9/8/2022

3.1 Computing the Value of the Firm ’s Equity, E

• Of all the computations related to the WACC, computing the value of the
Chapter Three firm’s equity is the easiest: As long as the company is publicly listed, take E to
be the product of the number of shares outstanding times the current value
Determining the value of the firm per share.
This chapter discusses the computation of the five components of the • As an example, consider El Paso Pipeline Partners (EPB), a New York Stock
WACC -the market value of the firm’s equity E, Exchange company that owns gas pipelines and gas storage facilities. On 29
June 2012, EPB has 205.7 million shares outstanding, each trading at $33.80.
-the market value of debt D , The equity value of the company is $6.953 billion.
-the firm ’s tax rate TC , the firm’s cost of debt rD , and
-the cost of equity rE.

1 2

3.2 Computing the Value of the Firm ’s Debt, D • For purposes of computing the weighted average cost of capital,
our definition of debt excludes other debt-like items such as
• We compute the value of the firm’s debt by the market value of its financial
debt minus the market value of its excess liquid assets. A common pension liabilities and deferred taxes. Though we consider these
approximation for this number is to take the balance sheet value of the items as debts, it is hard to attach a cost to them; we prefer to
firm’s debt minus the value of the firm’s cash balances and minus the value
of its marketable securities. Here ’s an example for Kroger:
approximate the WACC by using only financial obligations net of
liquid assets.
• It is not uncommon for a company to have negative net debt this
occurs.
• when the company has more cash and marketable securities than
debt. When this occurs, we set D in the WACC computation to be a
negative number. Both Intel and Whole Foods Markets are
3
examples: 4

3.3 Computing the Firm ’s Tax Rate, TC

• In the WACC formula, TC should measure the firm’s marginal tax rate, but it
is common to measure it by computing the firm’s reported tax rate. Usually
this should cause no problems, as the following example shows:

• The tax rate for Whole Foods is reasonably stable at 38% to 41%. In our
WACC computation we would most likely use the current tax rate or the
average over the past several years.
• Sometimes, however, this doesn’t work, as the following example shows:

5 6

1
 9/8/2022

3.5 Computing the Firm ’s Cost of Debt, rD

We now turn to calculating the cost of debt rD. In principle, r D is
the marginal cost to the firm (before corporate taxes) of borrowing
an additional dollar.
 Companies like Merck are very good at placing their income There are at least three ways of calculating the firm’s cost of debt.
in comfortable tax venues, and it appears that a reasonable We will state them briefly below and then go on to illustrate the
estimate for the tax rate is some where between 13% and application of two of
15%. the methods that although they may not be theoretically perfect are
often used in practice:
 In 2010, a year of low income for Merck, these tax-planning strategies
evidently did not work. Assuming that Merck’s future profitability is  As a practical matter, the cost of debt can often be approximated by
indicated by the two good years 2009 and 2011, we would most likely taking the average cost of the firm’s existing debt. The problem with
assume that Merck’s company’s future tax rate TC is in the range of this method is that it runs the danger of confusing the past costs
with the future anticipated cost of debt that we actually want to
the 2009 and 2011 tax rate.
7 measure. 8

 We can use the yield of similar-risk, newly issued corporate

 We can use a model that estimates the cost of debt from data about the
securities. If a
company is rated A and has mostly medium-term debt, then we firm ’s bond prices, the estimated probabilities of default, and the
can use the average yield on medium-term, A-rated debt as the estimated payoffs to bondholders in case of default.
firm’s cost of debt.
 Note that this method is somewhat problematic because the This method requires a lot of work and is mathematically non-trivial; we
yield on a bond is its promised return, whereas the cost of debt is postpone its discussion until Chapter 28. For cost of capital calculations it
the expected return on a firm’s debt. would be used in practice only if the firm we are analyzing has significant
 Since there is usually a risk of default, the promised return is
amounts of risky debt. The first two methods above are relatively easy to
generally higher than the expected return. Nevertheless, despite
the problematics, this method is often a good compromise. apply, and in many cases the problems or errors which are encountered in
these methods are not critical.
9 10

As a matter of theory, however, both of these methods fail to make proper There are several aspects of our calculations worth noting:

risk-adjustments for the cost of the firm’s debt.

The third method, which involves computing the expected return on a firm’s
debt, is more in line with standard financial theory, but it is also more difficult
to apply. It may not, therefore, be worth the effort. In the remainder of this
section, we apply the first two of these methods to calculate the cost of debt
for U.S. Steel and Merck.

Method 1: U.S. Steel ’s Average Cost of Debt • When calculating the average cost of debt r D from the financial statements, it is important
For U.S. Steel we compute the average cost of debt: to include all financial debt, without distinguishing between short-term and long-term items.

11 12

2
 9/8/2022

We treat liquid assets such as cash and cash equivalents as negative debt and  This is because we believe that historical costs of debt have little predictive
subtract them from the firm’s debt. The idea here is that the firm could use its power for future costs.
cash to pay off part of its debt, so that the effective debt financing of the firm is 3.6 Two Approaches to Computing the Firm ’s Cost of Equity, rE
its financial debt minus cash. However, the implementation of this particular  The equation for the weighted average cost of capital is WACC = E
piece of theory is largely a judgment call—we may not want to attribute all
/( E + D )* r E+ E /( E + D )* r D*(1 − T C ). Thus far in this chapter
we have discussed the estimation of four of the five parameters of
cash to the possibility of paying off debt, and we may want to compute the firm the WACC equation: E , D , T C, r D . We now come to the most
’s cost of borrowing as opposed to the interest it earns on cash. Were we to use problematic of the computations related to the WACC
the average cost of debt for U.S. Steel as a prediction of its future cost of debt r parameters—the computation of the cost of equity r E. There are
two approaches
D, we would most likely use the current cost rD = 4.90% in the WACC to r E that can readily be computed:
computation. 13 14

The Gordon dividend model computes rE based on current dividend Div0,

current stock price P0, and the anticipated growth of future dividends g : 3.7 Implementing the Gordon Model for rE
The Gordon dividend model derives the cost of equity from the following
 The capital asset pricing model (CAPM) computes rE based on the risk-free rate rf, the deceptively simple statement: The value of a share is the present value of
expected return on the market E ( rM), and a firm-specific risk measure β:
the future anticipated dividend stream from the share, where the future
anticipated dividends are discounted at the appropriate risk- adjusted cost
Where
rf = the market risk-free rate of interest
of equity rE .
E ( r M) = the expected return on the market portfolio
The simplest application of the Gordon model is the case where the
anticipated future growth rate of dividends is constant.
Each model has its variations and problems, which are discussed (adnauseam ?) in the next
two sections. 15 16

Thus—given a constant anticipated dividend growth rate, we derive the

Suppose that the current stock price is P0 , the current dividend is Div0 , and
Gordon model cost of equity:
the anticipated growth rate of future dividends is g. The Gordon model
states that the stock price equals the discounted (at the appropriate cost of Solving the above equation for rE gives the Gordon formula for the cost of

equity rE ) future dividends: equity:

Note the proviso at the end of this formula: In order for the infinite sum on the
first line of the formula to have a finite solution, the growth rates of the
Provided that | g | < r E , the expression can be reduced to
dividends must be less than the discount rate. In our discussion of the Gordon

model with supernormal growth rates (see below) we return to the case where
(we will spare you this derivation, which is based on a formula for geometric
this is not true. To apply this formula, consider a firm whose current dividend is
series usually studied in high school).
Div0 = $3 per share, whose share price is P0 = $50. Suppose the dividend is18
17

3
 9/8/2022

Anticipated to grow by 12% per year. Then the firm ’s cost of equity rE is 17.6%: It is also the most widely used cost of equity model, the reasons being both
its theoretical elegance and its implementation simplicity. The CAPM
derives the firm’s cost of capital from its covariance with the market return.
The classic CAPM formula for the firm ’s cost of equity is

Where rf =the market risk-free rate of interest

E ( r M) = the expected return on the market portfolio
3.8 The CAPM: Computing the Beta,β

 The capital asset pricing model (CAPM) is the only viable alternative to the
 In the remainder of this section we focus on measuring the firm’s β;the
Gordon model for calculating the cost of capital.
next section shows how to apply the CAPM to find the firm’s cost of equity
19 rE . 20

3.9 Using the Security Market Line (SML) to

• rf equal to the risk-free interest rate in the economy (for example, the yield on
Calculate Merck ’s Cost of Equity, rE
• In the capital asset pricing model, the security market line (SML) is used to Treasury bills). We leave the question of whether to use the short-term or long-
calculate the risk-adjusted cost of capital. In this section we consider two
SML formulations. term rate open until section 3.11. For the moment, for illustrative purposes, we
• The difference between these two methods has to do with the way taxes
are incorporated into the cost of capital equation. use r f = 2%.
Method 1: The Classic SML • E ( r M ) equal to the historic average of the market return, defined as the
The classic CAPM formula uses a security market line (SML) equation that
ignores taxes: average return of a broad-based market portfolio. There is an alternative
approach based on market multiples; both of these are discussed below. For the
• Here rf is the risk-free rate of return in the economy and E ( RM) is the
expected rate of return on the market. The choice of values for the SML current section, we use E ( rM ) = 8%. The following spreadsheet illustrates the
parameters is often problematic. A common approach is to choose:
classic CAPM cost of equity computation for Merck ’s cost of equity:
21 22

• Note that the Tax-adjusted cost of equity has a lower intercept and a higher
slope than the classic CAPM:

• The intercept is r f(1 − T C ) instead of r f . This intercept is lower than the r f

 Method 2: The Tax-Adjusted SML intercept of the classic CAPM.• The slope is E ( rM) − rf(1 − TC) instead of E (
The classic CAPM approach makes no allowance for taxation. Benninga-Sarig rM ) − rf. This slope can be written as the classic CAPM slope plus TCrf: E ( r
(1997) show that the SML has to be adjusted for the marginal corporate tax rate M) − r f (1 − TC) = [ E ( rM) − rf] + TCrf Another way to write the tax-adjusted
in the economy. Denoting the corporate tax rate by TC , the tax-adjusted SML is cost of equity is:

This formula can be applied by substituting rf(1 − TC) for rf in the classic CAPM.
23 24

4
 9/8/2022

• This rewriting makes clear that the difference between the classic rE and the • Although the tax-adjusted CAPM is more consistent with an economy with
tax-adjusted r E is a function of the corporate tax rate TC , the risk-free rate rf , taxation, we confess that given the uncertainties surrounding cost of capital
and the equity beta β.14 For Merck, the tax-adjusted approach gives a computations the difference between the classic CAPM and the tax-adjusted
somewhat higher cost of equity: CAPM may not be worth the trouble.

3.10 Three Approaches to Computing the Expected Return on the Market,

E( rM)
 There are three major approaches to computing E ( rM):
• The historical return on a major market index
• The historical market risk premium on the market index
• The Gordon model
25 26

 All three approaches are illustrated in this section, and their effect on
computing Merck ’s cost of equity is illustrated at the end of this section.

1. E ( rM) as the Historical Average Return on a Market Portfolio

• A simple approach to computing E ( rM) is to take it as the average
of the historical returns of a major market index.
• In the computation below we illustrate this approach by using
Vanguard ’s 500 Index Fund as a proxy for the market.
• The annualized return on this fund since 1987 is 8.27%. We can
take this as a reliable proxy for the historical annual average
return from holding the S&P 500:

27 28

2. Computing the Market Risk Premium E ( rM) − rf

Directly
• We can also compute the market risk premium directly. This requires a bit
more work: In the spreadsheet below we show the monthly returns on the
S&P 500 and monthly interest paid on U.S. Treasury bills. The average
annualized risk premium on the S&P 500 is 4.40%.

29 30

5
 9/8/2022

• Applying the risk premium directly to the computation of Merck’s cost of 3. Calculating the Expected Return on the Market
equity gives a cost of equity r E close to 5% (note that we still haven’t settled Using the Gordon Model
the question of r f): • Setting E ( rM) = 4.40% approximates the historic market return in the United
States for 1987–2012. Historic averages are appropriate if we think that the
future anticipated rates of return will correspond to the historic average.
• On the other hand, we may want to take current market data to calculate
directly the future anticipated market yield. We can do this computation by
using the Gordon model. Recall from section 3.6 that the model says that the
cost of equity rE is given by:

• This formula also applies to the market portfolio, so that we can write:

31 32

• interpreting Div0, P0, and g to be the current dividend, price, and growth rate
of the market portfolio. Assume that the firm pays out a constant proportion a
of its earnings as dividends; then, indicating by EPS0 the current earnings per
share, Div0 = a*EPS0. Interpreting g to be the earnings growth of the firm, we
can write:

• The term on the right-hand side of this equation, P 0/EPS 0, is the price
earnings ratio of the market. We can use this formula to compute E ( r M),
and thus tie the cost of equity to currently observable market parameters.
Here is an implementation: 33 34

The End

35