346* ÿ0.222 ÿ5.114*
PRO +/ÿ/+ 0.081 0.193 ÿ0.441 ÿ2.013* 0.116 0.020 0.112 0.269 ÿ0.467 ÿ2.188* 1.582 0.320
q na/na/+ 0.004 1.908* 0.051 1.861*
RD na/ÿ/na ÿ1.493 ÿ2.409* ÿ2.708 ÿ3.703*
STR na/+/na 0.193 2.008* 0.228 2.502*
SZ +/+/ÿ 0.033 3.256* ÿ0.008 ÿ1.288 ÿ0.459 ÿ2.202* 0.034 3.226* ÿ0.006 ÿ0.921 ÿ0.542 ÿ2.146*
VAR ÿ/+/na 0.186 0.783 0.039 0.359 0.215 0.969 0.027 0.271
Adj. R2 0.123 0.231 0.169 0.152 0.264 0.173
F-ratio/prob. 4.48*/0.000 8.42*/0.000 8.56*/0.000 3.22*/0.000 5.44*/0.000 4.11*/0.000

Notes:
* Indicates values that are significant at the 5% or higher levels. The estimates reported here are obtained by using the OLS regression procedure in LIMDEP (version
7.0). The t-values are corrected for heteroscedasticity. In the predicted influence column, x/y/z indicates that the corresponding variable is predicted to have x, y and z

1151
influence on DIV96, FL96 and INV respectively. na means that there is no prediction.
 1152
APPENDIX H
OLS Regression Estimates Obtained when Dividend Payout Ratio and Financial Leverage are Represented by AVDIV and AVFL Respectively

OLS Estimates Obtained Before Adding Industry Dummy Variables OLS Estimates Obtained After Adding Industry Dummy Variables

AVDIV AVFL INV AVDIV AVFL INV

Predicted
Variables Influence Coefs. t-values Coefs. t-values Coefs. t-values Coefs. t-values Coefs. t-values Coefs. t-values

Constant na 0.017 0.131 0.153 1.500 2.114 0.910 0.152 1.651 0.419 7.040* 3.080 0.948
DEFTX na/ÿ/na ÿ0.897 ÿ3.815* ÿ0.952 ÿ4.160*
AVDIV na/+, ÿ/ÿ 0.194 3.183* 1.442 1.284 0.195 3.437* 1.271 1.250
AVFL +/na/+ 0.305 2.833* 0.087 0.035 0.322 3.127* ÿ0.092 ÿ0.033
GRO na/na/+ ÿ3.860 ÿ1.183 ÿ4.128 ÿ1.121
INDDY +/na/na 0.586 3.033* 0.439 5.831*

ADEDEJI
INDFL na/+/na 0.546 2.871* ÿ0.178 ÿ2.031*
INV ÿ/+/na 0.007 1.839* 0.008 1.522 0.007 2.037* 0.009 1.620
IRRTX +, ÿ/na/na 4.238 1.877* 4.726 1.909*
LIQ +/na/na 0.029 1.048 0.024 0.897
OVPRO ÿ/na/na ÿ0.105 ÿ3.549* ÿ0.129 ÿ4.401*
PRO +/ÿ/+ ÿ0.002 ÿ0.007 ÿ0.721 ÿ3.285* 1.241 0.212 ÿ0.012 ÿ0.042 ÿ0.739 ÿ3.414* 2.530 0.497
q na/na/+ 0.043 1.885* 0.048 1.790*
RD na/ÿ/na ÿ1.191 ÿ2.522* ÿ2.014 ÿ3.435*
STR na/+/na 0.198 2.081* 0.222 2.380*
SZ +/+/ÿ 0.003 5.441* ÿ0.006 ÿ1.131 ÿ0.452 ÿ2.115* 0.034 5.211* ÿ0.005 ÿ0.865 ÿ0.514 ÿ2.036*
ß Blackwell Publishers Ltd 1998

VAR ÿ/+/na 0.007 0.042 0.068 0.515 0.041 0.246 0.054 0.436
Adj. R2 0.227 0.335 0.169 0.229 0.344 0.168
F-ratio/prob. 8.29*/0.000 13.51*/0.000 8.57*/0.000 4.67*/0.000 7.50*/0.000 4.00*/0.000

Notes:
* Indicates values that are significant at the 5% or higher levels. The estimates reported here are obtained by using the OLS regression procedure in LIMDEP (version
7.0). The t-values are corrected for heteroscedasticity. In the predicted influence column, x/y/z indicates that the corresponding variable is predicted to have x, y and z
influence on DIV96, FL96 and INV respectively. na means that there is no prediction.
 DIVIDEND PAYOUT RATIOS OF UK FIRMS 1153
dence reported by these studies is that the variables used in them are inconsistent with the
pecking order hypothesis. For instance, while the hypothesis predicts an interaction between
dividend payout ratio and investment, the studies only test the interaction between divi-
dends and investment. Similarly, instead of the interaction between financial leverage and
investment, the studies focus on the interaction between debt and investment. Since a high
level of dividends or debt may not necessarily imply a high dividend payout ratio or financial
leverage, it may not be right to assume that the nature of interactions observed among divi-
dends, debt and investment will apply to dividend payout ratios, financial leverage and in-
vestment.
3 Before the refund of Advance Corporation Tax (ACT) was abolished in July 1997, tax exempt
institutions (e.g. pension funds) received tax refunds from the Inland Revenue at the rate of
20% on all dividends that they received. This probably encouraged such institutions to prefer
dividends to capital gain. The system also probably encouraged such institutions to pressurise
firms to pay dividends.
4 In general, previous studies which used dummy variables to test industry influence do not par-
tition their samples in the way described above. Instead, they exclude small industries (i.e. in-
dustries with few firms) and use one of the large industries in their samples as the control group
(e.g. Baskin, 1989; and Bennett and Donnelly, 1993). This method is not used in the present
study because it will lead to the exclusion of 43 firms which are from `small industries'. Apart
from the benefit of a large sample size, another advantage of the method used in this study is
that it does not require a decision to be made about which industry to use as the control group.
This point is important because the result of a test of industry influence may be sensitive to the
industry chosen as the control group. In the method used in this study, the control group con-
sists of firms from a number of industries. Given its large size and broad composition, the group
can be viewed as a surrogate for the rest of the economy. Its purpose in the test is to serve as a
benchmark for determining whether the value of each of the dependent variables in equations
(4)^(6) for an industry is significantly different from the corresponding value for the rest of the
economy.
5 The investment variable referred to here (INV) is similar to the one used by Baskin (1989) and
Allen (1993) which is defined as the ratio of total assets at the end of the study period to the
corresponding value at the beginning of the study period. Another variable which could have
been used to measure investment is the amount of total expenditure on new fixed assets. Baskin
argued against the use of such measure on the ground that it ignores investment in working
capital. He suggests that the measure of investment based on total assets is theoretically more
appropriate because it includes investment in both fixed assets and working capital (see Ba-
skin, 1989, p. 32). The average value of the investment variable (INV) referred to here is large
because the distribution of the variable is highly skewed. The value of the variable is less than 1,
or 100%, for about 65% of the sample. The median value of the variable is 0.72, or 72%.
6 In preparing the summary reported in Table 3, the estimates obtained with AVDIV and
AVFL are given more weight than those obtained with DIV96 and FL96, in accordance with
the argument of the previous studies referred to earlier. For instance, if a variable is significant
in the estimate obtained when AVDIV is used as the dependent variable, it is reported in Table
3 that the variable is significant, even if the variable is not significant in any of the estimates
obtained when DIV96 is used as the dependent variable. But if a variable is significant only
once in the three estimates obtained with DIV96 but not significant in the estimate obtained
with AVDIV, it is reported in the table that the variable is not significant. However, if a vari-
able is significant in two of the three estimates obtained with DIV96, it is reported that the
variable is significant, even if the variable is not significant in the estimate obtained with AV-
DIV. The same rule is applied to the estimates obtained with FL96/AVFL and INV.
7 The liquidity variable (LIQ) has a generally high correlation with the financial leverage vari-
ables (FL93, FL94, FL95, FL96 and AVFL) ö see Table 2. This probably explains why the
variable is not significant in the results referred to here.
8 The results of the tests are included in Appendices A-H.

REFERENCES
Allen, D.E. (1993), `The Pecking Order Hypothesis: Australian Evidence', Applied Financial Eco-

ß Blackwell Publishers Ltd 1998

1154 ADEDEJI
nomics, Vol. 3, pp. 101^12.
Ang, J.S. and M. Jung (1993), `An Alternate Test of Myers' Pecking Order Theory of Capital
Structure: The Case of South Korean Firms', Pacific-Basin Finance Journal, Vol. 1, pp. 31^46.
Baskin, J. (1989), `An Empirical Investigation of the Pecking Order Hypothesis', Financial Man-
agement (Spring), pp. 26^35.
Bennett, M. and R. Donnelly (1993), `The Determinants of Capital Structure: Some UK Evi-
dence', British Accounting Review, Vol. 25, pp. 43^59.
Bradley, M., G.A. Jarrell and E.H. Kim (1984), `On the Existence of an Optimal Capital Struc-
ture: Theory and Evidence', Journal of Finance, Vol. 39 (July), pp. 857^78.
Chang, R.P. and S.G. Rhee (1990), `The Impact of Personal Taxes on Corporate Dividend Policy
and Capital Structure Decisions', Financial Management (Summer), pp. 21^31.
Chui, A., N. Strong and J. Cadle (1992), `The Empirical Significance of Tax Effects on the Valua-
tion of Dividends: The UK Evidence', Journal of Business Finance & Accounting, Vol. 49, No. 4
(June), pp. 515^32.
DeAngelo, H. and R. Masulis (1980), `Optimal Capital Structure Under Corporate and Personal
Taxation', Journal of Financial Economics, Vol. 8 (March), pp. 3^29.
Dhrymes, P.J. and M. Kurz (1967), `Investment, Dividends and External Finance Behaviour of
Firms', in R. Ferber (ed.), Determinants of Investment Behavior (Columbia University Press,
New York).
Donaldson, G. (1961), Corporate Debt Capacity: A Study of Corporate Debt Policy and the Determination of
Corporate Debt Capacity (Boston, Division of Research, Harvard Graduate School of Business
Administration).
Fama, E.F. and H. Babiak (1968), `Dividend Policy: An Empirical Analysis', Journal of the Amer-
ican Statistical Association (December), pp. 1132^61.
________________________ (1974), `The Empirical Relationships Between the Dividend and Investment Decisions of
Firms', American Economic Review, Vol. 64, No. 3, pp. 304^18.
Gill, J. and J.P. Green (1993), `Dividend Policy and its Relationship to Investment and Financing
Policies: UK Empirical Evidence', Working Paper (University of Ulster).
Greene, W.H. (1993), Econometric Analysis (2nd ed.,Macmillan).
Jensen, M.C. (1986), `Agency Costs of Free Cash Flow, Corporate Finance and Takeovers', Amer-
ican Economic Review, Vol. 76, pp. 323^29.
Lasfer, M.A. (1995), `Agency Costs, Taxes and Debt: The UK Evidence', European Financial Man-
agement, Vol. 1, pp. 265^85.
________________________ (1996), `Taxes and Dividends: The UK Evidence', Journal of Banking and Finance (April),
pp. 455^72.
Levis, M. (1989), `Stock Market Anomalies: A Reassessment Based on UK Evidence', Journal of
Banking and Finance, Vol. 13, pp. 675^96.
Lintner, J. (1956), `Distribution of Incomes of Corporations Among Dividends, Retained Earn-
ings and Taxes', American Economic Review (May), pp. 97^113.
Litzenberger, R.H. and K. Ramaswamy (1982), `The Effects of Dividends on Common Stock
Prices: Tax Effects or Information Effects', Journal of Finance, Vol. 37, pp. 429^43.
Madura, J. (1995), International Financial Management (4th ed., West Publishing Company, New
York).
Marsh, P. (1979), `Equity Rights Issues and the Efficiency of the UK Stock Market', Journal of
Finance, Vol. 34 (September), pp. 839^62.
________________________ (1982), `The Choice Between Equity and Debt: An Empirical Analysis', Journal of Finance,
Vol. 37 (March), pp. 121^44.
Masulis, R.W. and A.N. Korwar (1986), `Seasoned Equity Offerings: An Empirical Investiga-
tion', Journal of Financial Economics, Vol. 15, pp. 91^118.
McCabe, G.M. (1979), `The Empirical Relationship Between Investment and Financing: A New
Look', Journal of Financial and Quantitative Analysis, Vol. 14 (March), pp. 119^35.
McDonald, J.G., B. Jacquillat and M. Nussenbaum (1975), `Dividend, Investment and Financing
Decisions: Empirical Evidence on French Firms', Journal of Financial and Quantitative Analysis,
Vol. 10 (December), pp. 741^55.
Mikkelson, W.H. and M.M. Partch (1986), `Valuation Effects of Security Offerings and the Issu-
ance Process', Journal of Financial Economics, Vol. 15, pp. 31^60.
Miller, M. and F. Modigliani (1961), `Dividend Policy, Growth and the Valuation of Shares',
Journal of Business, Vol. 34, pp. 411^33.

ß Blackwell Publishers Ltd 1998

 DIVIDEND PAYOUT RATIOS OF UK FIRMS 1155
________________________ and K. Rock (1985), `Dividend Policy Under Asymmetric Information', Journal of Fi-
nance, Vol. 40 (September), pp. 1031^51.
Myers, S.C. (1977), `Determinants of Corporate Borrowing', Journal of Financial Economics (No-
vember), pp. 147^76.
________________________ (1984), `The Capital Structure Puzzle', Journal of Finance, Vol. 39 (July), pp. 575^92.
________________________ and N.S. Majluf (1984), `Corporate Financing and Investment Decisions When Firms
Have Information Investors Do Not Have', Journal of Financial Economics (June), pp. 187^
221.
Noronha, G.M., D.K. Shome and G.E. Morgan (1996), `The Monitoring Rationale for Dividends
and the Interaction of Capital Structure and Dividend Decisions', Journal of Banking and Fi-
nance, Vol. 20, pp. 439^54.
Scott, D. (1977), `Bankruptcy, Secured Debt and Optimal Capital Structure', Journal of Finance,
Vol. 32 (March), pp. 1^19.
Smirlock, M. and W. Marshall (1983), `An Examination of the Empirical Relationship Between
the Dividend and Investment Decisions: A Note', Journal of Finance, Vol. 38 (December), pp.
1659^67.
Theobald, M. (1978), `Intertemporal Dividend Models ö An Empirical Analysis Using Recent
UK Data', Journal of Business Finance & Accounting, Vol. 5, No. 1 (Spring), pp. 123^32.
Titman, S. and R. Wessels (1988), `The Determinants of Capital Structure Choice', Journal of Fi-
nance, Vol. 43 (March), pp. 1^19.
Viswanath, P.V. (1993), `Strategic Considerations, the Pecking Order Hypothesis, and Market
Reactions to Equity Financing', Journal of Financial and Quantitative Analysis, Vol. 28, No. 2
(June), pp. 213^34.
Weston, J.F. and T.E. Copeland (1992), Managerial Finance (9th ed., Dryden Press, London).

ß Blackwell Publishers Ltd 1998