UNIT ONE- RISK ANALYSIS AND THE OPTIMAL CAPITAL

BUDGET

To find competition and improve operational efficiency and profitability,

companies add new lines of business to existing projects or investments. The
projects can be expansionary projects or replacement projects or both.

For most of such projects the cash flows are uncertain, and this uncertainty
creates risk.

In this unit, we discuss procedures used to assess risk, incorporate risk in the
capital budgeting process and conclude by discussing how firms determine
optimal capital budgets.

Introduction to Project risk

Risk analysis is important in all financial decisions, especially those relating to

capital budgeting.

The higher the risk, the higher the rate of return required to compensate for the
higher risk.

Project risk can be assessed from three separate and distinct levels of risk
analysis, which are:

1) Stand – alone risk, which views a project’s risk in isolation, that is without
regard to portfolio effects. A project is evaluated only on its own merit.

2) Within firm risk, also called project’s corporate risk, which views project
risk within the context of the firm’s portfolio of already existing and running
projects.

3) Project’s market risk, which views the project risk in the context of the
firm’s stockholders holding diversified portfolio of stocks.
The coefficient of correlation is a very crucial measure of risk, because it
moderates the level of risk that a specific project contributes to corporate risk
and market risk. Coefficient of correlation measures the size and direction of the
relationship between two variables.

Definition of terms used in risk analysis

σP – standard deviation of the rate of return of the project in question, measured

as the standard deviation of the project’s IRR or standard deviation of the
project’s NPV. It is a measure of the project’s stand alone risk. We shall discuss
four techniques of measuring stand alone risk.

σF - standard deviation of the rate of return on the firm’s assets before the firms
takes on the project under viability evaluation. Rate of return in this context is
measured as [(Net Income + Interest)/(Debt + Equity)]. If the company has been
in existence for sometimes (σF) is based on average returns realised by existing
projects over the past few years. A high corporate risk ( σF) implies high chances
of bankruptcy.

The estimate of a corporate risk for the past years can be done statistically, but
changes in the firm’s situation may make its expected future corporate risk to
differ from its past levels, in which case subjectively estimated ( σF) would be
preferable for use in the analysis.

Remember that when using historical returns this way, we are assuming that the
historical returns will prevail into the future thus they are a good prediction the
returns that will be generated in the future.

σM - standard deviation of the market’s returns. Market returns are average

returns of national stock market index such as the Zimbabwe Stock Exchange
(ZSE).
rP,M - correlation coefficient between the rate of return on the project and the
return on the stock market index. Are the two returns dependent or independent
of each other? Most projects’ returns are positively correlated with returns in the
market/economy.

rP,F - coefficient of correlation between the rate of return on project and the
average rate of return on the firm’s other assets. Are the two returns dependent
or independent of each other? Are they negatively or positively correlated? Is
the project related to the firm’s core business or it is outside the core.

bP, F - project’s corporate beta coefficient – project’s corporate beta is a measure

of the project’s contribution to the firm’s corporate risk. It is found by
regressing the project’s expected returns under different economic states against
expected returns on firm’s other assets.

bP, M project’s market beta. It is a measure of the contribution of the project to

the risk borne by the firm’s stockholders, who are assumed to hold diversified
portfolios. The beta could be found by regressing the project’s returns against
returns on the market index under different economic states.

From the above equations, we can notice that the project’s stand alone risk ( σP)
is also an input variable in the formulae for project’s corporate beta and
project’s market beta.

Contrary to most people’s believe that stand alone risk and corporate risk are
not important to stockholders whose aim is to maximise their wealth, stand
alone and corporate risk are important for the following reasons:

1) Undiversified stockholders, including owners of small businesses, are more

concerned about corporate risk than about market risk.
2) Corporate risk is a good indicator of the likelihood of financial distress, so it
is an important risk measure to many company stakeholders including
customers, suppliers, creditors etc; and

3) Firm’s stability is important to other stakeholders, including managers,

customers, suppliers, creditors and the community in which it operates. Firms
that are in serious danger of bankruptcy, even of suffering low profits and
reduced output, have difficulty attracting and retaining good managers and
workers. Also, both suppliers and customers are reluctant to depend on weak
firms, and as such firms have difficulty borrowing money except at higher
interest rates.

A. Stand – Alone Risk

As we have seen, stand alone risk σP is an important determinant of both

corporate and market risk therefore, it is still critical for the company to
correctly estimate its stand alone risk.

The starting point for analysing a project’s stand alone risk is to determine the
project’s cash flows.

The uncertain cash flows can be determined through informal judgements or

through complex economic and statistical analytical tools or through a
combination of both methods.

In capital budgeting, most of the individual cash flows used to produce the
yearly expected cash flows are subject to uncertainty and the level of
uncertainty must be quantified and this can be done through using four stand
alone risk assessment techniques – sensitivity analysis, scenario analysis, Monte
Carlo simulation and decision trees.
Before project risk can be quantified be it on a stand- alone, corporate level or
market level, cash flows have to be estimated as we did on the excel for the RIC
watering project.

The input variables and how to calculate the cash flows and the resultant NPV
and IRR on excel sheet were discussed in class and excel sheet is available on
Tsimeonline. It was noted that most variables for estimating cash flows – sales
units, selling price, variable costs etc – are uncertain, and that when input
variables deviate from expected values, the yearly cash flows as well as the
NPV and IRR will also change.

As written on the excel sheet - the sales units, selling price and variable costs,
etc – are the expected values taken from respective probability distributions of
the input variables. The distributions of the input variables can either be
relatively tight reflecting small standard deviation and low risk or they can be
flat denoting a great deal of uncertainty about the variables under consideration.

Therefore, the variables’ distributions and their correlations with one another
affect the project’s NPV and IRR distributions, thus the project’s stand alone
risk. Note that anything that affects a project’s stand alone risk also affects a
project’s corporate and market risk because σP appears in the formulae for
calculating project’s corporate and market betas.

Techniques for assessing stand – alone risk

There four techniques for measuring stand -alone risk (sensitivity analysis,
scenario analysis, Monte Carlo simulation and decision trees)

I. Sensitivity Analysis

It is a technique which indicates exactly how much the NPV or IRR will change
in response to a given change in a key input variable, other things held constant.
Sensitivity analysis begins with a base case situation developed using each
input’s expected value. For example, the values for unit sales, sales price, fixed
costs, and variable costs are the expected or base case values and the resulting
NPV $12 076 984 and IRR 25% are the base case NPV and IRR (see excel
sheet).

Sensitivity analysis is concerned with finding out how sensitive the NPV is to
changes in key input variables. For example, in the RIC’s watering project the
analysts will have to identify key input variables that affect NPV of the project
the most, and if possible take corrective measures such as entering into hedging
contracts to mitigate against unfavourable movements in key input variables.

Please note the word input variables in this context is used to refer to the several
variables that are added and subtracted to arrive at annual cash flows that are
used in capital/project appraisals.

In sensitivity analysis, the level of stand- alone risk is denoted by the gradient of
the sensitivity lines. The steeper the sensitivity line, the greater the risk.

Sensitivity analysis testing the sensitivity of NPV to various input variables

within the same project or across a number of projects. For example, through
the later approach the analyst can be able to tell among the projects being
evaluated, which project for instance is very sensitive to changes in sales units,
selling price, etc?

Limitations of sensitivity analysis.

Firstly it ignores some useful information and the possibility of using

contemporary risk hedging techniques. For example, a proposed coal mine can
have an NPV which is highly sensitive to changes in output and sales price.
Sensitivity analysis will result in a decision to reject the project. However, the
company might have hedged against the risk or if it has not yet done so, the
company might actually be having at its disposal available options to hedge
against the risks of output and sales price fluctuations, in which case the mining
venture might be safe despite the steep sensitivity lines.

Secondly, sensitivity analysis ignores the range of likely values of the key input
variables as reflected in the inputs’ probability distributions.

II. Scenario Analysis

Sensitivity analysis does not take into take into account the range of likely
values of key variables as reflected in input variables’ probability distributions.

Scenario analysis is a technique that considers both the sensitivity of NPV to

changes in key input variables and the range of likely input values.

It incorporates 3 discrete NPV values:

1) Worst case scenario NPV – in which the least possible NPV is calculated
using lowest possible unit sales, lowest possible selling price, the highest
possible variable costs, highest possible construction costs and so on.

2) The average or the most likely NPV. The NPV of $12.076 million was
calculated assuming all input variables take their expected values.

3) Best case NPV – which could be obtained assuming the highest possible unit
sales, highest possible selling price, lowest possible variables and construction
costs etc.

To illustrate using the RIC watering system project, the company might be very
confident about all other input variables except sales quantity and selling price.
For the project, the lowest possible unit sales figure is estimated to be 15 000
units and the maximum to be 35 000 units.

On the other hand, the lowest possible price is estimated to be $1700 and the
best price is to be $2700.
To do scenario analysis, the company will have to calculate the NPV expected
under the 3 scenarios. Each scenario has a specific probability of occurring. The
table below shows how the NPVs under the three scenarios are treated to
calculate the expected NPV, NPV standard deviation and NPV coefficient of
variation.

To get the NPV under the worst case scenario, go to the excel sheet and change
the sales figure in year 2021 to 15 000 units and the selling price to $1 700. All
other variables remain the same, because we have assumed that the managers
are confident that the figures as they are correct estimates.

To get the NPV under the best case scenario, again go to the excel sheet and
change the same sales units cell to 35 000 and the same selling price cell to $2
700.

Scenario Sales Volume Sales Price NPV Probability of Product

outcome

Worst Case 15 000 $1700 ($11 447) 0.25 ($2861)

Most likely 25 000 $2200 $12076 0.50 $6,038

case

Best case 35 000 $2700 $43 575 0.25 $10,890

Expected NPV $14 070

Standard Deviation of NPV $19,555

Coefficient of variation of NPV [$19 555/$14

070] = 1.4

To get the expected NPV, sum the final column.

You find the standard deviation as follows:

Variance = 0.25(-10, 79-13,956)^2 + 0.5(12,076 – 13,956)^2 + 0.25(41,752 –
13,956)^2

Standard deviation = (Variance) ^0.5

Standard deviation = $18,421

Scenario analysis provides managers with information on the maximum

possible loss that the company can suffer under the worst case. For example in
the table above, the maximum loss possible is $10.079 million. The company
will have to make an assessment of its current financial position and see if it can
be absorb and withstand a loss of this magnitude without endangering the
company to dire financial consequences.

A closer look at project’s stand alone risk can be made since scenario analysis
gives managers the E (NPV), standard deviation of NPV, and coefficient of
variation.

The coefficient of variation can be compared with the average coefficient of

variation of the company to get an idea of the project’s relative stand-alone risk.

Two major limitations

1) Firstly, the analysis considers only 3 discrete NPVs for the project (worst
case NPV, most likely NPV and the best case NPV); although in reality there
are an infinite number of possibilities.

2) Secondly, scenario analysis naively assumes that the variables treated as

uncertain are perfectly correlated implying that all inputs are at their worst or
best levels at the same time. In reality, the probability that all worst case and all
best case inputs would occur at the same time is extremely low. Therefore,
scenario analysis typically overstates the extremes – the worst case NPV is too
low, and the best case NPV is too high.
Monte Carlo simulation

 It ties together sensitivities and probability distributions of input variables.

 The first step in Monte Carlo Simulation is to create a model that develops
project cash flows and that calculates NPV.
 The next step is to specify the probability distribution for each uncertain
input variable, that is, must specify a probability distribution for sales units,
and for other input variables. Also specify the correlations between input
probability distributions.
 With the two steps above in place, the Monte Carlo Software chooses at
random a value for each uncertain variable, based on its specified probability
distribution.
 The values generated for each input variable along with fixed inputs such as
tax rate and depreciation charges, are then used by the model to determine
the yearly net cash flows which are used to determine the project’s NPV for
that computer run
 Steps 3 and four above are repeated many times eg 1000 times, resulting in
1000 NPVs which are then used to form a probability distribution with its
own expected NPV and standard deviation. Other NPV distribution statistics
such as maximum NPV, minimum NPV, NPV range, probability of NPV
being greater than zero, coefficient of variation etc. are also generated by the
Monte Carlo software.
 For the RIC project, let us assume that the selling price follows a normal
distribution with mean $2,200 and standard deviation $167 and we specify
this in the system.
 Lets also assume that the sales units will follow a triangular distribution with
mean figure of 25 000 units, highest figure of 40,000 units and lowest figure
of 15,000 units. We assume all other variables to take their expected values.
 After 1000 runs, the software can generate the following NPV distribution
statistics for the RIC project.

 Expected NPV $12,096

 Maximum NPV $46,755
 Minimum NPV ($13,888)
 NPV range $60,643
 Probability of NPV>0 85.7%
 Probability of NPV<0 14.3%
 Coefficient of variation $10 724/12 096 = 0.89

Limitations

1. The major problem is specifying each uncertain variable’s probability

distribution and the correlations among the distributions.
2. The other problem with both scenario and simulation analysis is that there is
no clear cut decision rule. We end up with expected NPV and the relative
distribution with no clear cut criterion to indicate whether a project’s
profitability as measured by its expected NPV is sufficient to compensate for
its risk as measured by CVNPV and standard deviation of NPV.
3. Finally, scenario and simulation analysis both focus on project’s stand alone
risk, ignoring the effects of diversification both among the projects within
the firm and by investors in their personal investment portfolios. For
example, a project can have high stand-alone risk, but if these returns are not
correlated with returns on the firm’s other assets or with the market in
general, then the project might not be very risk in terms of either corporate
risk or market risk.

Decision trees
 Its emphasis is both on measuring and reducing project risk.
 Sometimes it is not feasible for the company to make an outright investment
outlay at time zero. Instead, the firm can stagger the capital outlays over a
period of time such that the outcome at each stage affects the next stage. The
project can be abandoned if an earlier stage shows that the project is not
feasible, hence avoiding further investment on non-viable projects thus
minimising overall company losses.
 For example suppose company X is considering the production of an
industrial robot for the television manufacturing industry. The net investment
for the project, of $11.5 million, can be broken into 3 stages:

Stage 1

 At t = 0, the marketing department conducts a $500,000 study of the

market potential for robots in television assembly lines. The probabilities
of their being market or no market for the product are shown in the
decision tree.

Stage 2

 If it appears that a sizeable market exists, then at t = 1, the company

spends $1,000,000 to design and fabricate several prototype robots. The
robots would then be evaluated by television engineers, and their reaction
would determine whether the firm should proceed with the project. The
probabilities of proceeding or stopping at this stage are shown on the
decision tree diagram.

Stage 3

 If reaction to the prototype robots is good, then at t = 2, the company

proceeds to build a production plant at a net cost of $10,000,000. If this stage
were reached, the project would generate high, medium, or low net cash
 flows over following 4 years. The probabilities of having high, medium and
low net cash flows are shown on the decision tree diagram.

The Cash inflow stages

 Under high cash flows, the company would have net cash flows of $10
million per year from year 3 to 6.
 Under medium cash flows, the firm would have net cash flows of $4 million
from yr 3 to 6
 Under low cash flows, the company will have negative net cash flows of $2
million in yr 3 through yr 6

The company uses a discount rate of 11.5%.

The figure attached shows how the company can do a decision tree analysis for
the project.

 Since the expected NPV is negative, it would appear that the project
should be rejected.

 The coefficient of variation is very large of (7,991/ (338)) = -23.64,

suggesting the project is extremely risky in terms of stand- alone risk.

 There is also a 0.144 + 0.320 + 0.20 = 0.664 probability of incurring a

loss.

 Based on the above analysis, the project would be totally unacceptable.

 However, the use of decision trees allows the incorporation of the project
abandonment option in the capital budgeting process.

 In the third branch of the attached decision tree, the company can
abandon the project at the beginning of year 4, if negative cash flows of
 $2,000,000 have been realised at end of year 3, and if based on latest
information available at that time, the company suspects the negative
cash flows to continue through year 6.

 The effect of abandoning the project at beginning of year 4 is to increase

the expected NPV from negative ($338 000) to positive $166 000. Also,
by abandoning the project at year 3 if net cash flows are negative, the
company can lower the standard deviation of NPV from $7,991,000 to
$7,157,000.

 Therefore, incorporating the abandonment option into the decision tree

analysis can change the project’s NPV and hence its ultimate
reject/accept decision.

 The abandonment option will change the branch to look like this:

3
0.3 Joint Prob NPV Prob X NPV

($2,000) 4 Stop 0.144 (10,883) (1,567)

Takeaways from decision tree analysis

 Managers can reduce project’s stand alone risk if they structure the
decision process to include several decision points rather than just one.

 By staggering the decision making points across the duration of the

project, decision tree analysis acknowledges that capital budgeting is a
dynamic process and makes possible the use of latest or current
information in the decision making process, thereby increasing the quality
of decisions made and minimizing level of losses incurred.
B. Project’s Corporate Risk
 Recall that a project’s corporate risk is the project’s contribution to the
firm’s overall corporate risk. Put another way, project’s corporate risk
reflects the impact of the project on uncertainty about the firm’s total cash
flows.
 A project with a corporate beta of 1.0 would be as risky as the firm’s average
asset; a beta greater than 1.0 would signify more than average risk; and a
beta less than 1.0 would signify low project’s corporate risk.

Formula: bP, F = [(σp/σF)]rP,F

 A negative project’s corporate beta means that the project’s returns move in
the opposite direction to the returns of the firm’s other assets/projects. It
implies that when the returns on a company’s existing projects/assets
increase, the returns of a project with negative corporate beta fall, and vice-
versa.

Estimating the corporate beta for the RIC watering project

 Let us assume that RIC is already a diversified company that wants to add
the watering control project line to its existing business lines that include
brick manufacturing, cement manufacturing and cattle ranching.
 One approach that RIC can use to estimate its watering project’s corporate
beta is to find the average historical returns - [(net income + interest) / (Debt
+ Equity)] - of a group of companies X, Y and Z, for example, whose only
line of business is the provision of watering services to the public. Using
this approach, RIC is taking average returns of pure play companies X, Y
and Z as a proxy of the likely returns that its watering project will generate
when commissioned.
 The average historical returns of the pure play companies will be regressed
against historical returns generated by the company A’s existing projects
(mentioned above) to get an estimate of RIC watering project’s corporate
beta.

Importance of incorporating project’s corporate beta in risk

analysis
 Making rejection/acceptance decisions merely on the basis of a project’s
high stand-alone risk may lead to overly incorrect decisions.
 A project’s corporate beta takes into account the potential of overall
corporate risk reduction that comes with holding a diversified portfolio of
projects/assets.
 However, even though this approach can give information to management
regarding the level risk of the envisaged project in relation to the average
risk of the company’s existing projects, the project’s corporate risk beta does
not provide the company with enough information to calculate a project
discount rate that commensurate with the risk of the project.

 This limitation is addressed by the incorporating project’s market risk in

capital budgeting.

C. Project’s market risk

 Recap: It is a measure of the contribution of the project to the risk borne by

firm’s stockholders who are assumed to hold diversified portfolios.
Formula: Formula: bP, M = [(σp/σM)]rP,M

RIC Watering Project’s market beta

 As stated previously RIC is a diversified company with three projects
currently running. The three projects constitute RIC’s current business are
contributing to the company’s current beta.
 Therefore, the RIC project if implemented will be the company’s fourth
business line. Let us also assume that RIC finances its capital investments
through a 50:50 combination of debt and equity. The risk free rate of return
is 8% and average return in the market is 13%. RIC currently has a beta of
1.8 on its stock. Cost of debt (Kd) is 10%. The total capital investment
invested in the existing projects is $103.2 million and the watering project
needs $25.8 million.
 It can be deduced that since the company has a market beta of 1.8 and that
since the company currently has three operational projects, the returns
generated by the existing projects account for the company’s overall beta of
1.8. Put in another way, the average of the existing projects’ individual betas
gives the company’s current stock beta of 1.8
 By adding the watering system to the company’s investment lines, the
company beta may change. The change will depend on the new project’s
market beta and the proportion of total capital invested in the new project.
 The immediate question that follows from the above discussion is “How do
we estimate the watering project’s market beta?”
From the above information we can calculate the following
a) The firm’s new stock beta
$103.2/$129*1.8 + $25.8/$129*2.5
=1.94
b) The firm’s current cost of equity Ke = Rf + (Rm - Rf)b
= 8% + (13% - 8%) 1.8
=17%
c) Firm’s new cost of equity = 8% + (13% - 8%)1.94
= 17.7%
d) The firm’s overall cost of capital before the new project
= 0.5*0.6*10% + 0.5* 17%
=11.5%
e) The company’s new WACC after adopting the watering project
= 0.5*0.6*10% + 0.5*17.7%
=11.85%
f) The new project’s WACC can be obtained using the following formula:

[Amount invested in existing projects/ total amount to be invested with new

projected included]*current firm’s overall WACC + [Amount to be
invested in the new project/ total amount to be invested with new projected
included]*new project’s WACC = Company’s overall WACC after with
new projected included.
0.8(11.5%) + 0.2X = 11.85%
0.2X = 13.25%

Methods of estimating a project’s market beta

We use two methods, which are the pure play method and the accounting beta
method.

Pure Play method

In the pure play method, the company tries to find one or more non-
integrated, single product companies in the same line of business as
the project being evaluated.

For example, RIC will have to find publicly traded firms that only
deal with watering control systems and we assume the pure play
companies have the same business risk as the watering project. RIC
should then determine the betas of these firms through the regression
process find the average of these betas and use it as a proxy for the
project’s beta.

However, betas of companies are also affected by capital structure

and corporate tax rates.

Therefore the average beta that we got from the regression process
will have to be adjusted for the differences’ in the capital structures
and tax rates of pure play companies versus the tax rate and capital
structure of RIC. The adjustment is not necessary if RIC has the same
capital structure and tax rate as the pure play companies.

The adjustment is done using the Hamada equation.

Hamada formula: bL = bU [1 + (1-T) (D/S)]

Making bU the subject, the equation changes to

bU = bL/ [1 + (1-T) (D/S)]

Where bU stands for beta of an unlevered firm, b L is beta of a levered firm, T is

the tax rate and D and S, respectively, stand for the proportion of debt and
equity in the capital structure.

To illustrate, assume that RIC’s analysts have identified three publicly owned
companies engaged only in the production and distribution of watering control
systems. Further, assume that average stock beta (note that because of one line
of business, the stock beta can also be called project market beta), is 2.23; that
their average debt to equity ratio, D/S is 0.4/0.6 = 0.67; and their average tax
rate is 36%.
We cannot conclude that the project’s appropriate leverage adjusted beta is
2.23, because we already know RIC’s D/S ratio of 0.5/05 =1 and tax rate of
40% are different from those of the pure play firms.

To adjust for the differences in financial leverage and tax rates, we use the
Hamada equation and proceed with the following four steps.

Step 1: Note that the proxy firms’ average stock beta of 2.23 reflects their
average D/S ratio of 0.67 and their average tax rate of 36%

Step 2: We insert the values in step 1 above into the Hamada equation to get the
3 firms’ unlevered beta.

bU = bL/ [1 + (1-T) (D/S)]

bU = 2.23/[1 + (1-0.36)(0.4/06)]

= 1.56

Step 3: We can now find what the proxy firms’ stock betas would be if they had
the same capital structure and tax rate as RIC.

bL = bU [1 + (1-T) (D/S)]

bL= 1.56[1 + (1 – 0.4)(0.5/0.50)

=2.50

Step 4: Determine the project’s cost of equity and its weighted average cost of
capital, using 2.5 as the estimate of its beta.

KE = Kf + (Rm - Kf)b

8% + (13% - 8%)2.5 = 20.5%

The RIC watering control systems average cost of capital WACC P = 0.5(10%)
(0.6) + 0.5(20.5%) = 13.25%
WACCp = WdKd(1 – T) + WeKe

The major limitation of the pure play method is that it can be difficult to find
pure play proxy companies.

In summary, through analysing the project’s market risk RIC can come up
with the following important conclusions:

 If RIC takes on the new project, its corporate beta would rise from 1.8
to 1.94;
 Its cost of equity would rise from 17% to 17.7%;
 Its corporate cost of capital would rise from 11.5% to 11.85%; and the
new project would have to earn at least 13.25% to avoid lowering the
RIC’s stock price.
 In appraising the watering project, the right discount rate to use is
13.25% not 11.85%.
 11.85% is the discount rate for average risk projects, given that beta of
the watering project which is high implying more risk, the appropriate
discount rate for use in appraising the watering project is 13.25%.

The accounting beta method

When it is difficult to find single product, publicly traded firms for the pure play
approach, then companies can use the accounting beta method.

The accounting beta entails regressing the company’s basic earning power
(earnings before interest and taxes/total assets) against the average basic
earning power for a large sample of companies.

The accounting beta method can also be used to different divisions and projects’
within firm risk.