In: Finance
1.onsider what you feel would be an optional capital structure.
a.Would your recommendation differ depending on the type of industry? Why/why not?
b.What would you expect the impact to be on earnings per share (EPS) at an optimal capital structure? Is this impact important to you? Why/why not?
Capital Structure Across Industries
1. Introduction
1.1 Purpose of this Study
From the number of published papers concerning capital structure, it is apparent
that this is an important, difficult and complex subject. Every company would
like a capital structure which is best fitted to their current situation that minimizes
the cost of capital. Without good knowledge of the variables that control capital
structure, more insight about what might be the optimal debt ratio cannot be
revealed.
1.2 The Structure of this Study
The paper begins with a short presentation of the existing theory used as a theoretical
approach later in this text. Next, the capital structure hypothesis to be
addressed will be formulated, followed by an introduction to the background
information and methods used in the study. The main part of this paper starts
with a presentation of the descriptive statistics, followed by separate regressions
and the evaluations for the five industries studied. Subsequently, the pooled
regression with industry dummies is presented. The empirical part of this paper
ends with a pooled regression on the 50 largest companies in the sample. In the
conclusion the main results are discussed and areas for further research will be
suggested.
2. A Short Review
2.1 Existing Theory
In the first theorem from Modigliani and Miller (1958), that assumes a perfect
arbitrage free capital market, they claim that the value of a firm is independent of
the capital structure. Miller (1977) found a balance between supply and demand
of debt under both private and corporate taxes. Through an introduction of real
life market imperfections like tax, but also expected bankruptcy costs, Myers
(1984) claimed that the firm value is dependent on the debt ratio. This aspect of
the capital structure is known as the static trade-off theory, since companies
balance their debt ratio against bankruptcy costs and tax savings.
Another capital structure theory is the pecking order theory (e.g. described in
Myers (1984)). This theory is a funding hierarchy, explained by transaction and
issuing cost argumentation. A firm requiring capital for a new project will use its
retained earnings first. Then they should issue debt and lastly issue new equity.
The pecking order theory is also explainable using information asymmetry argumentation,
described in Ross (1977), Myers and Majluf (1984), and Harris and
Raviv (1990). The asymmetric information distribution between firm managers
and investors brings the signalling behaviour connected with capital structure
into focus (hidden information problem). Myers and Majluf’s model (1984)
explains that equity issues are known as a bad signal for investors, since they are
particularly lucrative for already overvalued companies. Hence, this theory
supports the pecking order funding hierarchy.1
Barclay and Smith Jr (1995) argue that the size of a firm will affect the capital
structure for two reasons: first, the fixed issuance costs for public issues are large,
resulting in a significant scale economy effect in favour of large firms. Second,
large firms are more likely to have foreign operations, and to manage their
currency exposure, they want to use foreign debt. As many non-US debt markets
are not that liquid, especially for long-term debt, large companies will use mostly
foreign short-term debt and hence influence the long-term debt ratio. These
hypotheses are contradictory, as we will point out later in the study.
The costs of debt and equity are not constant over time. According to von
Nitzsch and Rouette (2006), firms should finance their company with equity at
times when it is cheap compared to debt, and vice versa. Therefore, purely opportunistic
reasons also lie behind the capital structure.
As there is no model that summarizes all the theories concerning capital
structure into one, it is difficult to decide which theory that affects the capital
structure of each firm that is studied. There are many factors that affect a firm
simultaneously.
2.2 Empirical Studies
Since there are many complicated factors, economists are working to explain the
main driving factors behind the capital structure of firms. Welch (2004) introduces
a dynamic model where stock returns and net issuing activity explain most
of debt ratio changes. With a static view concerning company variables such as
turnover and profit, Barclay and Smith (2005) presents a review of the evidence
for a target capital structure.
The number of capital structure papers focusing on industrial sectors is relatively
small compared to other types of capital structure research. A paper from
MacKay and Phillips (2005) and an extensive paper from Fan et al. (2003) come
close to the problem we wish to address as they present the capital structure of
firms depending on the industries as well as several other variables. Hall et al.
(2000) have an industry and capital structure study of small- and medium-sized
unquoted enterprises (so-called SMEs) in the United Kingdom. We have used
their paper as a benchmark for our study, owing to the fact that the capital structure question of US publicly quoted companies, which are not expected to
experience the constrained pecking order, is a natural continuation of their work.
3. Capital Structure Hypotheses
3.1 Capital Structure and the Selected Industries
Every industry experiences its own set of economic conditions. For instance, if a
company is operating in an industry with very volatile earnings, it tends to have
more equity as a buffer against possible bankruptcy (Balakrishnan and Fox, 1993).
Additionally, industries are subject to different challenges within technology
development, environmental regulations, etc.
The construction industry is known as being sensitive to general market
conditions, since building projects are associated with high initial expenses.
Building plans are often cancelled in periods of economic downturn. The food
industry is assumed to be more stable, owing to the fact that food is a basic need.
Simultaneously, this is supposed to be an industry with relatively hard competition.
Oil and gas production is a capital-intensive industry with, at present, very
high operating margins. As a representative for a mature industry with a relative
neutral growth outlook, we have chosen the chemical industry. The information
technology (IT) industry is our representative for the new economy. The IT firms
have a relatively small amount of fixed assets compared to the other industries
(mostly human capital), and they have a strong market outlook. Titman (1984) has
shown that capital structure can be used to commit the investors to liquidate only
in those states where the net gain of liquidation exceeds the costs to customers.
The customer’s cost rises from the inability to obtain the product, parts or service
for already acquired products. The industry dummies may measure the extent of
interaction between product market characteristics and the debt levels. Firms
where this effect is pronounced, such as computer and automotive firms, are
expected to have less debt, ceteris paribus, than firms where this effect is less
important, such as industries with homogenous goods.
3.2 Hypothesis Formulation
Hall et al. (2000) find industry differences in unquoted small- and medium-sized
UK enterprises. In addition to significant variables from the capital structure
framework, we expect to find differences between industries in our study.
Through our empirical study, we want an answer to these hypotheses by means
of statistical hypothesis testing:
H0: There is no significant difference in capital structure between
industries.
H1: There is a significant difference in capital structure between
industries.
The null hypothesis, as every other statistic in this study, will be evaluated at a 5%
significance level. The determinants that drive the debt ratio, the various capital
market imperfection hypotheses and incomplete contracts theory, are by definition
unobservable. Agency costs of management cannot, for instance, be
measured directly. The other variables we use will in some cases contain several
hypotheses in each variable, for instance profitability will contain both the
bankruptcy cost hypothesis, documented by Warner (1977) and Andrade and
Kaplan (1998), and the free cash flow hypothesis of Jensen (1986). Both hypotheses
predict a positive relation between debt ratio and profitability. It will also
contain Myers and Mailuf (1984) asymmetric information hypothesis which
predict a negative relation between debt and profitability.
The size variable will likely contain both the asymmetric information hypothesis
and bankruptcy costs hypothesis, since a larger firm is more transparent and
less likely to go bankrupt. A more transparent firm will find equity capital less
costly, with reduced expected bankruptcy costs. It is therefore difficult to set up
conclusive a priori conjectures about predictions for the independent variables.
However, we can use previous studies like Hall et al. (2000) and Harris and
Raviv (1990) to predict the signs of the regression coefficients. As it is difficult to
decide which of the debt ratio factors connected with company size that is most
effective, there is uncertainty about the size sign. The predicted significant
coefficients are collateral (+), profitability (+/−), growth (−), size (+/−) or age (−).as
Capital Structure and Earnings Per Share
To this point in this book we have assumed that the capital structure
decision was made based on the joint interests of management and the
stockholders, and that a primary consideration was the effect of debt
issuance on the value of the firm. Thus we have made use of two value
formulations:
without investor taxes and with investor taxes:
Now we will shift to a focus on the effect of a change in capital
structure on the finn's EPS. The implication ofthe analysis is that high PIE
firms will tend not to issue debt because of the adverse effect of debt
issuance on the finn's EPS. Management should not make decisions based
solely on the effect on earnings per share (EPS) but the reality is that the
effect of a decision on EPS does influence choices that businesses make.
Will a capital structure change increase the firm's earnings per share?
The basic formulation for the effect ofsubstituting debt for equity
(with no investor taxes and no costs offinancial distress) is:
The value ofthe stock after the debt issuance is:
Since the debt proceeds are assumed to be distributed to the stockholders,
the total value ofthe shareholders' position after debt issuance is:
Total Wealth ofShareholders =S +B=VL
These formulations are based on the cash flow and earnings streamst
with and without debt. They are not based on the immediate effect of debt
issuance on EPS.
But assume that there is a value S· that is determined by the market
after the debt issuance and share repurchase and the S· is not necessarily
equal to S as determined above. Now the total shareholder wealth after
issuance is:
Total Wealth ofShareholders = S· + B
but now S· + B may not be equal to S + B as previously computed.
Appendix A shows that the EPS is decreased by debt issuance and
share repurchase if(l-te)kj is larger than EPS where P is the price paid for
P
the shares and EPS is the earnings per share before debt issuance.
Appendixes Band C show the derivation ofP, the stock price after
debt issuance. Since the debt issuance also increases ris~ the stockholders
who do not sell their shares will likely consider themselves to be harmed
(lower EPS and more risk) by the debt issuancet if the effect on EPS and
risk to common stock are the only two considerations.
Example
Assume N = ItOOOtOOO shares and Total Income = $8tOOOtOOO so
that EPS = $8; the stock can be purchased at $100.
EPS te=.35t kj =.10t -=.08t and(l-te)~=.065.
P
Since (l-te)kj is smaller than EPS we expect the debt issuance to be
P
beneficialt based on the EPS effect.
If$50tOOOtOOO of.10 debt is issued (500tOOO shares purchased) the
after tax cost ofinterest is $3t250tOOO. The new EPS is:
EPS = 8tOOOtOOO-3t250tOOO 4t750tOOO =$9.50
, ltOOOtOOO - 500tOOO 500tOOO
We have EPS = .08 is larger than (l-t)ki = .065 and EPS is increased by P
debt issuance and share repurchase from $8 to $9.50. The EPS increases
since E:S islarger than (l-te)kj•
We are not recommending that the capital structure decision be
based on the effect of debt issuance on EPS. However, it would be
surprising if management did not consider the effect of debt issuance on
EPS when the capital structure decision is being made.
A Changed Example: P =$200
Now assume the same facts except the stock can be bought at $200
and EPS =$8 so that EPS =.04 and now we expect that debt issuance will
P
not be beneficial based on EPS. Assume the stock price before consideration
ofdebt issuance and proposed share repurchase was $182.50.
If $50,000,000 of .10 debt is issued (250,000 shares purchased at a
price of$200), the after tax cost ofinterest is $3,250,000. The new EPS is:
EPS = 8,000,000 - 3,250,000 4,750,000 =$6.33
1,000,000 - 250,000 750,000
The EPS is reduced from $8 to $6.33 at the same time that risk is
increased by the substitution ofdebt for stock.
If the finn's EPS is smaller than (l-t)kj, the debt issuance will not
P
be desirable for the remaining shareholders (those who do not sell) from the
viewpoint ofthe changed EPS.
If all shareholders sold a pro-rata percentage oftheir stock, the debt
issuance might again be deemed desirable, but the tax authorities would
treat the distribution as a dividend, not a capital gain.
A Value Perspective
Assume that before the debt issuance the value ofthe firm (Vu) was
$182.50 (1,000,000) = $182,500.000. With $50,000,000 of debt, teD =
$17,500,000 and
VL =182,500,000 + 17,500,000 = $200,000,000.
The total value ofthe outstanding stock after the $50,000,000 debt
issuance is now:
S=VL - B=200,000,000 -50,000,000 =$150,000,000
The total value ofthe shareholders' wealth after debt issuance is:
Total Wealth ofShareholders =150,000,000 + 50,000,000 =$200,000,000
an increase of$17,500,000 over the value with no debt. The value per share
after debt issuance is:
150,000,000 =$200.00
750,000
up from $182.50 before the debt issuance. But the EPS calculations showed
a decrease in EPS from $8 to $6.33.
In theory, the analysis leading to the capital structure decrease
should be based on value, and not on the effect on EPS. When there is an
adverse effect for today's EPS, the firm's future growth will be spread over
fewer shares and the value will accrue to the 750,000 outstanding shares
rather than the initial 1,000,000 shares.
The Stock Value With Growth
Assume the stock price was $182.50 per share before debt issuance
and that 250,000 shares are purchased at a price of $200 per share. Before
debt issuance and share repurchase the owners of 750,000 shares not
repurchased would have a value of:
750,000 ($182.50) =$136,875,000
After the debt issuance and share repurchase these investors would
have a value of:
750,000 ($200.00) =$150,000,000
But the EPS would be decreased from $8.00 to $6.33.
Assume the dividend before debt issuance was $1, the cost ofequity
was .15 and the growth rate was .14452.
P= I $182.50
.15 - .14452
The $1,000,000 of dividends will grow to $14,900,000 in 20 years
(g = .14452) and the value of the ftrm will be $2,715,000,000. If the
$50,000,000 debt obligation grows at .065 (after tax) to be $176,000,000 the
stockholder value will be $2,539,000,000 for the remaining shareholders
versus .75 (2,715,000,000) = $2,036,000,000 without the debt.
The debt substitution is desirable from a value perspective even
though the EPS is initially decreased, if future growth of the ftrm is
considered.
The Relationship ofPo and p.
Ifthe stock is repurchased at a price of P =PI with B of debt, what
is the relationship ofPo and PI? Po is the initial stock price.
Appendix B shows that:
For the example, where Po = $182.50, teB = .35(50,000,000) =
$17,500,000, and N = 1,000,000 shares, the value ofPI is:
PI = 182.50 + 17,500,000 =182.50+17.50 =$200.00
1,000,000
This agrees with the calculation for the value per share that was
previously computed for this example.
Appendix C shows that ifthe stock is repurchased at a price of Po.
(rather than PI) that:
v - B(l-t ) P _ u c 1- B N--
Po
For the example if Vu =$182,500,000, Po = $182.50 and B =
$50,000,000 we have:
PI = 182,500,000 - 50,000,000(.65) 150,000,000 $206.60
1000000 _ 50,000,000 726,027
" 182.50
The ex-post price is larger since the 273,973 shares are purchased at
a lower price of$182.50 (rather than $200.00 per share).
EPS Targets
Each calendar quarter, security analysts estimate the EPS of firms
and the CFO's ofthese firms then tend to try to have their firms meet the
targets to avoid a threat to their stock prices.
If a firm will be slightly short ofthe target it can implement a debt
issuance, share repurchase strategy, before the end ofthe quarter.
Example
X = $153.85 and $153.85 (1-~) = $100, N = 50, EPS = $2, tc = .35,
P=$20
Assume the target EPS is $2.10 and the CFO does not want to
disappoint the analysts. With no actions the firm will earn $2 per share.
Assume $400 ofdebt can be issued to yield .08. The interest is $32
and the debt proceed can be used to buy 20 shares. The new EPS is:
EPS = (153.85 - 32)(1-.35) =79.20 =$2.64
50-20 30
The $2.10 target is easily met.
Ifthe debt is issued near the year end so that the interest for the year
ofissuance equals zero we would have:
EPS = 100 = $3.33.
50-20
It is not necesS8ty to issue as much debt as $400. A smaller amount
would accomplish the objective ofthe finn earning $2.10 per share.
The June 27, 2001 Wall Street Journal reported that Merrill Lynch
warned that the earnings for the second quarter would be between $.52 and
$.57 a share. The earnings forecasts had been for $.82. The firm's stock
price dropped from $66.45 to $58.91 (down $7.54 or 11.3%).
It is not uncommon to have a desirable acquisition of a finn vetoed
by management because the acquisition would dilute (reduce) a finn's EPS.
Consider the following situation (there are zero operating synergies):
Earnings (after tax)
Shares
EPS
Initial Price ofa Share
Parent Firm
$3,360
1,600
$2.10
$ 20
Target Firm
$400
200
$ 2
$ 50
Shares of Parent to be issued: 3 shares for each Target share (600
shares). This is a "nominal" .20 premium: 60 -1 = .20. The pro-forma EPS
50
are $1.71.
3,760 Pro-forma EPS = -- = $1.71
2,200
Assume the reduction in EPS from $2.10 to $1.71 as a result ofthe
merger is not acceptable.
Now assume that Target can be acquired for $10,000 cash
(consistent with the $50 price and 200 shares outstanding).
Parent acquires Target and issues $10,000 of.06 debt. The after tax
interest cost is 600(1-.35) =$390. The new total earnings net ofinterest are
3,760 -390 =$3,370. The new pro-forma EPS is:
EPS = 3,370 =$2.11
1,600
Now there is a very small amount ofearnings accretion.
Conclusions
This chapter has focused on the effect of debt issuance on EPS.
While theory focuses on value, management practices frequently pay
attention to the effect debt issuance has on EPS. EPS can be manipulated by
a strategy ofdebt issuance and share repurchase. The capital structure deci-
sion should not be based on EPS, but the effect on EPS is likely to be
considered by management.
The result ofthe analysis is that we conclude that a high PIE firm (a
low EPS fIrm) will tend to be more reluctant to issue debt in substitution
p
for stock knowing that the effect on EPS will not be favorable.