1

Stern School of Business

New York University

Corporation Finance (B40.2302) Tony Marciano

Lecture 3: Capital Budgeting under Uncertainty

Intro. How does the DCF approach change in this world:

1. Use expected cash flows
2. Use risk-adjusted discount rate
A. Risk/Return Trade-off
B. Using an Asset-Pricing Model (CAPM,...)
B1. Risk Premium portion (risk)
B2. Risk-free portion (timing)
3. Putting it together with an example
4. Computing Betas to be used for APMs

Additional Issues:
5. Personal Investments
6. Embedded Options
A. The Value of Options at a High Level (More later)
B. Decision Trees (*)
C. Option Adjusted Spread (OAS) (*)
7. P/E Multiples

1
 2

Introduction to Uncertainty

So far in this course, we have only concerned ourselves with a world of perfect certainty -- ie, all
cash flows could be projected perfectly. We did this to demonstrate some of the important
concepts regarding NPV analysis -- namely, what cash flows count, how to obtain the discount
rate with the appropriate timing, and how to perform the mathematical operation.
But this is obviously not a realistic assumption -- we live in a world where we can only predict a
range of possible outcomes. Welcome to the world of uncertainty. In this world, we need to
readdress how to obtain the correct cash flows and obtain the proper discount rate; the
mathematical operations, however, remain the same from lecture 2 (except when we get to
options).

1 Use Expected Cash Flows

The only significant change in the determination of cash flows is that we will plug in the
expected cash flow.
NPV = Σ (Expected CFt)/(1+r)t - Investment

Ex: An Investment costs $1000. You know that it will produce a cash flow in two years of either
$2100 or $100 -- each possibility equally likely. You know the appropriate annual discount rate
is 8%. Is this a good investment?
A: Expected Cash flow = .5*2100+.5*100=1100.
NPV = -1000+1100/1.08^2 = -57. So reject project.

In the above case, the discount rate was given. But in the real world, this is not the case.
Determining the proper discount rate to use in the NPV calculation is a big part of the game
when performing DCF analysis under uncertainty.

2
 3

2 Adjusting the discount rate for risk

2.1 Risk/Return Tradeoff

At this point, we will add the second of the four major assumptions of the course: Investors
prefer a dollar guaranteed over a dollar in expectation. This aspect of human behavior is called
risk aversion. (Remember that this is due to the diminishing returns of wealth). What it means to
us is that the riskier an investment is, the higher the expected return must be in order to attract
these risk-averse investors. Therefore, the manager of a firm must use a higher discount rate for
projects of greater risk since investors demand more return to compensate them for this added
risk.

The following table shows how risk and return relate for four types of securities:

Average US Annual Returns (1926-1992)

Avg Return Risk Premia Std Dev

Stocks (Com) 12.4 8.6 20.5
Corp Bonds 5.8 2.0 8.4
Govt Bonds 5.2 1.4 8.5
Treas Bills 3.8 0.0 3.3

The risk premium is the additional return above the riskless rate (here the T-bill) required for a
particular investment based on the additional risk associated with it. In this table, the market
seems to demand over 8% in premium to go from short t-bills to stocks and over 7% in premium
for stocks over bonds during this time period. These premia seemed to have held up for many
years but recent data shows a decrease in the risk premium of at least a point – so that many
assume a market risk premium of stocks over bonds of about 6%.

In this course, we will assume that returns from historical events are expected to have a 7-8%
premium but current premia are 6-7%.

3
 4

2.2 Using an APM to determine the discount rate

So, when choosing a discount rate for a risky project in our NPV analysis, we need two things:
1) How do we measure the risk of the project
2) How do we numerically translate the risk into a required return number.

These are the jobs of a good Asset Pricing Model (APM). Is there an APM which perfectly
performs these duties? We don't know. So, does that mean we should forget about the whole
thing and just go on to capital structure? NO! The models we have are better than using nothing.
They all capture the kind of risk we care about -- namely, market-wide forces -- and give a
reasonable conversion of risk to return. Even if they are not perfect, they force managers to take
into consideration the important issues.

We will not derive any APMs in this course. Instead, we will merely apply them as formulas for
translating risk into return for our problems. We begin with the most famous one -- CAPM.

2.2.1 Using CAPM for Capital Budgeting

From Investments, you learned that the following formula translates risk into return:

Ri = Rf + βi * (Rm-Rf)
where Ri is the required return on the investment, Rf is the riskless rate, βi is the beta of the
investment representing both the volatility of the investment and its correlation to the market,
and Rm-Rf is the risk premium for the market. For our purposes, we can rewrite this equation as:

Required Return = Riskless Rate + β * (Risk Premium)

where β is the measure of risk inherent in the investment. A β of 0 is a riskless project (and
thus Rf is the proper discount rate) and a β of 1 is equally risky as the entire market (and thus
Rm, the market return, is the proper discount rate).

The exact definition of β=COV(i,m)/VAR(m)

where COV(i,m) = ρim* σi * σm

In more intuitive terms, β measures the investment’s sensitivity to the market. E.g., a β of 1
means the investments moves in synch with the market, a β of 0.5 means the investment moves
only half as much as the market and a β of 2 means the investment moves twice as much as the
market, on average. So, an investment with a β of 2 is expected to move twice as much as the
market does at a given time.

NOTE: From the definition of β, we see that an investment which is very volatile but
independent from changes in the market economy -- ie, a correlation of zero to the market -- has

4
 5

a β of 0 and therefore is discounted at the riskless rate. This risk is said to be firm-specific or
idiosyncratic and is ignored for the purposes of public corporations where investors can diversify
it away. The only risk that matters for a project which is part of a larger portfolio is its market-
related or undiversifiable or systematic risk.

So, CAPM answers the two questions:

1) The appropriate measure of risk is β
2) The way risk gets translated into required return is by simply multiplying β by the risk
premium and adding this to the no-risk required return.

Example 1: You have a project which costs $1000 and generates an expected cash flow in one
year of $1150. You determine its β to be 1.5. The one year risk-free rate is 3.5%. Assume from
above that the market risk premium is 8.6%. (Always assume this unless you are explicitly told
otherwise. Remember that the risk premium is the difference between the market return and the
risk-free rate.) Should you undertake this project?
A: NPV = -1000+1150/(1+Ri)
Ri = 3.5% + 1.5 * 8.6% = 16.4%
NPV = -12. So reject the project.
(Note that return numbers used are for the future not the past. Whatever has happened last year
is irrelevant in determining proper discount rates. For example, if the market was flat last year,
you do not use a risk premium of 0.)

Example 2: You have a project which costs $1000 and generates an expected cash flow in one
year of $1150. It has a volatility of 30% (much larger than the one above), but is uncorrelated to
the market. Should you take it?
A: NPV = -1000+1150/(1+Ri)
Ri = 3.5% + 0 * 8.6% = 3.5%
NPV = 111. So take it.

5
 6

2.2.2 Other Asset Pricing Models (*)

As you may know, the CAPM is not the only model for pricing risk. The APT (Arbitrage
Pricing Theory) and ICAPM (Intertemporal CAPM) also supply formulas as well a others.
Additionally on the empirical side, Fama-French have related expected returns to size and book-
to-market factors as well as market return.
Whichever you use, you will use consistently with how you would use it for any investment --
ie, multiply your investment's sensitivity to each factor in the model by the factor risk premium
and sum over each factor.
Nonetheless, the default in this class is CAPM due to its pervasiveness.
Behavioral Finance issues have become increasingly emphasized in recent times, but we will
hold off on addressing these issues until we fully understand the models based on efficient
markets.

2.2.3 Longer Term Cash flows

Unfortunately, the CAPM equation is a short term model whereas most firm projects are long
term. How can we adjust the CAPM model to deliver longer term discount rates. Longer term
projects are surely riskier than shorter term projects, but this is captured by the compounding of
the increased required return over multiple periods. The main problem revolves around the
timing of the risk-free rate. In a short-term one-period problem, using the T-bill rate as the risk-
free does quite well. However, for longer maturities, we would like to determine the expected
risk-free rate over the entire period. So, it makes to use the same maturity Treasury bond yield
since this captures the shape of the yield curve and the market's prediction of the trend of the
risk-free rate. If you do this, however, you should subtract the risk premium of Treasury bonds,
which we showed to be 1.4% historically. You will be responsible only for applying this concept
not deriving it.

So, the recipe for choosing the risk-adjusted discount rate via CAPM can take one of two forms:
The preferred method is:
• If the risky cash flow comes within one year, use the T-bill rate and plug into CAPM.
• If it arrives later, use the following formula:
R = T-bond rate of same maturity + β * (Market Risk Premium over This Bond)

[AS AN ASIDE:
The alternative described from the text footnote is (only different for the multi-year scenario):
• If the risky cash flow comes within one year, use the T-bill rate and plug into CAPM.
• If it arrives later, use the following formula:
R = (T-bond rate of same maturity – 1) + β * (Market Risk Premium over Short T-Bill)]

6
 7

2.2.4 Multiple Cash Flows

Generally, investments have cash flows coming in different periods. To the extent that these cash
flows are heterogeneous in the sense that they have different risk characteristics, you should
treat each cash flow independently and use the appropriate discount rate for each type. If the
cash flows are spotty -- ie, coming in very spaced out intervals -- you should use a different rate
for each time period. Alternatively, you can calculate a project-wide discount rate and achieve
the same answer. The latter approach is recommended where the project's cash flows arrive
each period and have reasonably homogeneous risk -- which is true under most circumstances. If
you are given a project-wide risk-adjusted time-adjusted discount rate, then it is reasonable to
apply it to all the cash flows of the project.

3 Putting it together with an example

Overall Recipe
I. Cash Flows
A. Determine all relevant cash flows (size and timing)
How does each characteristic of the project effect my economic wealth and when?
B. Take average of each cash flow type for each time period

II. Discount Rate

A. Get timing right for the rate to match each cash flow
B. Adjust the rate for the risk of the cash flow using APM

III. NPV Calculation

A. If NPV>0, do it; else, don't!
B. Value the embedded options

7
 8

EXAMPLE:
An investor is determining whether to rent for a year the Grape Poupon vineyard. Assume the
land is rented in Year 0 and the rent is paid up-front, the wine is produced in Year 1, and then
sold at the end of Year 5. The vineyard is 100 acres and the owner asks $2000 per acre. The
vineyard is expected to yield enough grapes to produce 20,000 bottles of wine. Harvesting costs
in Year 1 are $1000 per acre. Production costs in Year 1 are $3/bottle. Both costs are paid at the
end of year 1. The wine is held in inventory (assume no cost) for four years and is sold in Year
5. In five years the price of this type of wine will be correlated with the state of the economy.
There is a 25% chance the wine can be sold for $20/bottle, a 50% chance it can be sold for
$25/bottle, and a 25% chance that the price in five years will be $50. The  of the price of wine
is 0.6. The 3-month T-bill is 3.5%, the one-year Treasury note rate is 5%, and the 5-year T-bond
has a rate of 7.9%. The market price for risk is 7% over T-bonds. Would you rent the vineyard
and produce the wine?

Answer:
First we must determine the cash flows.
Time 0: Pays rent of 2000*100 = 200,000
Year 1: Pays harvesting costs of 1000/acres * 100 = 100,000
Pays production costs of 20000 bottles * 3 = 60,000
Year 5: Earns 20000*(.25*20+.5*25+.25*50) = 600,000

Now, let's determine the discount rates:

The Year 1 cash flow is certain, so we could really use the one-year riskless rate of 5%, but
generally people just lump all the cash flows together and use one discount rate for all.
The Year 5 cash flow, on the other hand, is risky and correlated to the market with a β=.6.
As reflected in section 2.2.3, there are two acceptable ways to derive the discount rate. I will use
the simpler method:
R = Treasury Bond + β * (Market Risk Premium over same Treasury Bond)
= 7.9 + 0.6 * 7 = 12.1%

Doing the operation gives negative NPV either way, so don’t do it.

NPV (Simplified and Common Method) = -200000 - 160000/(1.121) + 600000/(1.121) 5 = -3790

so don’t do it.

NPV (Academic Method) = -200000 - 160000/(1.05) + 600000/(1.121) 5 = -13440, also negative

so still no.

8
 9

4 Computing Betas

So far, we have assumed that you knew the beta of a project. This is very rarely given to you
(except for occasional exam and homework problems). In this section, we will discuss how to
obtain the proper beta to be used in the CAPM.

NOTE: Sometimes a time-adjusted, risk-adjusted cost of capital is given to you. If so, this
number is usable as long as it is the risk-adjusted discount rate that your investors will use when
doing DCF analysis.

This is a question of practicality. There are three ways to compute beta:

1) Calculate it statistically. (We will not address this and I do not recommend it since it is
difficult and quite imprecise.)
2) Deduce it. For example, if the risk is purely idiosyncratic, use a beta of 0.
3) Look it up. However, this approach requires some adjustments. The method for performing
these adjustments is described below.

There are sources available for looking up the beta for a firm.
A "beta book" is supplied by Merrill Lynch and Value Line. Also, the Center for Research in
Security Prices (CRSP) at the University of Chicago calculates an equity beta based on the firm's
historical sensitivity to the market. You can simply look up your firm in one of these books and
pull out the beta of the firm. But there are two problems with this.
1) These books only list total firm betas -- not project betas. Therefore, you want to use a beta
that is most similar to your project at hand as opposed to an entire firm which may be in several
businesses. For example, if your firm is a conglomerate and is interested in expanding in the oil
industry, using your whole firm beta will not give an accurate beta since your firm beta is an
average of each of your project's betas.
2) The correct beta to use is the asset beta -- not an equity beta. That is, we care about the risk
of the project independent of how levered our position in it is. (Remember that at this point in
the course, we are ignoring capital structure.)
We will address each of these problems in order.

4.1 Finding the best betas (The "identical twin" method)

(In this section, we will assume that all firms are unlevered -- that assets and equity are the same
thing -- in order to put off these issues to the next part.)
In order to solve the first problem, you should begin by describing the business that this project
is in. Then, find publicly traded firms whose income comes predominantly from this business.
These will be your "identical twins" or comparable firms ("comps"). Then, look up the betas in a
"beta book" or CRSP or both for these firms. Taking the average of these would then give you a
ballpark figure for the beta of this investment. As you can see, this methodology is as much art
as science.

9
 10

4.2 Converting equity betas into asset betas

One major problem with the above approach is that the beta figures we've obtained for the
"comps" are equity betas whereas our formula requires an asset beta. We need a mathematical
transformation for computing asset betas from equity betas.

Fortunately, the CAPM tells us that the beta of a combined investment is just the weighted
average of the individual betas (where the beta is weighed by market value). Since the assets of a
firm is just the combination of its debt and equity, we can use the following formula:

βA = D/(D+E) βD + E/(D+E) βE

where D and E are the market values of Debt and Equity:

E = Price/share * # shares
D = Market Value of outstanding debt

So, you can see that the beta of equity is not the same as the beta of assets -- which is what we
want. The relationship between the two becomes clearer when rearranging the above expression
as:
βE = βA + D/E (βA-βD)
= (Business Risk) + (Financial Risk)

So, in the case of an unlevered firm (D=0), the equity beta equals the asset beta and perfectly
captures the business risk of the project. However, levered firms (D>0) will have higher betas
for their equity than what is appropriate to perfectly capture the risk of the project. In these
cases, we want to use the equation above to derive the asset beta. This approach is called
"Unlevering Beta".

Since we want the asset beta but only have the equity beta, we need to plug into the equation
above the appropriate values. D and E seem easy enough to get, but the debt beta is not trivially
obtained.

For firms with very little chance of default, an appropriate debt beta is zero. And for the example
we will do below, this is precisely what we are going to assume (as well as for the American
Chemical case). In these cases, the above equation reduces to:
βA = βE E/(D+E)

10
 11

In the cases where a firm has risky debt, however, there are basically two methods for obtaining
a debt beta.
1) Back the beta out using the CAPM equation by plugging in the expected return on the bonds.
The problem with this method is that we observe the "promised return" or yield not the expected
return which captures the possibility of default. That is, we want to know the expected return
which uses expected cash flows -- not the yield on the security which uses the maximum cash
flows that the debtholders will receive. To the extent that default probabilities are difficult to
determine, this method becomes quite imprecise.

2) Another equally imprecise but much simpler approach is to use the historical betas associated
with the debt's bond rating. That is, bonds of equal bond rating should, in the mind of the rating
agency responsible, should have similar default probabilities. So, given the bond rating, just look
up the appropriate debt beta from the table below.

Betas for various classes of debt

Type of Bond Beta

Govt (1-5 yrs) 0.08
Govt (6-10 yrs) 0.13
Corporate (Aaa) 0.20
Corp (Aa) 0.20
Corp (A) 0.21
Corp (Baa) 0.23
Lower Grade 0.31

11
 12

EXAMPLE:
(For the purposes of this example we will assume that debt will not default and that the
appropriate debt beta is 0.)

The magazine National Review is considering a project aligned with its current business. You
are the CFO and are given the expected cash flows projected over time if the project is
undertaken. In order to determine whether the project should be accepted, you need to know the
proper discount rate to use. In turn, you need to know what is a suitable project beta to plug into
the CAPM formula. You remember what you learned in business school and proceed as follows.

The National Review is a bi-weekly magazine that promotes conservative opinion. Since the
magazine is private, obtaining the project beta requires the use of data from other firms. The
best-case scenario is to use a portfolio of publicly-traded firms that focus solely on delivering
intellectual opinion via the magazine outlet. These comparable firms could be considered twins
of National Review. The problem is that such comparables do not exist. So, one must use
broader sources. One excellent source is Value Line Investment Surveys. Their publishing
industry contains 17 firms which include publishers of books, magazines and newspapers, as
well as printing firms and so forth. However, there are only five firms with significant magazine
operations (at least 20% of sales from magazine operations). Table 1 lists these five publicly-
traded firms. All five firms have widely different subscribers. Moreover, each firm has business
units in addition to its magazine operations. The common characteristics across these firms,
however, is their magazine publications. Though the number of firms here is small, other
operations of each firm tend to cancel one another out such that the overall portfolio resembles
that of the magazine industry.

Next, you obtain the listed betas for these 5 "comps". You pull off the betas from Value line as
well as from CRSP. These figures are reported in Table 2. The CRSP and Value line betas are
comfortably close to one another. (Also, notice that the betas are around 1. What does that
mean?)

Next, you know that these numbers are equity betas while you want asset betas. Remembering
the equation above, you obtain the market values of debt and equity for each of the five firms
and choose as the equity beta the average of the Value Line and CRSP numbers. (See Table 3).
You plug these into the "Unlevering Beta" calculation of E/(D+E)* βE giving the asset beta
column in Table 3. Taking a simple average of these five figures gives the resulting asset beta of
0.89.

Now, you can go on to obtain the project discount rate and consequently evaluate the project.

NOTE: You should use the risk measure associated with the project you are investing in -- not
necessarily the risk measure for your firm.
12
 13

TABLE 1

Description of Comparable Firms

Firm Description
------------------------------ -------------------------------------------------------------------------------

McGraw-Hill Publisher of books and magazines (e.g., Business Week) and

markets information for finance and consulting practices.
Magazine division accounts for more than 20 % of sales.

Meredith Publisher of several magazines and books. Also owns television

stations and a residential real estate franchise network. Magazine
group accounts for 80% of sales.

Playboy Enterprises Primary business is publication of Playboy magazine. Also

operates an entertainment group and product and marketing event
group.

Reader’s Digest Assoc. Publisher of magazine and books and markets home entertainment
products. Reader’s Digest accounts for more than 25% of sales.

Time-Warner Publishes magazines (Time, People, Fortune, etc.) and books,

produces movies, owns cable TV systems and pay TV-networks,
and records music. Magazine division accounts for more than
20% of sales.

Source: Value Line Investment Survey (December 3, 1993). McGraw-Hill, Meredith, Playboy
Enterprises and Reader’s Digest Association are in Value Line’s publishing industry; Time-
Warner is in the entertainment industry.

13
 14

TABLE 2

Betas for Comparable Firms to National Review

Firm Value Line Beta CRSP Beta

------------------------------ ------------------------ ---------------------

McGraw Hill 1.00 0.97

Meredith 1.05 1.19

Playboy 0.70 0.91

Reader’s Digest Assoc. 1.10 0.98

Time-Warner 1.30 1.56

Average Beta 1.03 1.12

Median Beta 1.05 0.98

Notes: Value Line betas come from the December 3, 1993 issue. The CRSP betas are based on
a regression analysis of each stock’s monthly return on the value-weighted index of all NYSE,
AMEX, and NASDAQ stocks. The CRSP betas are calculated using the 60 months of stock
return history over the 1988-1992 period. Stock price data for Reader’s Digest Assoc. is
available for only 38 months during this period.

14
 15

TABLE 3

Asset Betas for Comparable Firms to National Review

Firm Equity Beta Debt Equity Asset Beta

--------------------------- -------------- -------- ----------- -------------

McGraw Hill 0.99 892 3,405 0.79

Meredith 1.12 136 591 0.91

Playboy 0.81 1 196 0.80

Reader’s Digest Assoc. 1.04 8 4,906 1.04

Time-Warner 1.43 8,980 16,579 0.93

Average Asset Beta 0.89

Median Asset Beta 0.91

Notes: Equity beta for each firm is the average of the Value Line and CRSP betas from Table 7.
Debt and equity figures come from Standard & Poor’s Compustat service. Asset betas are
calculated using the asset beta equation described in the text.

15
 16

There are certain situations where deviation from this Valuation approach is accepted.

5 Personal Investments
Our APM approach only applies to cases where the cash flows of the project can be combined in
a large portfolio of well-diversified investments. In the case of private projects which can not be
placed in a diversified portfolio, the investor must use a personal discount rate which captures
her/his distaste for risk. Here, introspection is required and the Fisher separation theorem no
longer holds. But fortunately, for most corporations, the above methodology works well.

6 Options (next section for details)

The Economist
January 8, 1994

HEADLINE: Optional investing

HIGHLIGHT:
Can the theory of how to price financial options improve the way bosses invest
their companies' money?

BODY:
TALK to a company boss about financial options and he might smile at the
prospect of exercising his generously priced share-options. Or shudder, maybe,
at the reminder of the firm's incomprehensible hedging strategy. He will
almost certainly not think about his recent decision to build a new widget
factory, or to postpone the launch of the exciting new fax-cum-roller-blade
that marketing says customers are baying for.

He should, however. For when he took those decisions, he probably obeyed the
rules of financial-options pricing theory. Or so says a new book * by two
economists, Avinash Dixit of Princeton University and Robert Pindyck of the
Massachusetts Institute of Technology's Sloan School. In it, they attempt to
integrate options theory and the traditional model of how firms decide where
to direct their investment. Their merged theory, they say, better explains
what really goes on inside firms, and clears up some real-world riddles the
old approach left unsolved.

* "Investment under Uncertainty", Princeton University Press, January 1994

At business schools, managers learn to calculate the "net present value" (NPV)
of a mooted investment. They forecast future profits, then discount them using
a "discount rate" to reflect the higher value that money has today than in the
future (usually the interest rate on government bonds plus a bit to offset the
riskiness of the investment). That gives the present value of future profits.
If this number is bigger than the present value of the costs of investing in
the project -- that is, the NPV is positive -- then go ahead, says orthodox
theory. Otherwise, think again.

All investment calculations rely on predicting uncertain future profits. But

the traditional theory also assumes, implicitly, that investments are a now-
or-never choice. That is unrealistic, say Messrs Dixit and Pindyck. Mostly,
managers have some choice about when to invest. Waiting may mean missed

16
 17
opportunities. More often, in an uncertain world, it offers a valuable chance
to learn more about the likely fate of the project.

The ability to delay an irreversible investment is like a financial "call

option", say the two economists. The firm has the right, but not the
obligation, to buy (invest in) an underlying asset (the profits from the
project) at a price (the investment cost) at a future time of its choosing.
This option has a value. When the firm makes the investment it exercises (or,
in financial-market jargon, "kills") its option.

It follows, then, that the cost of that killed option (the value of waiting
for better information) ought to be included when calculating NPV. Before a
project goes ahead, the present value of profits should exceed the investment
costs by at least the value of keeping the option alive.

In the money

The price of a financial option is extremely sensitive to uncertainty about

the value of the asset it gives the right to buy. So, too, is the value of
what the authors call firms' "real options" -- that is, untapped investment
opportunities. Uncertainty about input costs, interest rates, taxes and prices
pushes up the real option "price" and increases the advantages of the firm of
waiting.

It is profitable to kill a financial option as soon as the underlying asset

value exceeds the option price plus investment cost -- that is, when the
option is what is known as "in the money". But financial-market traders mostly
hold on to their options in the hope that they will yield even bigger profits
by becoming "deep" in the money. The same logic applies to firms' real
options.

Orthodox investment theory suggests that firms should invest in a project as

soon as its NPV is positive. These two insights from option-pricing theory
suggest something quite different: that it will often make sense for firms to
invest in a project only when its NPV is very large.

In fact, that is what usually happens in practice. Most bosses do their basic
NPV sums, but add in a margin to help them feel comfortable. For instance,
they may discount predicted profits using a "hurdle" required rate of return
that is often two or three times the standard discount rate. Applying options-
pricing theory to a big sample of typical business projects, the two
economists found that such hurdle rates are perfectly sensible. However, the
new "real options" theory allows bosses to set them on a more rational basis
than gut instinct.

The theory also explains why firms often respond slowly to changes in tax,
interest rates or demand -- all of which, orthodox theory suggests, should
elicit an instant response. Bosses are more interested in changes in
uncertainty. Consider, for illustration, a government that cuts interest rates
and taxes to boost investment, but as a result sows the seeds of longer-term
economic troubles. It may actually increase uncertainty and cause firms to
invest less.

The theory will be most useful in areas where the uncertainties are relatively
apparent. The sums are already being churned over in the computers of some oil
firms, for instance. There is one main risk in developing an oil field -- a

17
 18
change in the oil price. Property developers, electric utilities and drugs
companies are also investigating how to apply the theory in practice. It may
work less well in, say, judging whether to invest in finding a cure for
cancer, or in evaluating strategic decisions where long-term goals matter more
than financial results.

Another problem is the theory's complexity: it has taken decades to persuade

managers to use even the simple disciplines of the traditional NPV model. And,
while the price of financial options is obvious, there will always be debate
about the price of real options, which will be based on managers' assumptions
about the future, not on market supply and demand.

Even so, the theory is likely to prove a big advance on the old NPV method.
Though its aim is to improve company decision-making by showing how bosses can
learn from financial markets, it can also help people in financial markets to
understand firms better. Projects that a firm has not yet invested in may be
at least as valuable to it as the ones that are going ahead -- especially in a
volatile market. Might bosses boost their firms' share price by telling
analysts about all the lucrative things they are not doing?

18
 19
Heard on the Street

American Greetings Stock Gets Warm `Hello'

From Investment Pros Who Cite Mass Market
By John R. Dorfman and Wendy Bounds

06/01/1994
The Wall Street Journal
PAGE C2

American Greetings scored a coup when Bill Clinton's White House

placed an order for 275,000 Christmas cards in 1993. During the
previous 12 holiday seasons, the Reagan and Bush administrations
had relied on arch-rival Hallmark Cards.

But it's not high-brow business that has fans excited about
American Greetings stock. It's the grass-roots popularity of
American Greetings cards at places like Wal-Mart and Walgreen.
That's where the growth is, says Frederic E. Russell, a Tulsa,
Okla., money manager who owns more than 50,000 shares.

"In 1980, mass merchants or discounters had 45% of the

greeting-card market in the United States. Today, they have 62%,"
Mr. Russell says. " American Greetings , as the leading
distributor to mass merchants, is well-positioned to exploit this
trend."

Brokerage houses seem to agree. Of nine firms that follow

American Greetings stock, eight rate it a "buy," and six of them
call it a "strong buy," according to Zacks Investment Research.
Yet unlike many brokerage-house favorites, American Greetings
doesn't carry a sky-high stock price.

At yesterday's price of 28 1/8, American Greetings shares sell

for 16 times the past four quarters' earnings. That's about in
line with the stock market's average price/earnings ratio these
days.

Yet the company's growth appears better than average. Wall Street
looks for earnings to rise to about $2.02 a share in the current
fiscal year, which ends in February 1995, from $1.54 in the past
fiscal year. Next year, analysts expect earnings of $2.26 a
share.

Mr. Russell says the Big Three of the card industry -- Hallmark,
American Greetings and Gibson Greetings -- control more than 90%
of the market. All three are doing well, as time-pressed
Americans turn to cards to get their messages across.

Hallmark is closely held, though rumors abound that it might make

a public offering soon. American Greetings and Gibson are
publicly traded. Mr. Russell favors American Greetings because of
its strong balance sheet (debt was recently only 8% of capital)
and its strength in the discount end of the market.

The stock is down about 18% from the peak of 34 1/4 it reached in

19
 20
December. One reason: Doubts about the profit potential of the
company's new CreataCard kiosks.

American Greetings has sprinkled the country with some 8,000 such
kiosks, where customers can prepare customized cards for friends
or loved ones. "They've preempted everyone by being the first out
there," says Jeffrey S. Stein, analyst with McDonald & Co., who
speaks for the bulls. But bears argue that the kiosks don't seem
to be catching on enough to boost profits much.

"There's concern about CreataCard having only eight cards a day

at certain outlets, rather than 10," says Leo Murphy, analyst
with Pioneering Management in Boston. "That's really dancing
around the fringes."

Even if the kiosks were to bomb, Mr. Murphy thinks the company
can show solid, steady 12% earnings growth. If the kiosks are a
success, he sees growth at 14% to 15%. The stock in his opinion
is worth at least 34, and as much as 40 if CreataCard works out
well.

"The company keeps punching out 15% earnings gains," says Edward
Cimilucca, director of research at J.W. Seligman, a New York
money-management firm. The greeting-card field, he says, is "a
very nice business," and CreataCard is "an authentic new product"
that should contribute to profits before long.

Mr. Cimilucca likes the way American Greetings has reduced debt
and improved its profit margins. Debt is below $100 million now,
down from about $250 million two years ago, while operating
margins should exceed 13% of sales this year, compared with about
12% last fiscal year, he says.

What could go wrong? One threat would be a price war, like the
one that raged through the industry in the late 1980s.

A second possible threat is Eastman Kodak Co.'s planned launch of

its Creation Station later this year. The Kodak unit will permit
consumers to personalize cards, with the added lure of using
their own photographs. The two companies may be forced to compete
for floor space, with retailers willing to make room for only one
such device.

But those dangers don't appear major. For now, the tide seems to
be going American Greetings ' way, as more card shoppers turn to
mass merchandise outlets. That should buoy American Greetings at
the expense of rival Hallmark, which depends primarily on its own
specialty card shops, says Rudy Hokanson, an analyst with C.J.
Lawrence/Deutsche Bank Securities Corp.

--- American Greetings (OTC symbol: AGREA)

Business: Greeting cards and gift items
Year ended Feb. 28, 1994: Revenue: $1.78 billion Net Income:
$113.7 million*; $1.54 a share
Fourth quarter, Feb. 28 1994: Per-share earning: 53 cents vs. 47 cents
Average daily trading volume: 379,292 shares

20
 21
Common shares outstanding: 74 million
*includes $17.2 million charge from accounting change

21
 22

Valuation Cookbook Version II:

(Risky World w/o Frictions)
Tony Marciano
NYU Stern School
1) Set up
a) Get projections information
b) Choose Horizon
c) Choose Long Term Growth
i) I often assume inflation
d) Determine methodology (WACC,APV,FCF to Equity)
i) More on this later
ii) For now, assume simple FCF to all investors and get aggregate enterprise value
2) Cash Flow Projections
a) Project Earnings
i) Project Sales
ii) Multiply by historical ratio to get to Gross Profit
iii) Project Depreciation
(1) CAPX straight-lined over proper time
iv) EBIT = Gross Profit – Operating Expense (Depr, SGA)
v) EBIAT = EBIT * (1-TaxRate)
b) Convert Earnings to FCF (Free Cash Flow)
i) Project Inc NWC w/ NWC = % of Sales
ii) FCF = EBIAT + DEPR – CAPX – IncNWC
c) Remember these are cash flows in expectation
d) More later. For now this is it.
3) Determine Discount Rate (The big incremental change)
a) Get Beta
i) Find Comps with Equity betas
ii) Unlever their equity betas to get to asset beta = B(equity)*%Equity
(1) Use their leverage ratio
(2) Book debt but Market Equity (P*#shares)
iii) Take median (or mean)
b) Get Risk-free Rate
i) Treasury rate of the maturity that matches the investment’s maturity
c) Plug into CAPM
i) Discount Rate = Rf + MRP*Beta
ii) MRP = 6 now and 7 for older cases
4) Mechanics
a) Discount Cash Flows at the Discount Rate
b) If there is an infinite horizon where a terminal value (TV) needs to be calculated
then take the last cash flow and divide over (r-g). Note that the result needs to be
placed one period before the final cash flow.

22