CHAPTER 12 : CASH FLOW ESTIMATION AND RISK ANALYSIS The basic principles of capital budgeting were covered in Chapter

ting were covered in Chapter 11. Given

a project’s expected cash flows, it is easy to calculate the primary decision
Home Depot Inc. (HD) has grown phenomenally in recent years, and criterion— the NPV—as well as the supplemental criteria, IRR, MIRR,
that growth continues. At the beginning of 1990, HD had 118 stores with payback, and discounted payback. However, in the real world, cash flow
annual sales of $2.8 billion. By early 2008, it had 2,234 stores and annual numbers are not just handed to you—rather, they must be estimated
sales of $77 billion. Stockholders have benefited mightily from this growth based on information from various sources. Moreover, uncertainty
as the stock’s price has increased from a split-adjusted $1.87 in 1990 to surrounds the forecasted cash flows, and some projects are more
$40 in early 2007, or by 2,039%. uncertain and thus riskier than others. In this chapter, we review examples
However, the more recent news has not been as good. In the face of that illustrate how project cash flows are estimated, discuss techniques for
a declining housing market, the company has struggled. In May 2008, it measuring and then dealing with risk, and discuss how projects are
announced the closing of 12 underperforming stores. Still, despite the poor evaluated once they go into operation. Finally, we discuss techniques to
housing market, the company continues to open new stores in areas it use when evaluating mutually exclusive projects that have unequal lives.
thinks the stores will do well. It costs, on average, over $20 million to When you finish this chapter, you should be able to:
purchase land, construct a new store, and stock it with inventory.  Identify “relevant” cash flows that should and should not be
Therefore, it is critical that the company perform a financial analysis to included in a capital budgeting analysis. L
determine whether a potential store’s expected cash flows will cover its  Estimate a project’s relevant cash flows and put them into a time
costs. line format that can be used to calculate a project’s NPV, IRR, and
Home Depot uses information from its existing stores to forecast other capital budgeting metrics.
its new stores’ expected cash flows. Thus far, its forecasts have been  Explain how risk is measured and use this measure to adjust the
outstanding, but there are always risks. First, a store’s sales might be less firm’s WACC to account for differential project riskiness.
than projected, especially if the economy weakens. Second, some of HD’s  Correctly calculate the NPV of mutually exclusive projects that
customers might bypass the store altogether and buy directly from have unequal lives.
manufacturers through the Internet. Third, its new stores could
“cannibalize,” or take sales away from, its existing stores. To avoid 12-1 CONCEPTUAL ISSUES IN CASH FLOW ESTIMATION
cannibalization while still opening enough new stores to generate Before the cash flow estimation process is illustrated, we need to discuss
substantial growth, HD has been developing complementary formats. For several important conceptual issues. A failure to handle these issues
example, it recently rolled out its Expo Design Center chain, which offers properly can lead to incorrect NPVs and thus bad capital budgeting
one-stop sales and service for kitchen, bath, and other remodeling and decisions.
renovation work; and in 2007, it acquired a Chinese home improvement 12-1a Cash Flow versus Accounting Income
chain to jump-start its operations in that nation. We saw in Chapter 3 that there is a difference between cash flows and
Rational expansion decisions require detailed assessments of the accounting income. We also saw that cash is what people and firms spend
forecasted cash flows, along with a measure of the risk that forecasted or reinvest; so the present value of cash flows, not accounting income, is
sales might not be realized. That information can then be used to the basis of a firm’s value. That’s why, in the last chapter, we discounted
determine the risk-adjusted NPV associated with each potential project. In net cash flows, not net income, to find projects’ NPVs
this chapter, we describe techniques for estimating projects’ cash flows, as Many things can lead to differences between net cash flows and net
well as projects’ risks. Companies such as Home Depot use these income. First, depreciation is not a cash outlay, but it is deducted when net
techniques on a regular basis when making capital budgeting decisions. income is calculated. Second, net income is based on the depreciation rate
the firm’s accountants decide to use, not necessarily on the depreciation
rate allowed by the IRS, and it is the IRS rate that determines cash flows. maintenance expenses, and the shiny new trucks also might improve the
Moreover, if a project requires an addition to working capital, this directly company’s image and reduce pollution. Replacement analysis is
affects cash flows but not net income. Other factors also differentiate net complicated by the fact that almost all of the cash flows are incremental,
income from cash flow, but the important thing to keep in mind is this: For found by subtracting the new cost numbers from the old numbers. Thus,
capital budgeting purposes, the project’s cash flows, not its accounting the fuel bill for a more efficient new truck might be $10,000 per year
income, is the key item. versus $15,000 for the old truck. The $5,000 savings is the incremental
cash flow that would be used in the replacement analysis. Similarly, we
12-1b Timing of Cash Flows would need to find the difference in depreciation and other factors that
affect cash flows. Once we have found the incremental cash flows, we use
In theory, capital budgeting analyses should deal with cash flows exactly them in a “regular” NPV analysis to decide whether to replace the asset or
when they occur; hence, daily cash flows theoretically would be better to continue using it.
than annual flows. However, it would be costly to estimate and analyze
daily cash flows, and they would probably be no more accurate than 12-1e Sunk Costs
annual estimates because we simply cannot accurately forecast at a daily
level out 10 years or so into the future. Therefore, we generally assume A sunk cost is an outlay that was incurred in the past and cannot be
that all cash flows occur at the end of the year. Note, though, for projects recovered in the future regardless of whether the project under
with highly predictable cash flows, it might be useful to assume that cash consideration is accepted. In capital budgeting, we are concerned with
flows occur at midyear (or even quarterly or monthly); but for most future incremental cash flows—we want to know if the new investment
purposes, we assume end-of-year flows. will produce enough incremental cash flow to justify the incremental
investment. Because sunk costs were incurred in the past and cannot be
12-1c Incremental Cash Flows recovered regardless of whether the project is accepted or rejected, they
are not relevant in the capital budgeting analysis. To illustrate this concept,
Incremental cash flows are flows that will occur if and only if some specific suppose Home Depot spent $2 million to investigate a potential new store
event occurs. In capital budgeting, the event is the firm’s acceptance of a and obtain the permits required to build it. That $2 million would be a sunk
project and the project’s incremental cash flows are ones that occur as a cost—the money is gone, and it won’t come back regardless of whether or
result of this decision. Cash flows such as investments in buildings, not the new store is built. Not handling sunk costs properly can lead to
equipment, and working capital needed for the project are obviously incorrect decisions. For example, suppose Home Depot completed the
incremental, as are sales revenues and operating costs associated with the analysis and found that it must spend an additional $17 million, on top of
project. However, some items are not so obvious, as we explain later in the $2 million site study, to open the store. Suppose it then used as the
this section. required investment $19 million and found a projected NPV of negative $1
million. This would indicate that HD should reject the new store. However,
12-1d Replacement Projects that would be a bad decision. The real issue is whether the incremental
$17 million would result in incremental cash inflows sufficient to produce a
Two types of projects can be distinguished: expansion projects, where the positive NPV. If the $2 million sunk cost is disregarded, as it should be, the
firm makes an investment, such as a new Home Depot store, and true NPV will be a positive $1 million. Therefore, the failure to deal
replacement projects, where the firm replaces existing assets, generally to properly with the sunk cost would lead to turning down a project that
reduce costs. For example, suppose Home Depot is considering replacing would add $1 million to stockholders’ value.
some of its delivery trucks. The benefit would be lower fuel and
12-1f Opportunity Costs Associated with Assets the Firm Owns into the company’s existing business. Manufacturers also can experience
cannibalization. Thus, if Cengage Learning, the publisher of this book,
Another issue relates to opportunity costs associated with assets the firm decides to publish another introductory finance text, that new book will
already owns. For example, suppose Home Depot owns land with a market presumably reduce sales of this one. Those lost cash flows should be taken
value of $2 million and that land will be used for the new store if HD into account, and that means charging them as a cost when analyzing the
decides to build it. If HD decides to go forward with the project, only proposed new book.
another $15 million will be required, not the typical $17 million because Dealing properly with negative externalities can be tricky. If Cengage
HD would not need to buy the required land. Does this mean that HD decides not to publish the new book because of its cannibalization effects,
should use $15 million as the cost of the new store? The answer is no. If might another publisher publish it, causing our book to lose sales
the new store is not built, HD could sell the land and receive a cash flow of regardless of what Cengage does? Logically, Cengage must examine the
$2 million. This $2 million is an opportunity cost— something that HD total situation, which is more than a simple mechanical analysis.
would not receive if the land was used for the new store. Therefore, the $2 Experience and knowledge of the industry is required to make good
million must be charged to the new project, and a failure to do so would decisions.
artificially and incorrectly increase the new project’s NPV. If this is not One of the best examples of a company fouling up as a result of not
clear, consider the following example. Assume that a firm owns a building dealing correctly with cannibalization effects was IBM’s response when
and equipment with a market (resale) value of $10 million. The property is transistors made personal computers possible in the 1970s. IBM’s
not being used, and the firm is considering using it for a new project. The mainframe computers were the biggest game in town, and they generated
only required additional investment would be $100,000 for working huge profits. But IBM also had the technology, entered into the PC market,
capital, and the new project would produce a cash inflow of $50,000 and for a time was the leading PC company. However, top management
forever. If the firm has a WACC of 10% and evaluates the project using only decided to rein back the PC division because managers were afraid it
the $100,000 of working capital as the required investment, it would find would hurt the more profitable mainframe business. That decision opened
an NPV of $50,000/0.10 ¼ $500,000. Does this mean that the project is a the door for Microsoft, Intel, Dell, Hewlett-Packard, and others; and IBM
good one? The answer is no. The firm can sell the property for $10 million, went from being the most profitable firm in the world to one whose very
which is a much larger amount than $500,000. survival was threatened. This experience highlights the fact that while it is
essential to understand the theory of finance, it is equally important to
12-1g Externalities understand the business environment, including how competitors are
likely to react to a firm’s actions. A great deal of judgment goes into
Another potential problem involves externalities, which are defined as the making good financial decisions.
effects of a project on other parts of the firm or the environment. The
three types of externalities—negative within-firm externalities, positive Positive Within-Firm Externalities
within-firm externalities, and environmental externalities—are explained Cannibalization occurs when new products compete with old ones.
next. However, a new project also can be complementary to an old one, in which
Negative Within-Firm Externalities case cash flows in the old operation will be increased when the new one is
As noted earlier, when retailers such as Home Depot open new stores introduced. For example, Apple’s iPod was a profitable product; but when
that are too close to their existing stores, this takes customers away from Apple made an investment in another project, its music store, that
their existing stores. In this case, even though the new store has positive investment boosted sales of the iPod. So if an analysis of the proposed
cash flows, its existence reduces some of the firm’s current cash flows. This music store indicated a negative NPV, the analysis would not be complete
type of externality is called cannibalization because the new business eats unless the incremental cash flows that would occur in the iPod division
were credited to the music store. That might well change the project’s NPV
from negative to positive.

Environmental Externalities
The most common type of negative externality is a project’s impact on
the environment. Government rules and regulations constrain what
companies can do, but firms have some flexibility in dealing with the
environment. For example, suppose a manufacturer is studying a proposed
new plant. The company could meet the environmental regulations at a
cost of $1 million, but the plant would still emit fumes that might cause ill
feelings in its neighborhood. Those ill feelings would not show up in the
cash flow analysis, but they still should be considered. Perhaps a relatively
small additional expenditure would reduce the emissions substantially,
make the plant look good relative to other plants in the area, and provide
goodwill that would help the firm’s sales and negotiations with
governmental agencies in the future.
Of course, everyone’s profits depend on the earth remaining healthy, so
companies have an incentive to do things to protect the environment even
though those actions are not required. However, if one firm decides to
take actions that are good for the environment but costly, its products
must reflect the higher costs. If its competitors decide to get by with less
costly but less environmentally friendly processes, they can price their
products lower and make more money. Of course, more environmentally
friendly companies can advertise their environmental efforts, and this
might—or might not—offset the higher costs. All of this illustrates why
government regulations are necessary, both nationally and internationally.
Finance, politics, and the environment are all interconnected.

12-2 ANALYSIS OF AN EXPANSION PROJECT

In Chapter 11, we analyzed two projects, S and L. We were given the cash
flows and used them to illustrate how the NPV, IRR, MIRR, and payback are
calculated. Now we demonstrate how cash flows are actually estimated,
using our old Project S to demonstrate the procedure. We explain the
process in Table 12-1. Look at it as we discuss the analysis. Note that the
dollars are in thousands; we omitted
three zeros to streamline the presentation. Also note that we used Excel to If for some reason the firm decided to use straight-line depreciation, it
make Table 12-1. We could have used a calculator and plain paper, but could write off a constant $225 per year. Its total cash flows over the entire
Excel is much better in dealing with arithmetic. You don’t need to know 4 years would be the same as under accelerated depreciation; but under
Excel to understand the discussion; but if you plan to work in finance—or straight line, those cash flows would come in a bit slower because the firm
in almost any business function—you should learn something about it. would have higher tax payments in the early years and lower tax payments
The column headers in the table, the A through I, and the row headers, later on.
1 through 38, designate cells, which contain the data. For example, the We calculate the annual cash flows for Project S over the 4 years in
equipment needed for Project S will cost $900, and that number is shown Columns F, G, H, and I, ending with the net cash flows shown on Row 22.
in Cell E4 as a negative. The equipment is expected to have a salvage value The numbers in Cells E22 through I22 amount to a cash flow time line, and
of $50 at the end of the project’s 4-year life; this is shown in Cell I19.1 The they are the same numbers used in Chapter 11 for Project S. Since the
new project will require $100 of working capital; this is shown in Cell E5 as numbers are the same, the NPV, IRR, MIRR, and Payback shown in Cells
a negative number because it is a cost and then as a positive number in C31 through C34 are the same as those we calculated in Chapter 11.
Cell I21 because it is recovered at the end of Year 4. The total investment The Excel model used to create Table 12-1 is part of the chapter Excel
at Time 0 is $1,000, which is shown in Cell E22. model available on the text’s web site. We recommend that anyone with a
Unit sales of Project S are shown on Row 7; they are expected to decline computer and some familiarity with Excel access the model and work
somewhat over the project’s 4-year life. The sales price, a constant $10, is through it to see how the table was generated. Anyone doing real-world
shown on Row 8. The projected variable cost per unit is given on Row 9; it capital budgeting today would use such a model; and our model provides a
generally increases over time due to expected increases in materials and good template, or starting point, if and when you need to analyze an actual
labor. Sales revenues, which are calculated as units multiplied by price, are project.
given on Row 10. Variable costs, equal to units multiplied by VC/unit, are
given on Row 11; and fixed costs excluding depreciation, which are a 1 The equipment will be fully depreciated after 4 years. Therefore, the $50
constant $2,000, are shown on Row 12. estimated salvage value will exceed the book value, which will be zero. This
Depreciation is found as the annual rate allowed by the IRS times the $50 gain is classified as a recapture of depreciation, and it is taxed at the
depreciable basis. As noted in Chapter 3, Congress sets the depreciation same rate as ordinary income
rates that can be used for tax purposes and these are the tax rates used in
the capital budgeting analysis. Congress permits firms to depreciate assets 12-2a Effect of Different Depreciation Rates
by the straightline method or by an accelerated method. As we will see,
profitable firms are better off using accelerated depreciation. We discuss If we replaced the accelerated depreciation numbers in Table 12-1 with
depreciation more fully in Appendix 12A; but to simplify things for this the constant $225 values that would exist under straight line, the result
chapter, we assume that the applicable accelerated rates for a project with would be a cash flow time line on Row 22 that has the same total flows.
a 4-year life are as given on Row 24 of the depreciation section of the table However, in the early years, the cash flows resulting from straight-line
and that straight-line rates are as given on Row 27. Thus, we assume that if depreciation would be lower than those now in the table; and the later
the firm uses accelerated depreciation, it will write off 33% of the basis years’ cash flows would show higher numbers. You know that dollars
during Year 1, another 45% in Year 2, and so forth. These are the rates received earlier have a higher present value than dollars received later.
used to obtain the cash flows shown in the table. Therefore, Project S’s NPV is higher if the firm uses accelerated
The depreciable basis is the cost of the equipment including any depreciation. The exact effect is shown in the Project Evaluation section of
shipping or installation costs, or $900 as shown in Cells E4, C24, and C27. Table 12-1—the NPV is $78.82 under accelerated depreciation and $64.44,
The total depreciation over the 4 years equals the cost of the equipment. or 18% less, with straight line.
 Now suppose Congress wants to encourage companies to increase their incremental cost; it is a sunk cost. Therefore, it should not enter into the
capital expenditures to boost economic growth and employment. What analysis.
change in depreciation would have the desired effect? The answer is to One additional point should be made about sunk costs. If the $150
make accelerated depreciation even more accelerated. For example, if the expenditure was actually made, in the final analysis, Project S would turn
firm could write off this 4-year equipment at rates of 50%, 35%, 10%, and out to be a loser: Its NPV would be $78.82 − $150 ¼ −$71.18. If we could
5%, its early tax payments would be lower, early cash flows would be somehow back up and Chapter 12 Cash Flow Estimation and Risk Analysis
higher, and the project’s NPV would be higher than that shown in Table 12- 371 reconsider the project before the $150 had been spent, we would see
1. that the project should be rejected. However, we can’t back up—at this
point, we can either abandon the project or spend $1,000 and go forward
12-2b Cannibalization with it. If we go forward, we will receive an incremental NPV of $78.82,
which would reduce the loss from −$150 to −$71.18.
Project S does not involve any cannibalization effects. Suppose, however,
that Project S would reduce the net after-tax cash flows of another division 12-2e Other Changes to the Inputs
by $50 per year. No other firm would take on this project if our firm turns
it down. In this case, we would add a row at about Row 18 and deduct $50 Variables other than depreciation also could be varied, and these
for each year. If this were done, Project S would end up with a negative changes would alter the calculated cash flows and thus NPV and IRR. For
NPV; hence, it would be rejected. On the other hand, if Project S would example, we could increase or decrease the projected unit sales, the sales
cause additional flows in some other division (a positive externality), those price, the variable and/or the fixed costs, the initial investment cost, the
after-tax inflows should be attributed to Project S working capital requirements, the salvage value, and even the tax rate if
we thought Congress was likely to raise or lower taxes. Such changes could
12-2c Opportunity Costs be made easily in an Excel model, making it possible to see the resulting
changes in NPV and IRR immediately. This is called sensitivity analysis, and
Now suppose the $900 initial cost shown in Table 12-1 was based on the we discuss it later in the chapter when we take up procedures for
assumption that the project would save money by using some equipment measuring projects’ risks.
the company now owns and that equipment would be sold for $100, after
taxes, if the project is rejected. The $100 is an opportunity cost, and it 12-3 REPLACEMENT ANALYSIS
should be reflected in our calculations. We would add $100 to the project’s In the last section, we assumed that Project S was an entirely new
cost. The result would be an NPV of $78.82 − $100 ¼ −$21.18, so the project. So all of its cash flows were incremental—they occurred only if the
project would be rejected. firm accepted the project. This is true for expansion projects; but for
replacement projects, we must find cash flow differentials between the
12-2d Sunk Costs new and old projects and these differentials are the incremental flows that
we analyze.
Now suppose the firm had spent $150 on a marketing study to estimate We evaluate a replacement decision in Table 12-2, which is set up much
potential sales. This $150 could not be recovered regardless of whether like Table 12-1, but with data on both a new, highly efficient machine
the project is accepted or rejected. Should the $150 be charged to Project (which will be depreciated on an accelerated basis) and the old machine
S when determining its NPV for capital budgeting purposes? The answer is (which is depreciated on a straight-line basis). Here we find the firm’s cash
no. We are interested only in incremental costs. The $150 is not an flows when it continues using the old machine, then the cash flows when it
decides to use the new one. Finally, we subtract the old flows from the
new to arrive at the incremental cash flows. We used Excel in our analysis; would, of course, be reflected in the differential cash flows on Row
but again, we could have used a calculator or pencil and paper. Here are 25.
the key inputs used in the analysis. No additional working capital is
needed.

The key here is to find the incremental cash flows. As noted previously,
we find the cash flows from the operation with the old machine, then find
the cash flows with the new machine, then find the differences in the cash
flows. This is what we do in Parts I, II, and III of Table 12-2. Since there will
be an additional expenditure to buy the new machine, that cost is shown in
Cell E13. However, we can sell the old machine for $400, so that is shown
as an inflow in Cell E14. The net cash outlay at Time 0 is $1,600, as shown
in Cell E23.
The net cash flows based on the old machine are shown on Row 11, and
those for the new one are on Row 23. Then on Row 25, we show the
differences in the cash flows with and without replacement—these are the
incremental cash flows used to find the replacement NPV. When we
evaluate the incremental cash flows, we see that the replacement has
an NPV of $80.28, so the old machine should be replaced.
In some instances, replacements add capacity as well as lower
operating costs. When this is the case, sales revenues in Part II
would be increased; and if that led to a need for more working
capital, that number would be shown as a Time 0 expenditure along
with a recovery at the end of the project’s life. These changes
12-4 RISK ANALYSIS IN CAPITAL BUDGETING risk and then consider the other two risk measures in a qualitative
manner.
Projects differ in risk, and risk should be reflected in capital budgeting
decisions. However, it is difficult to measure risk, especially for new Projects are generally classified into several categories. Then
projects where no history exists. For this reason, managers deal with risk in with the firm’s overall WACC as a starting point, a risk-adjusted
many different ways, ranging from almost totally subjective adjustments to cost of capital is assigned to each category. For example, a firm
highly sophisticated analyses that involve computer simulation and high- might establish three risk classes, assign the corporate WACC to
powered statistics. average-risk projects, add a 5% risk premium for higher-risk
Three separate and distinct types of risk are involved: projects, and subtract 2% for low-risk projects. Under this setup, if
1. Stand-alone risk, which is a project’s risk assuming (a) that it is the the company’s overall WACC was 10%, 10% would be used to
only asset the firm has and (b) that the firm is the only stock in evaluate average-risk projects, 15% for high-risk projects, and 8%
each investor’s portfolio. Stand-alone risk is measured by the for low-risk projects. While this approach is probably better than
variability of the project’s expected returns. Diversification is not making any risk adjustments, these adjustments are highly
totally ignored subjective and difficult to justify. Unfortunately, there’s no perfect
2. Corporate, or within-firm, risk, which is a project’s risk to the way to specify how high or low the adjustments should be.
corporation as opposed to its investors. Within-firm risk takes
account of the fact that the project is only one asset in the firm’s 12-5 MEASURING STAND-ALONE RISK
portfolio of assets; hence, some of its risk will be eliminated by
diversification within the firm. This type of risk is measured by the A project’s stand-alone risk reflects uncertainty about its cash
project’s impact on uncertainty about the firm’s future returns. flows. The required investment, unit sales, sales prices, and
3. Market, or beta, risk, which is the riskiness of the project as seen operating costs shown in Table 12-1 for Project S are subject to
by a well diversified stockholder who recognizes (a) that the uncertainty. First-year sales were projected at 537 units (actually,
project is only one of the firm’s assets and (b) that the firm’s stock 537,000, but we shortened it to 537 to streamline the analysis) to
is but one part of his or her stock portfolio. The project’s market be sold at a price of $10 per unit. However, unit sales would
risk is measured by its effect on the firm’s beta coefficient. almost certainly be somewhat higher or lower than 537, and the
price would probably turn out to be different from the projected
Taking on a project with a great deal of stand-alone or corporate $10 per unit. Similarly, the other variables would probably differ
risk will not necessarily affect the firm’s beta. However, if the from their indicated values. Indeed, all the inputs are expected
project has high stand-alone risk and if its returns are highly values, and actual values can vary from expected values.
correlated with returns on the firm’s other assets and with returns Three techniques are used to assess stand-alone risk: (1)
on most other stocks in the economy, the project will have a high sensitivity analysis, (2) scenario analysis, and (3) Monte Carlo
degree of all three types of risk. Market risk is theoretically the simulation. We discuss them in the following sections.
most relevant of the three because it is the one reflected in stock
prices. Unfortunately, market risk is also the most difficult to 12-5a Sensitivity Analysis
estimate, primarily because new projects don’t have “market
prices” that can be related to stock market returns. Therefore, Intuitively, we know that a change in a key input variable such
most decision makers do a quantitative analysis of stand-alone as units sold or sales price will cause the NPV to change.
Sensitivity analysis measures the percentage change in NPV that
 results from a given percentage change in an input, other
variables held at their expected values. This is by far the most
commonly used type of risk analysis, and it is used by most firms.
It begins with a base-case situation, where the project’s NPV is
found using the base-case value for each input variable. Here’s a
list of the key inputs for Project S:
 Equipment cost
 Required working capital
 Unit sales
 Sales price
 Variable cost per unit
 Fixed operating costs
 Tax rate
 WACC
The data we used back in Table 12-1 were the most likely, or base-case,
values; and the resulting NPV, $78.82, is the base-case NPV. It’s easy to
imagine changes in the inputs, and those changes would result in different
NPVs.
When senior managers review capital budgeting studies, they are
interested in the base-case NPV, but they always go on to ask the financial
analyst a series of “what if” questions: What if unit sales turn out to be
25% below the base-case level? What if market conditions force us to price
the product at $9, not $10? What if variable costs are higher than we
forecasted? Sensitivity analysis is designed to provide answers to such
questions. Each variable is increased or decreased from its expected value,
holding other variables constant at their basecase levels. Then the NPV is
calculated using the changed input. Finally, the resulting set of NPVs is
plotted to show how sensitive NPV is to changes in each variable
Figure 12-1 shows Project S’s sensitivity graph for six key variables. The
table below the graph gives the NPVs based on different values of the
inputs, and those NPVs were then plotted to make the graph. Figure 12-1
shows that as unit sales and price increase, the project’s NPV increases,
whereas the opposite is true for the other four input variables. An increase Figure 12-1, the slopes of the lines in the graph indicate how sensitive NPV
in variable costs, fixed costs, equipment costs, and WACC lowers the is to each input: The larger the range, the steeper the variable’s slope and
project’s NPV. The ranges shown at the bottom of the table and the slopes the more sensitive the NPV is to this variable. We see that NPV is very
of the lines in the graph indicate how sensitive NPV is to changes in each sensitive to changes in the sales price, fairly sensitive to changes in variable
input. When the data are plotted in costs, a bit less sensitive to units sold and fixed costs, but not very sensitive
to changes in the equipment cost or the WACC.
 If we were comparing two projects, the one with the steeper sensitivity $5,028.94. When we divide the standard deviation by the expected NPV,
lines would be riskier, other things held constant, because relatively small we get the coefficient of variation, 7.12, which is a measure of stand-alone
changes in the input variables would produce large changes in the NPV. risk. The firm’s average-risk project has a coefficient of variation of about
Thus, sensitivity analysis provides useful insights into a project’s risk. 2.0, so the CV of 7.12 indicates that this project is much riskier than most
of the firm’s other projects.
12-5b Scenario Analysis Our firm’s WACC is 10%, so that rate should be used to find the NPV of
an average-risk project. Project S is riskier than average, so a higher
In sensitivity analysis, we change one variable at a time. However, it is discount rate should be used to find its NPV. There is no way to determine
useful to know what would happen to the project’s NPV if all of the inputs the “correct” discount rate—this is a judgment call. However, some firms
turned out to be better or worse than expected. Also, we can assign increase the corporate WACC when they evaluate projects deemed to be
probabilities to the good, bad, and most likely (or base-case) scenarios, relatively risky and reduce it for low-risk projects. When the NPV was
then find the expected value and the standard deviation of the NPV. recalculated using a 12.5% WACC, the basecase NPV fell from $78.82 to
Scenario analysis allows for these extensions—it allows us to change more $33.62; so the project still passed the NPV test.
than one variable at a time, and it incorporates the probabilities of changes Note that the base-case results are the same in our sensitivity and
in the key variables. scenario analyses; but in the scenario analysis, the worst case is much
In a scenario analysis, we begin with the base-case scenario, which uses worse than in the sensitivity analysis and the best case is much better. This
the most likely set of input values. We then ask marketing, engineering, is because in scenario analysis, all of the variables are set at their best or
and other operating managers to specify a worst-case scenario (low unit worst levels, while in sensitivity analysis, only one variable is adjusted and
sales, low sales price, high variable costs, and so forth) and a best-case all the others are left at their base-case levels
scenario. Often the best and worst cases are defined as having a 25%
probability of conditions being that good or bad, with a 50% probability for
the base-case conditions. Obviously, conditions can take on many more
than three values, but such a scenario setup is useful to help in
understanding the project’s riskiness.
The best-case, base-case, and worst-case values for Project S are shown
in Figure 12-2, along with plots of the data. If the project is highly
successful, the combination of a high sales price, low production costs, and
high unit sales will result in a very high NPV, $7,450.38. However, if things
turn out badly, the NPV will be a negative $4,782.40. The graphs show the
wide range of possibilities, suggesting that this is a risky project. If the bad
conditions materialize, the company will not go bankrupt—this is just one
project for a large company. Still, losing $4,782.40 (or $4,782,400 since we
are working in thousands) would hurt the stock price.
If we multiply each scenario’s probability by the NPV under that scenario
and then sum the products, we will have the project’s expected NPV,
$706.40 as shown in Figure 12-2. Note that the expected NPV differs from
the base-case NPV. This is not an error—mathematically, they are not
equal. We also calculate the standard deviation of the expected NPV; it is
 profitability, and the standard deviation (or perhaps the coefficient of
variation) of the NPVs is used as a measure of risk.

12-5c Monte Carlo Simulation

Monte Carlo simulation, so named because this type of analysis grew out
of work on the mathematics of casino gambling, is a sophisticated version
of scenario analysis. Here the project is analyzed under a large number of
scenarios, or “runs.” In the first run, the computer randomly picks a value
for each variable—units sold, sales price, variable costs per unit, and so
forth. Those values are then used to calculate an NPV, and that NPV is Monte Carlo simulation is technically more complex than scenario analysis, but simulation
stored in the computer’s memory. Next, a second set of input values is software makes the process manageable. Simulation is useful; but because of its complexity, a
detailed discussion is best left for advanced finance courses.
selected at random and a second NPV is calculated. This process is
repeated perhaps 1,000 times, generating 1,000 NPVs. The mean of the
1,000 NPVs is determined and used as a measure of the project’s expected
  It is very difficult, if not impossible, to quantitatively measure
12-6 WITHIN- FIRM AND BETA RISK projects’ within firm and beta risks.
 Most projects’ returns are positively correlated with returns on
Sensitivity analysis, scenario analysis, and Monte Carlo simulation as the firm’s other assets and with returns on the stock market. This
described in the preceding section dealt with stand-alone risk. They being the case, because stand-alone risk is correlated with within-
provide useful information about a project’s risk; but if the project is firm and market risk, not much is lost by focusing just on stand-
negatively correlated with the firm’s other projects, it might stabilize the alone risk.
firm’s total earnings and thus be relatively safe. Similarly, if a project is  Experienced managers make many judgmental assessments,
negatively correlated with returns on most stocks, it might reduce the including those related to risk; and they work them into the
firm’s beta and thus be correctly evaluated with a relatively low WACC. So capital budgeting process. Introductory students like neat, precise
in theory, we should be more concerned with within-firm and beta risk answers; and they want to make decisions on the basis of
than with stand-alone risk. calculated NPVs. Experienced managers consider quantitative
Although managers recognize the importance of within-firm and beta NPVs, but they also bring subjective judgment into the decision
risk, they generally end up dealing with these risks subjectively, or process.
judgmentally, rather than quantitatively. The problem is that to measure  If a firm does not use the types of analyses covered in this book, it
diversification’s effects on risk, we need the correlation coefficient will have trouble. On the other hand, if a firm tries to quantify
between a project’s returns and returns on the firm’s other assets, which everything and let a computer make its decisions, it too will have
requires historical data that obviously do not exist for new projects. trouble. Good managers understand and use the theory of
Experienced managers generally have a “feel” for how a project’s returns finance, but they apply it with judgment.
will relate to returns on the firm’s other assets. Generally, positive
correlation is expected; and if the correlation is high, standalone risk will 12-7 UNEQUAL PROJECT LIVES
be a good proxy for within-firm risk. Similarly, managers can make
judgmental estimates about whether a project’s returns will be high when If a company is choosing between two projects and those projects
the economy and the stock market are strong (hence, what the project’s (1) have significantly different lives, (2) are mutually exclusive, and
beta should be). But for the most part, those estimates are subjective, not (3) can be repeated, the “regular” NPV method may not indicate
based on actual data. the better project. For example, suppose Home Depot is planning
However, projects occasionally involve an entirely new product line, to modernize a distribution center; it is choosing between a
such as a steel company going into iron ore mining. In such cases, the firm conveyor system (Project C) and a fleet of forklift trucks (Project
may be able to obtain betas for “pure-play” companies in the new area. F). The projects are mutually exclusive—choosing one means
For example, this steel company might get the average beta for a group of rejecting the other. Also, the distribution center will be used for
mining companies such as Rio Tinto and BHP, assume that its mining many years, so the equipment will be replaced when it wears out.
subsidiary has similar characteristics, and use the average beta of the Part I of Figure 12-3 shows the analysis that traditionally would
“comparables” to calculate a WACC for the mining subsidiary. While the be used to analyze the two projects. We see that Project C, when
pure-play approach makes sense for some projects, it is rare. Just think discounted at a 12% WACC, has the higher NPV and thus appears
about it. How would you find a pure-play proxy for a new inventory control to be the better project. However, the traditional analysis is
system, machine tool, truck, or most other projects? The answer is, you incomplete, and the decision to choose Project C is actually
couldn’t incorrect. If we choose Project F, we will have an opportunity to
Our conclusions regarding risk analysis are as follows: make a similar investment in 3 years; and if costs and revenues
remain at the Part I levels, this second investment also will be
profitable. If we choose Project C, we will not have the option to Electrical engineers designing power plants and distribution lines
make this second investment. Therefore, to make a proper were the first to encounter the unequal life problem. They could
comparison between C and F we must make an adjustment. We use transformers that had a relatively low initial cost but a short
discuss the two methods for making the adjustment in the life, or they could use transformers that had higher initial costs
remainder of this section. but longer lives. Transformers would be required on into the
indefinite future, so this was the issue: Which choice would result
12-7a Replacement Chains in the higher NPV over the long run? The engineers first found the
NPV of each project over its stated life and then found the
First, we can apply the replacement chain (common life) approach constant annual cash flow that this NPV would provide over the
as shown in Part II of Figure 12-3. This involves finding the NPV of project’s initial life. Since the projects would presumably be
Project F over 6 years, which is also the life of Project C, and then repeated indefinitely, those annuity payments would continue
comparing this extended NPV with the NPV of Project C over the indefinitely and the project that provided the higher payment
same 6 years. We see that on a common-life basis, F turns out to stream was the better option. This procedure was called the
be the better project. equivalent annual annuity (EAA) method. The EAAs of Projects C
and F are calculated in Part III of Figure 12-3. We first find the
projects’ traditional NPVs and then find the EAAs of those NPVs.
As you can see, Project F is the better choice, the same decision
reached by using the replacement chain approach.

12-7c Conclusions about Unequal Lives

The replacement chain and EAA methods always result in the

same decision, so it doesn’t matter which one is used. The EAA is
a bit easier to implement, especially when the longer project
doesn’t have exactly twice the life of the shorter one—and hence
more than two cycles are needed to find a common life. However,
the replacement chain method is often easier to explain to senior
managers. Also, it is easier to make modifications to the
replacement chain data to deal with anticipated productivity
improvements and asset price changes. For those reasons, we
generally use the replacement chain method when we work with
non engineers; but when engineers are involved, we show both
results.
Another question often arises: Do we have to worry about
unequal life analysis for all projects that have unequal lives? As a
general rule, the unequal life issue never arises for independent
12-7b Equivalent Annual Annuities (EAA) projects, but it can be an issue when we compare mutually
 exclusive projects with significantly different lives. However, the cannot be measured for most projects, stand-alone risk is the one
issue arises if and only if the projects will be repeated at the end on which we generally focus. However, firms subjectively consider
of their initial lives. Thus, for all independent projects and those within-firm and market risk, which they definitely should not
mutually exclusive projects that will not be repeated, there is no ignore. Note, though, that since the three types of risk are
need to adjust for unequal lives. generally positively correlated, stand-alone risk is often a good
proxy for the other risks.
TRYING IT ALL TOGETHER  Stand-alone risk can be analyzed using sensitivity analysis,
scenario analysis, and/or Monte Carlo simulation.
This chapter focused on estimating the cash flows that are used in  Once a decision has been made about a project’s relative risk, we
a capital budgeting analysis, appraising the riskiness of those determine a risk-adjusted WACC for evaluating it.
flows, finding NPVs when risk is present, and calculating the NPVs  If mutually exclusive projects have unequal lives and are
of mutually exclusive projects having unequal lives. Here is a repeatable, a traditional NPV analysis may lead to incorrect
summary of our primary conclusions: results. In this case, we should use replacement chain or
 Some cash flows are relevant (hence, should be included in a equivalent annual annuity (EAA) analysis.
capital budgeting analysis), while others should not be included.
The key question is this: Is the cash flow incremental in the sense
that it will occur if and only if the project is accepted?
 Sunk costs are not incremental costs—they are not affected by
accepting or rejecting the project. Cannibalization and other
externalities, on the other hand, are incremental—they will occur
if and only if the project is accepted.
 The cash flows used to analyze a project are different from a
project’s net income. One important factor is that depreciation is
deducted when accountants calculate net income; but because it
is a noncash charge, it must be added back to find cash flows.
 Many projects require additional net working capital. Net working
capital is a negative flow when the project is started but a positive
flow at the end of the project’s life, when the capital is recovered.
 We considered two types of projects—expansion and
replacement. For a replacement project, we find the difference in
the cash flows when the firm continues to use the old asset versus
the new asset. If the NPV of the differential flows is positive, the
replacement should be made.
 The forecasted cash flows (and hence NPV and other outputs) are
only estimates—they may turn out to be incorrect, and this means
risk
 There are three types of risk: stand-alone, within-firm, and market
(or beta) risk. In theory, market risk is most relevant; but since it