Lecture: VI 1

Project Analysis using Decision Trees and Options

Decisions on projects always involve uncertainty. How do

you deal with it?
• Probabilistic Analysis: Assess risk by estimating project
values for a range of outcomes for the underlying
variables:
- Sensitivity Analysis
- Scenario Analysis
- Breakeven Analysis
- Simulations
- Decision Trees
- Valuing the options inherent in the project:
- the option to delay a project
- the option to expand in the future
- the option to abandon the project
• Risk Analysis: Factor the risk into either the discount rate
or the expected cash flows explicitly, and calculate risk-
adjusted performance measures.

BAFI 402: Financial Management I, Fall 2001 A. Gupta

Lecture: VI 2

Sensitivity Analysis

• Project analysis is based on cash flows

• Cash flows are uncertain, and are based on assumptions
• Changes in the assumptions can change the decision

Sensitivity Analysis examines the sensitivity of a decision rule

(NPV, IRR, etc.) to changes in the assumptions underlying a
project.

The steps:
• Do a base case analysis, based on expectations about the
future.
• Identify key assumptions in the base case analysis -
these could be firm specific (revenue levels, operating
costs, etc.) or macroeconomic (tax rates, inflation, etc.).
• Change one key assumption at a time, and estimate the
decision criterion (NPV, IRR, etc.) - summarize the
impact of changing the key assumption on the decision
criterion in the form of a table or graph.
• Decide whether or not to take the project based on the
risk of changes in the key assumptions.

The key assumptions must be either:

• those that matter the most, in terms of affecting cash
flows (e.g. revenue levels), or
• those with the most uncertainty (e.g., it does not make
sense to vary depreciation rates if they are fixed!).

BAFI 402: Financial Management I, Fall 2001 A. Gupta

Lecture: VI 3

Example:
Van Ommeren is considering construction of a new Oil Jetty at
Rotterdam port for unloading petroleum products, at a cost of $40
million, which would save them $5 million every year in
unloading costs paid to the Rotterdam Port Trust. The jetty has an
expected life of 25 years. How sensitive is the project decision to
life expectancy of the oil jetty? (assume r=12%)

• For a 25 yr life, NPV = $-0.7 mil, so it should be rejected.

• However, NPV as a function of the project life is:
$1,000,000

$500,000

$0
20 22 24 26 28 30 32 34 36
NPV ($)

($500,000)

($1,000,000)

($1,500,000)

($2,000,000)

($2,500,000)
Project Life (yrs)

• Hence, if the actual life of the jetty is 35 years, it could

be a very profitable project!
• This shows how sensitive the NPV is to assumptions
about project life.

BAFI 402: Financial Management I, Fall 2001 A. Gupta

Lecture: VI 4

Limitations of Sensitivity Analysis:

• It provides results for a range of values of key
parameters, without providing any sense of the
likelihood of these values occurring - In the previous
example, how likely it is for the project life to be more
than 30 years is important in evaluating the project.
• Only one assumption is changed at a time - in the real
world, variables often move together. For instance,
revenues may be low in precisely the cases when
inflation is high and growth rates are low.
This can be corrected using scenario analysis, where
specific scenarios are developed for the future, and the
viability of the projects is considered under each scenario
(scenarios consider changes in many variables at one time).
One can also do a best case and a worst case analysis to
assess the potential upside and downside of the project.
• This analysis is used subjectively - due to different risk
preferences of decision makers, one might accept the
project while the other might reject it, based on the same
analysis.

BAFI 402: Financial Management I, Fall 2001 A. Gupta

Lecture: VI 5

Breakeven Analysis

Accounting Breakeven: The number of units a firm must sell

to ensure that it doesn’t make an accounting loss.
Total Fixed Costs
• Breakeven Units =
Sales Price - Variable Cost
Unit Unit

Financial Breakeven: The number of units a firm must sell to

have a zero NPV.
• Estimate the annual cash flow needed to make the NPV
zero.
• Back out the revenues needed to generate these cash
flows.
• Estimate the number of units that must be sold to create
this revenue.

Financial breakeven is not only a higher hurdle, but also a

more realistic hurdle, since it factors in the opportunity
cost of funds invested in the project.

Example: The Boeing 777 project.

Sales price per plane = $130 million
Variable cost per plane = $ 97.50 million
Total fixed costs = $2.35 billion
therefore,
Accounting breakeven = 72.31 planes
i.e., Boeing must sell at least 73 planes per year to cover its
fixed costs and not show an accounting loss.

BAFI 402: Financial Management I, Fall 2001 A. Gupta

Lecture: VI 6

Financial Breakeven:
• The free cash flows can be estimated for the project for
various sales levels (planes per year).
• For each of these cash flows, the NPV can be computed.
• The NPV can be plotted as a function of the sales level,
to see at what sales level the NPV is zero.

In this example, the financial breakeven comes to about 78

planes per year.

Caveats:
• Breakeven analysis by itself does not answer the
question of whether we should accept or reject the
project - it just provides additional information to use in
the decision
- it provides a measure of margin of safety the decision
maker has if the project is accepted.
- Once the project is accepted, it provides a useful
benchmark against which actual performance can be
compared.

BAFI 402: Financial Management I, Fall 2001 A. Gupta

Lecture: VI 7

Decision Trees

This is a presentation of the decisions and possible outcomes,

with probabilities, at each stage of a multistage project.

• Break the project into clearly defined stages.

• List all possible outcomes at each stage.
• Specify probabilities of each outcome at each stage.
• Specify the effect of each outcome on expected project
cash flows (the discount rates may also vary at each
stage).
• Evaluate the optimal action to take at each stage in the
decision tree, based on the outcome at the previous
stage and its effect on cash flows and discount rate,
beginning with the final stage and working backwards.
• Estimate the optimal action to take at the very first
stage, based on the expected cash flows over the entire
project, and all of the likely outcomes, weighted by their
relative probabilities.

Example:
Home Depot is considering introducing a new in-home computer
shopping service. Knowing little about the business, they propose
to make the investment in three stages:
• A market test on very few consumers, at a cost of $1 million.
The likelihood of success is believed to be 75%.
• A partial introduction, over one year, if the market test is
successful. In this test, some widely used items will be put on-

BAFI 402: Financial Management I, Fall 2001 A. Gupta

Lecture: VI 8

line and marketed it to a wider audience. This much tougher

test (50% chance of success) would cost $2.5 million.
• Full introduction at a cost of $15 million, if the previous test is a
success. This will happen only by the end of the second year,
and the service will start generating revenues only by the end
of the third year. The three possible outcomes, along with the
revenues (million $) over 4 yrs (yrs 3-6) and probabilities are:
- Huge Success (12, 15, 18, 21, prob 25%)
- Moderate success (7, 9, 11, 13, prob 50%)
- Disaster (-2, -4, -6, -8, prob 25%)
Should Home Depot embark on this project?

• Construct a decision tree

• List all the outcomes, the cumulative probabilities of
each outcome, and the corresponding NPVs.
Outcome Probability NPV
Full investment: huge success 0.09375 $23,066
Full investment: moderate success 0.18750 $7,978
Full investment: disaster 0.09375 -$26,252
Partial investment: stop 0.37500 -$3,222
Stop after market test 0.25000 -$1,000
• Compute expected NPV across all outcomes.
• In this case, the expected NPV is -$261, hence Home
Depot should not embark on the project.

BAFI 402: Financial Management I, Fall 2001 A. Gupta

Lecture: VI 9

Limitations of Decision Trees:

• Decision trees provide a wealth of information to the
decision maker, but they also require a wealth of
information. Estimating all the outcomes and the
probabilities is very difficult when the product or
service is new or unique, and the firm has no past
experience of similar projects.
• The project has to be analyzed in stages, hence decision
trees are difficult to apply when investments occur
gradually over time, rather than in clear stages.
• The outcomes have to be discrete (e.g., drug will be
approved by FDA - yes/no). If the outcomes are not
discrete (e.g., how high will the sales be), they have to
be discretized into broader classes for analysis (e.g., the
sales can be high/medium/low).

Hence, decision trees are most likely to work when:

• The firm makes decisions on the project in clear stages.
• The outcomes at each stage can at least be classified into
broader classes.
• The probabilities of the outcomes and the effect on cash
flows can be estimated at the start; this usually implies
that the firm has done similar projects in the past.

Very often used by oil companies in evaluating opportunities to

drill for oil.

BAFI 402: Financial Management I, Fall 2001 A. Gupta

Lecture: VI 10

Project Options

One of the limitations of traditional investment analysis is

that it is static, and does not do a good job of capturing the
options embedded in the investment:
• The first option is the option to delay taking a project,
when a firm has exclusive rights to it, until a later date.
• The second option is taking one project may allow us to
take advantage of other opportunities/projects (option
to expand) in the future.
• The last option embedded in projects is the option to
abandon a project, if the cash flows do not measure up.

These options all add value to projects and may make a

“bad” project (from traditional analysis) into a good one.

BAFI 402: Financial Management I, Fall 2001 A. Gupta

Lecture: VI 11

The Option to Delay

• When a firm has exclusive rights to a project or product

for a specific period, it can delay taking this project or
product until a later date.
• A traditional investment analysis just answers the
question of whether the project is a good one if taken
today.
• Thus, the fact that a project does not pass the NPV or
the IRR criteria today, does not mean that the rights to
this project are not valuable.
• Having the exclusive rights to a product or project is
valuable, even if the product or project is not viable
today.
• The value of these rights increases with the volatility of
the underlying business (i.e. uncertainty).
• The cost of acquiring these rights (by buying them or
spending money on development - R&D, for instance)
must be weighed off against these benefits.

Example: (Valuing a drug patent)

A pharmaceutical company has the patent rights to a new ulcer
drug for 20 years. Although the drug shows promise, it is still
very expensive to manufacture ($500 mil.) and has a small market
(PV of cash flows from introducing the drug now is $350 mil.). The
technology and the market are volatile, and the annualized
variance in the PV is 0.05. How valuable are these patent rights?

BAFI 402: Financial Management I, Fall 2001 A. Gupta

Lecture: VI 12

• The NPV of introducing the drug is negative, but the

rights to this drug may still be valuable due to the
uncertainty in the PV of the cash flows.
• This drug may be profitable some years from now (we
cannot rule that out due to the uncertainty).
• If this project becomes profitable, the firm has exclusive
rights to exploit this opportunity - otherwise, the firm
just lets the patent expire, which costs nothing
(remember, R&D costs are sunk costs!).
• Hence, these patent rights give the firm an upside
potential with some positive probability, and no
downside potential at all!
• Therefore, this possibility of a profit in the future, with
no possibility of a loss, must have some value.
• This value (of the patent rights) is the value of the option
to delay investing in the project till it is profitable.
• If the project doesn’t ever become profitable, the firm
would just let the option expire unexercised.
• The value of this (call) option, using option pricing
theory, comes to $51.02 million.
• Thus, this ulcer drug, which has a negative NPV if
introduced now, is still valuable to the owner!

Question: Would you offer to pay $51.02 million for the rights
to this drug today? Why/why not?

BAFI 402: Financial Management I, Fall 2001 A. Gupta

Lecture: VI 13

The Option to Expand

• Taking a project today may allow a firm to consider and

take other valuable projects in the future.
• Thus, even though a project may have a negative NPV,
it may be a project worth taking if the option it provides
the firm (to take other projects in the future) provides a
more-than-compensating value.
• These are the options that firms often call “strategic
options” and use as a rationale for taking on negative
NPV projects.

Example:
• Disney is considering investing $100 million to create a Spanish
version of the Disney channel to serve the growing Mexican
market.
• The PV of cash flows from this investment is only $80 million,
i.e., the new channel has a negative NPV (of -20 million).
• If the Mexican market turns out to be more lucrative than
currently anticipated, Disney could expand its reach to all of
Latin America with an additional investment of $150 million,
any time over the next 10 years. While current expectations are
that the PV of cash flows from having a Disney channel in Latin
America is only $100 million, there is considerable uncertainty
about both the potential for such a channel and the shape of the
market itself, leading to significant variance in this estimate.
Should Disney open the Spanish channel in Mexico?

BAFI 402: Financial Management I, Fall 2001 A. Gupta

Lecture: VI 14

• Disney can only expand into Latin America if it first

tests the channel in Mexico, i.e., not entering the
Mexican market rules out expanding into Latin
America.
• If Disney enters the Mexican market, it also gets the
option to expand into Latin America - if the market
turns out to be good, it can make profits by expanding,
otherwise, it just sticks to the Mexican market.
• Again, once having invested in Mexico, the firm has upside
potential (profits from Latin America if scenario turns
out to be good), with no downside potential (don’t
expand if things are not good!).
• This positive probability of a profit, with no probability
of losses, must have value.
• The value of this option to expand into Latin America
only arises if the firm enters Mexico.
• Therefore, the value of this option must be considered
as a benefit of taking up the Mexican project.
• Using option pricing theory, the value of this option to
expand is $45.9 million.
• Therefore, the NPV of the project (Disney channel in
Mexico), with the option to expand, is:
NPV = -20 million + $45.9 million = $25.9 million
• Therefore, take the project!

BAFI 402: Financial Management I, Fall 2001 A. Gupta

Lecture: VI 15

The Option to Abandon

• A firm may sometimes have the option to abandon a

project, if the cash flows do not measure up to
expectations.
• If abandoning the project allows the firm to save itself
from further losses, this option can make a project more
valuable.

Example:
• Disney is considering taking a 25-yr project which
- requires an initial investment of $250 million in a real estate
partnership to develop time share properties with a South
Florida developer,
- has a PV of expected cash flows of $254 million.
• While the NPV of $4 mil. is small, assume that Disney has the
option to abandon this project anytime in the next 5 years by
selling its share back to the developer for $150 million.
• The variance of the PV of cash flows is 9%.
How should Disney value this project?

BAFI 402: Financial Management I, Fall 2001 A. Gupta

Lecture: VI 16

• There is uncertainty about future cash flows.

• Disney has a fallback alternative - if things turn out to be
very unfavorable in the next 5 years, they can still
recover $150 million from the developer.
• There is no downside to this fallback alternative - if
things are good, they continue with the project, if things
are bad, they abandon the project and recover $150
million (remember, we don’t know which one of these will
happen! - in all these cases, its the uncertainty that gives
value to these options - in a world with perfect certainty, we
know what would happen, so there is no “optionality”).
• Therefore, Disney has a (put) option that must be
valued, and added to the NPV of the project.
• In the next 5 years, if the PV of the expected cash flows
from the project drops below $150 million, Disney will
abandon the project and collect $150 million.
• The value of this abandonment option can be estimated
to be $7.86 million (using option pricing theory).
• This option value ($7.86 mil.) must be added on to the
NPV of the project of $4 mil., yielding a total NPV of
$11.86 mil.

BAFI 402: Financial Management I, Fall 2001 A. Gupta