6 Uncertainty and Risk Analysis

CHAPTER 6
UNCERTAINTY AND RISK ANALYSIS
6.1 Introduction
In this age of advancing technology, successful managers must make
informed investment decisions that determine the future success of their
companies by drawing systematically on the specialized knowledge, accu-
mulated information, experience and skills of many people. In evaluating
projects and making choices between investment alternatives, every manager
is painfully aware that he cannot and will not always be right. Management
pressure is increased by the knowledge that a company's future depends on
the ability to choose with a high degree of consistency those investment and
market opportunities that have a high probability of success even though the
characteristics of future events are seldom precisely known.
In the previous chapters in the text investment analyses all were consid-
ered to be made under "no-risk" conditions. That is, the probability of suc-
cess was considered to be 1.0 for each investment evaluated. This means
that by expressing risk and uncertainty quantitatively in terms of numerical
probabilities or likelihood of occurrence, where probabilities are decimal
fractions in the range of zero to 1.0, we implicitly considered that the proba-
bility of achieving projected profits or savings was 1.0 for investment situa-
tions evaluated. We are all aware that due to risk and uncertainty from innu-
merable sources, the probability of success for many investments is
significantly different than 1.0. When faced with decision choices under
uncertain conditions, a manager can use informal analysis of the risk and
uncertainty associated with the investment or he can analyze the elements of
risk and uncertainty in a quantitative manner. Informal analysis relies on the
decision makers experience, intuition, judgement, hunches and luck to
determine whether or not a particular investment should be made. The quan-
282
Chapter 6: Uncertainty and Risk Analysis
283
. titative analysis approach is based on analyzing the effects that risk

and
uncertainty can haveon an investment situation by,using a logical
and con_
sistent decision strategy that incorporates the effects or iist<
and uncertainty
into the analysis results. use of the quantitative analysis approach
should
aot be considered to imply that the iriformal analyiis considerations
of
experience, intuition and judgement are not needed. on the
contrary, the
purpose of quantitative analysis of risk and uncertainty
is to provide the
decision maker with as much quantitative information as possible
concern-
ing the risks and uncertainties associated with a particular investment
situa-
tion, so that the decision maker has the best possible information
on which
to apply experience, intuition and good judgment in reaching
the final deci_
sion. The objective of investment decision making from an
Jconomic view_
point under conditions of uncertainty is to invest available
capital where we
have the highest probability of generating the maximum possible
future
profit. The use of quantitative approaches to incorporate
risk and uncer-
tainty into analysis results may help us be more successful in achieving
this
objective over the long run.
No matter how comprehensive or sophisticated an investment evaruation
may be, uncertainty still remains a factor in the evaluation.
F,ven though rate
of return or some other economic evaluation criterion may be calculated
for
a project with several significant figures of accuracy
usinj the best available
cost and income data, the decision maker may still feei uneasy
about the
economic decision indicated because he or she knows the
assumptions on
which the calculations are based are uncertain. If the economic
evaluation
method used does not reflect this uncertainty then every assumption
built
into an economic analysis is a "best guess', and the final economic
result is
a consolidation of these values. Making decisions on
the basis of such ,,best
guess" calculations alone can be hazardous. Consider
a manager who may
select investment alternative 'A" with a 20vo RoR over inv-estment ..B,,
which has a 157o RoR based on the "best guess" RoR calculation
approach. would this decision be justifiable if the probability
of success of
alternative 'A" was 50va (or one chance in two) compared with probabil-
a
ity of success of alternative "B" of 90va? It is evident that the manager
needs some measure of the "risk" involved in each alternative
in addition to
the "best guess" or most likely rate of return results.
Tlrcre ore seve ral different approaches that can be used to quantitatively
ittcorporate risk and uncertainty into analyses. These inctuie sensitivity
analysis or probabilis_tic sensitivity analysis to account
for uncertainty
associated with possible variation in project paramercr;, and expected
284 Economic Evaluation and lnvestment Decision Methods
value or expected net present value or rate of return analysis to account for
risk associated with finite prubabiility of failure., The,use of, sensitivity anaty-
sis is advocated for most economic analyses and the use of expected value
analysis is advisable if finite probability of projectfailure exists. Sensitivity
analysis is described in the first half of this chapter and expected value
analysis is described in the second half.
Sensitivity analysis is a means of evaluating the fficts of uncertairul,^ on
investment by determining how investment profitability varies as the param-
eters are varied that affect economic evaluation results. Sensitivity analysis
is a means of identifying those critical variables that, if changed, could con-
siderably affect the profrtability measure. In carrying out a sensitivity analy-
sis, individual variables are changed and the effect of such a change on the
expected rate of return (or some other decision method) is computed. Once
all of the strategic variables have been identified, they can be given special
attention by the decision maker. Some of the typical investment parameters
that often are allowed to vary for sensitivity analysis include initial invest-
ment, selling price, operating cost, project life, and salvage value. If proba-
bilities of occurrence are associated with the various levels of each invest-
ment parameter, sensitivity analysis becomes probabilistic sensitivity
analysis.
It may now be evident to the reader that the term "uncertainty" as used in
this text refers to possible variation in parameters that affect investment
evaluation. "Risk" refers to the evaluation of an investment using a known
ntechanism that incorporates the probabilities cf occurrence for success
and failure and/or of dffirent values of each investment parameter. Botll
uncertainty and risk influence almost all types of investment decision, but
especially investment involving research and development for any industry
and exploration for minerals and oil or gas.
6.2 Sensitivity Analysis to Analyze Effects of Uncertainty
As described in the previous section, sensitivity analysis refers to analysis

of how investment profitability is affected by variation in the parameters
that affect overall profitability. For a case where rate of return is the eco-
nomic criterion used to measure profitability, sensitivity analysis involves
evaluation of how rate of return varies with parameters such as initial
investment, profit per year, project life, and salvage value. It is frequently
Chapter 6: Uncertainty and Risk Analysis 2As
used to determine how much change in a variable would be necessary to

reverse the decision based on average-value or best-guess estimates. It usu-
ally does not taiie into consideration the likelihood of variation. The rate of
change in the total outcome relative to the rate of change in the variahle
being considered u'ill indicate the significance of this variable in the overall
evaluation.
Example 6-1 will introduce a single variable sensitivity analvsis. tri is
important for the reader to keep in mind that in this analysis, no dor.r'n-
stream effects are considered relevant to the evaluation. In other words, each
parameter is independent of the other so changing the magnitude of the cap-
ital investment will have no impact on any other operations or the magni-
tude of project operating costs, etc. To further illustrate, when the years of
income are reduced, there is no adjustment in the residual value of the
assets. Such numbers may increase or decrease but again, are neglected here
to simplify the introduction of this evaluation procedure.
EXAMPLE 6-1 Single Variable Sensitivity Anatysis

Annual profits of $67,000 are shown on the time diagram for ihis
$240,000 investment case with an expected salvage value of
$70,000 after five years. Evaluate the sensitivity of project RoR to
plus or minus 2o'h and 4oo/" varialions in initial investment, annual
profit, project life and salvage value.
Profits
C=$2a0,000 $67,000 $67,00Q $67,000 $67,000 $O7,OOO

L=$70,000
Solution:
Using the most expected cost and revenue parameters gives:
PW Eq: 240,000 = 67,000(P/A;,5) + 70,000(p/F;,5)
The "most expected" project ROR is 18%. How will this ,,most
expected" 18% ROR vary as parameters are changed?
A) lnitial lnvestment Sensitivity Analysis

lnitial 'Change in Percent Change ih 18.0%
investment Prediction ROR ROR Prediction
144,000 -40 42.0 133.3
192,000 -20 27.5 52.9
240,000 0 18.0 0
288,000 +20 11.2 -37.9
336,000 +40 5.8 -67.7
The percent variations in the ROR from changes in initial investment
costs are very significant. ln general, changes in parameters close to
time zero (such as initial investment and annual profit) have a much
more significant effect on investment ROR than changes in parame-
ters many Vears in the future from time Zero (such as salvage value).
B) Project Life Sensitivity Analysis

Project Change in Percent Change in 18.0%
Life Prediction ROR ROR Prediction
3 -40 5.6 -68.8
4 -20 13.4 -25.5
5 0 18.0 0
6 +20 20.9 16.3
7 +40 22.9 27.1
Note that this sensitivity analysis really involves changes in total
cash flow as well as project life. lf project life was longer (say 10 years
or more), changes in life would have a less sensitive effect on ROR.
C) Annual Profit Sensitivity Analysis

Annual Change in Percent Change in 18.0%
Profit Prediction ROR ROR Prediction
40,200 -40 3.6 -80.2
53,600 -20 11.0 -39.0
67,000 0 18.0 0
80,400 +20 24.8 37.9
93,900 +40 31.5 74.8
Chaorer 6: Unce(ainty and Risk Analysis 287
Percent variations in ROR due to changes in annual profit are very

significant beciuse the changes start occurring close to time zero.
lndividual parameters such as selling price. production rates and
ope;ating costs affect profit.
D) Salvage Value Sensitivity Anaiysis
Salvage Change in Percent Change in 18.0%

Value Prediction ROR ROR Prediction
42,004 -40 15.9 -11.9

56,000 -20 16.9 - 6.0
70,000 0 18.0 0
84,000 +20 19.0 5.4
98,000 +40 20.0 10.8
Sensitivity analysis shows that accuracy of salvage value is the

Ieast important of all the parameters that go into this ROR analysis
because salvage value occurs far in the future from ti,-r:e zero. Also,
in this case, cumulative salvage doliar value is small compared to
cumulative profit. ln some evaiuatjons this is not the case and sal-
i,age value has a much more sensitive effect.
6.3 The Range Approach to Sensitivity Analysis
The range approach involves estimating the most optimistic and most
pessimistic vzrlues (or the best and worst) tor each thctor in addition to esti-
mating most expected values. This approach will make investmcnt decision
ruraking easier fbr the cases where (1) a project appeats desirable even when
pessimistic values are used and therefore obviously should be adopted from
an economic viewpoint, (2) when a project appears to be undesirable even
when optimistic values are used and therefore rejection is dictated on eco-
nomic grounds. When a project looks good with optirnistic values but bad
with pessimistic values further study of the project and the risk and uncer-
tainty surounding the project should be made. Application of this method
is shown in Example 6-2.
EXAMPLE 6-2 Range Approach Sensitivity Analysis
Use the range approach to evaluate the investment described in

Example 6-1 for best and worst case sensitivity analysis using the
plrrs or minus 2Oo/" parameter variations with a five year life and a
minimum rate of return of 15"h.
Solution:
Best Case Expected Case Worst Case

lnvestment 192,000 240,000 288,000
Annual Profit 80,400 67,000 53,600
Salvage 84,000 70,000 56,000
Project Life in yrs. 5 5 5
ROR, % 36.4 18.0 3.7
The results indicate that the project is satisfactory for the best and
most expected conditions but unsatisfactory for the worst conditions.
More information is needed on the expected probability of occur-
rence of the worst case conditions to reach a valid and meaningful
decision.
The best, worst and most expe-cted sensitivity analysis results give very
useful information that bracket the range of project ROR results that can
reasonably be expected. This is the type of information that managers need
to reach investment decisions. It is very important to recognize that
although a project ROR greater than the minimum rate of return is predicted
for the most expected parameters, the result is based on parameters which
are subject to variation. This variation should be analyzed over the full
range of possible results utilizing the best engineering and management
judgments of people involved with a project.
6.4 Probabilistic Sensitivity Analysis

The application of probability distributions to relate sales volume and
prices, operating costs and other parameters to probability of occurrence
Chaoter 6: Uncertainty and Bisk Analysis 289
permits "probabilistic analysis" by the Monte Carlo simulation technique. A

brief description of the method follows, then the method will be illustrated.
A weakness of traditional tecirniques for the evaluation of projects is an
inabilitv to combine ilfcrmation fron-, a number of sources into a straight-
forivard and reliable profitability indicator. The major factor in this problern
is the large number of variables that must be consi.dered. Ai: additic,nal fac-
ior is that it is not possible to obLain accurate single-valued estirnates of
n:anr oi the r,ariable:.
Probability theory is the study of the uncertainty of events. A basic tool
of probability theory is the use of a range of values to describe variables
that cannot adequately be quantified by single value estimates. For example,
the determination of the least, greatest and most likely values of a variable
will more accurately quantify the variable than will the average value.
The distribution and relative possibilities of values assigned to a given
variable will remain characteristic of that varjable if factors affecting the
variable remain constant. Figure 6-l illustratejs three possible distributions
of values. The values are plotted on the horizontal axis and the respective
probabilities of their occurrence on the vertical axis. Analysis of the three
examples indicates that the uncertainty of the parameter described in Exam-
ple 3 is much greater than that in Example 2. The uncertainty of Example 3
is indicated by the w'ide range of vllues.
A m:{ority of the parameters in ieal evaluations otien give an intermecii-
ate ranse of parameter values as illustrated by Example l. Ideally we would
like to have a very smirll range tor all piuameters as illustrated by Example
2. In practice we get a combination of small, intermediate and large ranges
of parameter value variation for different parameters in actual evaluation sit-
uations.
The drstributions in Figure 6-1 all have their values symmetrically dis-
tributed around the most likely value (the familiar bell-shaped curve is an
example of this distribution). A distribution of this type is called the normal
distribuiion. If the most likely value is shifted to either side of the center of
the distribution, then it is refered to as a skewed distribution. The nonnal
distributiotr and the skewed distribution are botlt special types of a general
cLass of cttrves known as densitl, functions. In rhe remainder of this discr,ts-
sion normal distributions will be referred to as such and the terru density
function will be used to describe probability distributions that are not nor-
mal. The shape of the distributions will be determined by the nature of the
variables they describe.
uJ trl
O o
zu
zul - MEAN
(f
cE
tr
(r ,r'
f :)
o
() o
o
u- IL
o
tJ F
)d]
=
co co
o o
E
cc
(L o-
INTERMEDIATE RANGE
tl
UL LL SMALL RANGE
EXAMPLE 1 EXAMPLE 2
uJ
o
zul
0c
c(
f
o
o
IL
o
F
-)
6
c0
o
(r
(L
PARAMETER VALUES,
LARGE RANGE
EXAMPLE 3
Figure 6-1 Frequency Distribution Graphs lllustrated
(LL = Lower Limits; UL = Upper Limits)
In principle, a relative frequency distribution graph is converted to the

equivilent cumulative frequency distribution graph by moving from the left
.nO of the distribution to the right end and computing the total area that is
less than or equal to corresponding values of the parameter within the range
evaluated. The cumulative area to the left of a given parameter value divided
by the total area under the curve is the cumulative probability that a random
parameter value ivill be less than or equal to the given parameter value.
Figure 6-2 illustrates typical cumulative frequency distribution graphs
that would result from converting the relative frequency graphs shown in
Figure 6-1 to the corresponding plots. Note that Example 3 with a large
range of parameter values has a much flatter cumulative probability curve
than Example 2 which has a small range of parameter values. Example 2
Chapter 6: Uncertainty and F{isk Analysis
has a much higher percentage of the parameter values in a given

range of
curnulative probabitity of occurrence compared to Example 3.-
1.0 1.0
no
0.9
,_a
i 0.7
0.8
0.7
,! ,:1
:C 0.6 0.6
o-!
oi
>x 0.5 0.5
=*
E
o.4 0.4
f
O 0.3 0.3
4.2 0.2
0. 1
0.1
0.0 0.0
Pa:a;'le'er Values i-: Para;rete:- ', a jues tJL
lnt€rinei;ate fiange Small Range
Example 1
Example 2
'1.0
0.9
=
io'
4.7
.c.6
oI
.>; 0.5
6LJ
E= 0.4
tr
=
O 0.3
0.2
0. 1
0.0
LL parameter Values UL
Large Range
Example 3
Figure 6-2 Cumulative Frequency Distribution

(LL = Lower Limit, UL = gppsr Limit)
Now to describe the probabilistic sensitivity analysis approach, consider that
we are about to evaluate a new development project. we migtrt be interested
in
the effect on our economic evaluation criterion of changes in prolect
parame-
ters such as initial investment, product selling price per unit, priaucdon
rate or
sales projections, operating costs, project life and salvage value. For each of
these variables and/or other variables to be investigated, a frequeney distribu-
tion plot of probability of occurrence versus para=meter vatue similar ,o tr,"
examples in Figure 6-l is prepared by the person or persons most familiar with
and capable of projecting future values of the parameter involved. These fre-
quency distribution data are then converted to cumulative probability of occur-
rence versus parameter value graphs similar to the examples in Figure 6-2.
when these graphs are available for each parameter that is considered to vary,
the use of Monte carlo simulation is applied. This generally involves curve-fit-
ting a mathematical expression to describe each cumulative probability of
occrurence versus parameter value so that picking a random number between
0.0 and 1.0 analogous to the cumulative probability of occurrence will auto-
matically fix the parameter at the corresponding value. Different random num-
bers are selected to fix each of the different parameters being varied. Then
using the randomly selected parameter values, the economic analysis criterion
such as rate of return is calculated. Then you iterate and do the same thing over
again picking a new set of random numbers to determine a new set of parame-
ters used to calculate rate of retum. This is done over and over again for some-
where between 100 and 1000 times and then a histogram (frequency distribu-
tion plot) of these rate of return results versus probability of occurrence is
prepared. The number of iterations (RoR results) that are required is the num-
ber that will give the same shape of frnal RoR results frequency distribution
graph that would be obtained if many more iterations were made. To illustrate.
consider hypothetical cases I and 2 shown ih Figure 6-3.
30
s \o
() 25
8zs
o o
20 :20
o
o
15 o15
o
't0
=
o Ero
6 a6
o 5 -o-
o5l
o- (L1
0
5 't0 15 20 25
Rate of Return, 7o Flate of Return, 9,o
100 Simulations 500 or 1000 Simulations
(Note multimodal maxima) (Constant shape, unimodal)
Case 1 Case 2
Figure 6-3 Frequency Distribution of RoR for Varying simulations

Cnapter 6: Uncertainty and Risk Analysis
293
'rhe
histograrn fbr case t has murti-modar (more
than one maxima) peaks
indicating that statisticalry we have
not mad-e enough simurations to get ,a
rate of return frequency distribution
with constant shape. If the input data
frequency clisuibution is unimodal (having
one peak as'iitusirated in Figure
5-l; the ourtrrut rare of return freirue,cy ciistribuiion
plot str,;urd be uni_
mcdal. case 2 iiemonstrates for this hyporheticar
situation thar if we to to
5il0 or 1000 simulations we g"t a unimtdar
graph *.ith the same shape indi_
;;iting that 500 simulations are sufiicient
fbr anaryzing this hypotheticar
iuresrment using probabilistic sensitivity
analysis.
An iniportant thing to.note about probabilistic
sensitivity anarysis resurts
is that y'ou i-ro not gei a single resrrt,iut
instead you get a iarg" over which
the results vary as a function or prouaurtity
of occunJr""aJ*so you get a
m'st expected result. In many iuvestmenisituations
is more important rt,
tt.,tup" ot this curve
varue. For exampri
a most expecred RCR ofln:-ryl,rr.*p..i.o
25vc with ara,ge of po-rsibre ,.Ji;
a project with
RoR to positi'e 4c7o might be consideri from negative
ress cssirabt. tt un a r_,rcrject wirh
a i,lr)it expecied ROR oi lgvo and
a range u.f pos.;ible resurts from rQvo
25vc because the certainiy associateJ to
,rirtr, tn. tsz, mosi-expected RoR
inr"estment is greater.tho,t th"..rtointy
r.socrieted with the other investment.
Prohabilistic sensitivity anarysis .*.ot.i-tr,"
decision get a firmer
fecirng concerning rhoeff-ecis or.irirra "r"t.r^,"
uncertainty on economic anarysis
resuirs than an-v otlrer analysis approxsh gi1.sE.
The weak pr:int of trre prooauriistic
ariarysis method ries in the subjectire
assi,uning of probabiritie. of occurrence
to the rer,.els of pararneters that go
int. the arrarysis' it is gene'aily considered to be best to specifi-
'o::.u:r, of o.-.r.."n.. based on the
callv staie the probabililies
peopic invoh'ed rvith a project best judgement of
and then to ba.se the anaiysis on these
maies even though they are subjective esti_
in nature. In the finar analysis any
evaluatio, technique is only as gooo
as the estimates of the input parameters
and rnust be used in conjunction
with good engineering logic and managerial
judgement' Assigning probabilities
to"pararneter estimates is just one more
step in quantifying the assumptions
thai are made. These rechniques provicle
managcment with addirionai toors to
aiir in the decision-.uung'p.o."rr.
Folloiving is a simprified exampre that
ilrustrares the principles of appry-
ing probabilistic analysis to project
ROR sensitivity analysis.
EXAMPLE 6-9 A Simple probabilistic

Sensitivity Analysis
A new product to be produced by one
of two different processes. rt
is felt that there is a 60% prooaoitity
that the process serected wiil
have initial cost of $50,000, a life of five years and zero salvage
value. There is a 40"/" probability that an improved process will be
selected with iriitial cost of $4O,OOO, a life of five years and zero sal-
vage value. With either process there is a 50ol" probability that
annual profit will be $20,000 for the five year project life and a 25"/o
probability that the annual profits will be $15,000 or $25,000 per
year. Plot project ROR versus probability of occurrence assuming the
parameter values are independent of each other.
Solution:
The following table presents the possible combinations of different
investment costs, annual profits, probabilities of occurrence and ROR.
Annual Probability of
lnvestment Profit Occurrence P/Ai.s RoR (%)
50,000 15,000 (.6X.25) =.15 3.334 s.2
50,000 20,000 (.6X.50) =.so 2.500 28.7
50,000 25,000 (.6X.25) =.15 2.000 41.1
40,000 15,000 (.4X.2s) =.to 2.667 25.4
40,000 20,000 (.4X.50) =.zo 2.000 41.1
40,000 25,000 (.4X.25) =.10 1.600 56.0
1.00
The most probable project ROR is 41.1"/" with a 35% probability of
occurrence. The cumulative probability diagram shows a cumulative
probability o175"/" that the ROR will be 28.7% or above in Figure 6-5.
Instead of mathematically determining the probability of occurrence of the
various ROR results for this problem we could have used the general Monte
Carlo simulation technique to get the same results. The general idea of this
method as described earlier is to first develop curves for cumulative probabil-
ity of occurrence versus the economic ltarameter similar to Figure 6-4. Then
random numbers between zero and one are related to cumulative probability
of occurrence so that selection of a random number fixes the initial invest-
ment for a calculation. Selection of another random number selects the cash
flow for that calculation. ROR is then calculated using these values. This pro-
cedure is then repeated several hundred or maybe a thousand times until the
shape of the ROR versus probability of occurrence curve does not change
with additional calculations. For a large number of Monte Carlo simulations
the results using this technique for Example 6-3 will be identical to those
Chapter 6: Uncertainty and Risk Analvsis
295
given in Figure 6-5. The Monte carlo simulation data are evaluated by
form-
ing a histogram from the RoR results. For instance, if one thousand runs are
made tlren approximately tluee hundreci of the RoR rcsults ,rrorr.i bc 2g.i7o,
since we know the mathematical probability of occurrence of this resylt is .30.
In general, the variations ir.r input crata such as seiling price, oirL-raiing
cost. production rale, borrowed money interesr rete and so forth that are
evaluated will be continuous functions of cumulative probability of occur-
rence rather than the step functions illusrrated in Figure 6-4. Coniinuous
input data give a continuous RoR versus probabiliiy of occurrence his-
tograrn graph for lrlonte Carlo simulation in general.
1.00 1.00
.80 .80
Cumulative
Probability
.60 .60
of
Occurrence
.44 .40
.20 .24
0 0
15,000 25,000
Initial lnvestment Profit
(a) (b)
Figure 6-4 Cumulative Probability Diagram
Probability
.30
of
Occurrence
.20
.10
0
1.00
.80
Cumulative
Probability .60
of
Occurrence .40
.20
0L
0 10 20 30 40 50 60
ROR, %
Figure 6-5 Cumulative probability of Occunence
6.5 Expected Value Analysis (Economic Risk Analysis)

Expected value is defined as the dffirence between expected profits and
expected costs. Expected profit is the probability of receiving a certain profit
times the profit and expected cost is the probabitity that a certain cost will be
irtcurred times the cosr. If you define cost as negative profit and keep the signs
straight, you can do as some text authors do and define expected value as the
algebraic sum of the expected valuq of each possible outcome that could
occur if the alternative is accepted. Either definition leads to the same
expected value result, which sometimes is called a "risk adjusted" result.
Several examples of expected value analysis when time value of money con-
siderations are not relevant or significant will be presented first. Then time
value of money related examples will be illustra.ted.
EXAMPLE 6-4 Expected Value Analysis of a Gambling Game

A wheel of fortune in a gambling casino has s4 different slots in
which the wheel pointer can stop. Four of the s4 slots contain the
number 9. For $1 bet on hitting a g, the gambler wins $10 plus return
of the $1 bet if he or she succeeds. what is the expected value of
this gambling game? what is the meaning of the expected value
result?
Solution:
Probability of Success = 4/54
Probability of Failure = SOIS4
ExPected
:lffi ;;: *
Var ue
:,?ffi,'*,T -"-::l
The meaning of the -$0.195 expected value result is that it is the
rverage monetary loss per bet of this type that would be realized if
he gambler made this bet over and over again for many repeated tri-
rls. lt is important to recognize that the gambler is not going to lose
80.185 on any given bet. over a large number of bets, however, the
oss per bet would average $.18s. This result should make it evident
hen that a positive expected value is a necessary condition for a sat-
sfactory investment, but not a sufficient condition as will be dis-
:ussed later.
Chapter 6: Uncertainty ar':C Flisk Analysis 297
EXAMPLE 6-5 Expected Value Analysis of a Simplistic
I
. ,:,DrillingVenture
lf you spend $500,000 drilling a wildcat oil well, geologists esti-
I
mate the probability of a dry hole is 0.6 with a probabiiity of 0.3 that
i
! the well will be a producer that can be sold immediately for
l $2,000,000 and a probabiiity of 0.1 that the weil will produce at a rate
I that will generate a $1,000,000 immediate sale value. What is the
{ oroject expected value?
{
,
f Solution, All Values in Thousands of Dollars:
Expec'ied Va ue
I
I
I
: il:ffi :;ffi ; :lT:::::H, _ o 6(500)
I = +$200
or rearrangtng:
I
Expected value = 0.3(2,000) + 0.1(1 ,000) - 1.0(500) = +$200
I
Over the long run, investments of this type will prove rewarding,
I
but remember that the +$200,000 expected value is a statistical long-
term average profit that will be realized over many repeated invest-
I ments of this type. The expected value of an investment alternative is
i
the average profit or loss that would be realized if many investments
of this type were repeated. ln terms of Example 6-5 this means if we
dritied 100 wells of the type descrrbed, we expect statistics to begin
to work out and assuming our probabilities of occurrence are correct,
we would expect about 60 dry holes out of 100 wells with about 30
1
t wells producing a $2,000,000 income and about 10 wells producing
i a $1,000,000 income. This makes total income of $70,000,000 from
I
I '100 wells drilled costing
a total of $50,000,000 leaving totai profit of
tf
$20,000,000 after the costs, or profit per well of +$200,000, which is
the expected value result for Example 6-5.
Certainly if you have enough investments of this type in which to invest

It and enough capital to invest, statistics are very favorable to you, and you
lI would expect to come out ahead over the long run if you made many invest-
ments of this type. However, if the loss of the $500,000 drilling investment
on a dry hole would break you, you would be foolish to invest in this type
I
*
t
of project because there is only a 4avo chance of success on any given try.
only if you can stick with this type of investment for many times can you
expeci statistics to work in your fuvor. This is one important reason why
most large companies and individuals carry insurance oi various types even
though in nearly all cases the expected profit from self-insuring ii positive
and therefore favorable to the company or individual. If a disaster from
fire
can break a company cr individual financially, that company cannot
afford
to self-insure. This is why insurance companies spread large policies over
many insurance companies. If disaster does strike a large poti.y holder,
the
loss will be distributed over several companies, lessening the iikelihood
of
financial disaster for any one company. It is also the r"u.on most individuals
carry fire insurance, homeowners or tenant insurance, and car insurance.
The direct financial loss or lawsuit loss potential is so great that most
of us
cannot afford to carry that risk alone even though exp-cted value is
favor-
able to us if we self insure. The conclusion is thit a pisitive expected
value
is a necessary br.tt not a sfficient condition
for a satisfactory iivestment.
It should now be evident that although expected value has deterministic
meaning only if many trials are performed, if we consistently follow
a deci-
sion-making strategy based on selecting projects with positive expected
val-
ues, over the long run statistics will work for us and income
should be more
than sufficient to cover costs. on the other hand, if you consistently
take the
garnbler's ruin approach and invest or bet on investments or gambling
games with negative expected values, you can rest assured that
over the
long run, income will not cover your costs and if you stick with negative
expected value investments long enough, you will of course, lose all your
capital. This is exactly the situation that exists with all of the gambling
garnes in places such as Las vegas, Reno and Atlantic city. The
odds are
alrvays favorable to rhe house, meaning the gambling housl has a positive
expected value and therefore the gambler has negative expected value.
The
gambler has absolutely no realistic hope for success over the long
run under
these negative expected value conditions. He will lose all the money
set
aside to gamble with if he sticks with the games long enough..
The reader should notice that in Example 6-5, two different, but equiva-
lent equations were usec to calculate expected value as follows:
Let P = probability of success, I -p = probability of failure

Expected Value = (P)(Income (1 _ pXCost)
- Cost) - 6-2
or =(P)(Income)-(i.O)(Cost) 6-2a
C;,apter 6: Uncertainty and Risk Analysis
6.6 Expected NPV, Expected pVR, and Expected ROR Analysis

when time value of money considerations are significant,
expected Npv,
PVR and RoR anarysis are merhods of incrudrrg
ir" p."ir"uiii,;"s of suc_
cess and failure in analyses when costs and
revenues occur ar different
ptrints in time. If we use appropriate time ralue
of monel. present rvorth fac_
ttlrS ro convert costs and profits at different points
in time to lurnp sunr r.al_
ucs at time zero or some other chosen tinre, the
expected value anarysis
approach can be applied to determine if rhis type
crt.- investment woulcl be
snitable over the long run for many repeated investments
of the same type.
\{'ith e-rpected Npv and pvR *r u.., of course,
looking for alternatives
with a positive expected with expected RoR anarlsis we calcurate
the expected RoR value,'arue."i", that will rnake the expected Npv equation
cquai to zero. An acceptabre expected RoR must
be greater than the mini-
mum ROR.
EXAMPLE 6-6 Expected Vatue Apptied to ROR,

NpV and pVR
Analysis
A research and deveropment project is being considered.
ject is expected to have an initial ihvestment The pro-
cost of $90,000 and a
probability of 0.4 that annual profits of
$50,000 will be reatizeO during
the 5 year life of the project with a probabitity of
0.6 of faiting. Sal_
vage value is expected to be zero for success
or failure. Assume the
minimum discount rate is lOoh on a risk_free basis.
should the project be done? compare expected varue,

nel present value, expected rate of return and expectedexpected
present
value ratio analysis with corresponding non-risk adjusted
evaluation
results.
Solution, Values in Dollars:
C = $90,000 P=0.4 | =$50,000 . . . . I =$50,000

1............s
P=0.6
L=0
i
A) Expected vatue Anarysis rncruding Time varue of Money

at
i' = lgolo
3.7908
EV = 6.4150,000(P/A1
Oo/o,S) - 9O,O0O) - .6(90,000) = _$14,1g4
The negative expected varue indicates rejeci investing
in the project.
The expected varue approach is most usefur when
statisticaily
more complex models are utilized. ln many of these
cases, the vari-
ous outcomes that are possible in a model are defined,
asso-
rn this "no ", t*o
ciated chance factor is derived for each branch.
outcomes exist which can be rabered ,,success',
and ,fairure."
"r."rpr"
By car-
culating the NPV for each separatery, the chance factors (in
.outcome
this modeljust the probabirity of success and fairure)
can
to.the appropriate outcomes to determine the identicalbe appiied
expected
value (EV) as shown:
Expected value per outcome (Npv of outcome)(chance
= Factor)
summing ail of the project's expected varues per
outcome gives
the overall project expected value.
3.7908
EV Success = {50,000(p/Aj _
EV Failure = (_go,ooo)(0.
Oy",S) 90,000X0.4) =
39,g16
= -54,000
Project Expected Value
= -14,184
The same calculation was introduced in the expected
value solution
to this problem.
B) Expected Net present Value

This analysis is identicar to the (A) anarysis, except
we generaily
use Equation 6-2a to determine Expected Npv. Thi; just
is a rear_
ranged form of the Expected Value equation
from part,,A,,.
3.7908
ENPV = 0.4(50,00O)(ptA1O%,5)
- 1.0(90,000) = _$1 4,184
Since the ENpv is negative, we shourd not invest
in the project
from an economic viewpoint. Note the expected
varue anarysis in
oart (A) accounting for the time varue of money
ro7" p"iyear gave
e result identical with the EV result from Case "t
A.
Chapter 6: Unce(alnty and Risk Analysis
For these statistically simplistic models, another approach

for cal-
culating the ENPV is to first, risk adjust the cash trows, and
then, car-
culate ENPV using the same present worth equation format previ-
ously developed. This approach is helpful in making expected
value
calcuiations on a hand calculator but again, it shor.rli be
emphasized
that the resuiting cash flow siream does not represent the
expected
ca-sh flows for any single investment. lnstead, ihe calculated
values
reflect the average cash fiows that mignt be expected after
many
repeated investments of this type with success occurring 40%
of the
time and failure occurring 60% oi the time.
Cash Flow. -$90,0i,09.0)__!9900!04).. .. s0,000(0.4)
0
Risk Adjusted
Cash Ftows -$90,000
ENpv = -e0,000 + 20,000(priliijr, *14,184

=
This approach can be more difficurt to appry to more comprex
sta-
tistical models than summing the expected vaiues for all
outcornes.
lf you look at non-risk adjusted NpV (risk free NpV) which implic-
itly assumes 1007o probability of success, the project economics
look very acceptable
Risk Free NpV = (50,000)(pr"^133^,r,- (e0,000) +$ee,540 >

= 0
obviously adjusting for risk of failure or not adjusting for risk
of fail_
ure has a very significant impact on economic results
aid conclusions.
C) Expected Rate of Return
Expected RoR is the "i" value that makes Expected NpV equal
0.
Expected Present worth lncome @ "i" present worth cost ,,i,,=
- @ 0
50,000(P/Ai,S)(.4) - 90,000 = 0
by trial and error, "i" = Expected RoR g..ay" < i* 10% so reject
= =
Non-risk adjusted or risk free rate of return analysis gives a
differ-
ent conclusion:
50,000(P/A1,5) - 90,000 = 0
by trial and error, f i" = ROR = 47.60/o> i* = 10% so accept

This result is much greater than the 10% minimum RoR indicating
very acceptable economics. As a variation of expected RoR analysis,
some people account for risk by increasing the minimum ROR by
what they consider to be an appropriate amount. The difficulty with
this approach is that there is no consistent, rational way to adjust the
minimum RoR appropriately to account for risk in different projects.
For this example, the risk free RoR is about 42.6% (based on proba-
bility of success of 1.0 instead of 0.4) compared to the expected RoR
of 3.6%. Expected RoR of 3.6% compared to risk free minimum RoR
of 1oh indicates the project is economically unsatisfactory. To get the
same conclusions based on comparison of risk free RoR and risk
adjusted minimum RoR results requires increasing the minimum
ROR to about 132% (where l1yo/3.6/" = Xl4Z.6/q therefore X
=
13:2%). A majority of people who attempt to compensate for risk by
adjusting the minimum RoR end up significantly under-compensating
for risk. For this example many peopre wilr propose increasing the
minimum RoR inversely proportional to probability of success 6to.q)
from 10"/o to 25% when as previously discussed the increase should
be from 10% to about 1go%. However, there is no way to know this
without making an expected RoR type of analysis. Expected value"
analysis using RoR, NPV or. PVR is the preferred approach to incor-
porate risk into economic analysis calculations.
D) Expected Present Value Flatio Analysis

EPVR = ENPV / Expected PW Cost = -14,184/90,000 =
-0.16
The negative expected PVR indicates the project economics are
;nsatisfactory for the project parameters built into this analysis.
Risk Free PVR = +99,540/90,000 = +1..l1
consistent with the other criteria, the risk free pVR indicales very
rcceptable economics which is the opposite conclusion reached with
lxpected PVR.
Expected value anarysis in general involves

constructing a diagram
showing investmenr cosrs and all subsequ"r,
riilues that are anticipated. Standard si;mbolism"t*."^;;;;;, and dolrar
uses circles to designiite
chance nodes tiom which different delrees
of success and fairure may be
-rh,"rn ttr r-rccLlr. l''c sum of the proh.b'ilities of
occurrence on the dittbrent
branches ernanariirg from a .hu,.,"" node
must adil up a r.o. Trtese
e-rpec'ted r'.;lua L!trigrtuns" are someriutes
. called ,,rlecisiori rree d.iagrants,,
be-cause decisior: options concerning
whether to proceed in one or more
di ferent wa)'s or to terminate the project aiways
.^irt prio, to each chance
node where different degrees oi ,r.."r,
or failure may occur. These
diagrams often ha'e murtiple branches and
rook u.ry -u.h like a drawing
of a tree' which has led to the name "decision
ree analysis,, being used in
industry pracrice to refer to this type of analysis.
In typicai decision tree
anaiyses, at dift-erent stages of projects,
probabilities of success and failure
cha3ge, As you progress from the i"r.ur.t
or exproration stage of a project
to.c.eveiopmenr and production, risk
oi' failure ,rgniii.unrry. This
is illustrated in the follorving two examples. "hung",
EXAMPLE 6-7 Expected varue Anarysis

of a petroreum proiect
Use Expected NpV anarysis for a minimum
RoR of 2o"hto evar-
uate the economic potentiar of buying
,:
and driiling an oir lease with
the following estimated costs, ,."rlnr". and
suc-cess probabirities.
The lease would cost g.100,000 at time 0
and it is considered 100%
certain that a wefi wourd be driiled to the point
year later for a cost of $500,000. There
of .ornpr"tion on"
is a boz pronaoitity that welr
lggs will took good enough.to comprere the *!riril"ar
1 for
$400,000 compretion .o.i. tf tre wirr rogs are ,n.rtirtr.tory a
abandonment cost of $40,000 wi, be incurred an
at year 1. rf the we, is
completed it is estimated there will be a
50"/" prooroitity of generat_
ing production thar will give $450,000 per year
net income for years
2 through 10 and a 3S^y" probability of generating
net income for years 2 through io, witn $000,000 per year
a 1so/o piobanitityof the weil
comple-tion being unsuccessiur, due to
water or unforeseen compre-
tion. difficulties, giving a year 2 salvage
value of $2SO,OO0 for pro_
ducing equipment.
Solution, in Thousands of Doltars:

l=450 l=450
(D) (c)
C=40
Times 1 and 1+ are effectively the same point in time, the end of
year 1. Normally drilling and completion time are separated by weeks
or months which puts them at the same point in time wittr annual
periods. Times 1 and 1+ are separated on the diagram to make room
for the probabilities of occurrence associated with different events.
Expected Value
. *Determine the possibte different outcomes and the subsequent

NPV for each. Apply the probability of each outcome to the calcu-
lated NPV's giving ENpv per outcome. summing each gives the pro-
ject's overall ENPV.
There are four (4) possible outcomes:
A) Successful Development leading to incomes of $450 per year.
B) Successful Development leading to incomes of $300 per year.
C) Failure with a salvage of $250 at the end of year two.
D) Failure with an abandonment cost of $40 at year one.
For Case A, the chance factor is; (0.6X0.5)r
For Case B, the chance factor is; (O 6X0.35)\
For Case C, the chance factor is; (O.OiiO.f S)
\
I-100 - eoo(p /F 20,1) + a50(p/A
1)
e00(p tF 20"1) + 300(p/A
20,eXp/F zo rltlsol = +198.48
?l t-100 - 2['!11er ;;:j,] io.zr
' i = +33.13
9) t-100 - e00(ptF 20"1)-1-250(p/F i6"ill(0.00i""
= -60.87
D) t-100 - 540(P/F16"1)l (o 4)
= -219.99
Project Expected Net present Value
lftrln4 = -49.26
l---
Chapter 6: Uncertainty and Risk Analysis g0
Expected Net present Value (ENPV) at 2}o/o

4.031 4.031
.8333
{[450(PiA20,eX.5) + 300(piA20,ex.ss) + zs0(n{1X. j5) _ 400](.6
.8333
- 40(.4) - 500xP/Fz},i _ 100 = _49.26
Thisresurt is onry- srighfly ress than zero compared
ject costs of $1 miilion, therefore, srighily tc the totar pro.
unsatisfact"iy o,, break.
even economics are indicated.
Alternate Form of (ENpV) Equation

4.031 .8333 4.0g1 .ffi33
4s0(P t Azo, (. 5) (P/F20, 6) + 000(p/A2s e) (. 35) (p/Fio,
e) X.
1
r X o)
.6944 .8933 .8339
+zso (p'/ F 20,z) (. 1sX. 6) _ 4oopi ; ;;. _ +o pir"ll1; i: +1
1 X. 6)
.8333
-500(P/F20,1X1 .0) - 100 = -49.26
Risk Adjusting the Cash Flows
-500(1.0) 450(0.6X0.5) 450(0.6X0.s)

-jlBIS 3oo(0 6iio 3b)
cF(prob
Year
) _i00(1 0) ?i IBBIB\Y'Y/\v'
Bi[B ?Ei
tet
O 1 g_10
RiskAdj cF -1oo -zs6 ffi
ENPV = -100 -756(p/F20,1)+ 220.S(p/F29,2)
+'t 98(p/A2 0,A) (p ff
,0,r1
ENPV = -49.26
EXAMPLE 6-8 Expected value Economics of a New process
_ use Expected NpV and pVR anarysis for a minimum rate of return of
2a.0"/" to evaluate the economic potentiar
of buying ano Jevetoping the
rights to a new process with the foilowing
estimated costs, revenues
and success probabilities. The process rights
would cost g100,000 at
J
time zero, and, it is considered 100% certain.that experimental

devel_
9ql"ll pilot plant work wiil be done one year raier for
a cost of
$500,000. There is a 60.0% probabirity that the experimentar
develop-
ment results will look good enough to iake the project
to production flr
a $400,000 capital cost at year 6ne. (This capital cost is
estimated to
be incurred in the first six months of year one which is
closer to year
one than year z.) lf the experimental development results
are unsatis-
factory, a pilot prant abandonment cost of
$+o,ooo wiil bL-incurred at
year one- If the project is taken to production,
it is estimateci there will
be a 50.0% probabirity of generating production that
wirr tive g4s0,000
per year net positive cash flow for years
two through ten, i 35.0% prob_
ability of generating $300,000 per year net positive
cash frow for years
two through ten with a 15.0% probabirity of the project
deveropment
being unsuccessfur due to unforeseen iechnicai oiiticutties
giving a
year two salvage value of $250,000 for production
equipment.
Solution, in Thousands of Dollars:
l=450 l=450
(A)
!=1OO
lru
P=1.0 C- P=0.6 P=0.35 l=300 l=300
01
P=0.4 P=0.1 5
(D) (c)
C=40 Salv=250
Expected Value
..letermine the possibre different outcomes and the subsequent

l\lv. Igl "uch. App]y the probabirity of each outcome to the carcu-
lated NPV's giving ENpv per outcome. Summing gives the pro-
ject's overall ENpV. "r.n
There are four (4) possible outcomes:
A) successfur Deveropment reading to incomes of $450 per year.
B) Successful Development leadin! to incomes of per year.
c) Faiture with a sarvage of $2s0 ai tne end of y"ri$gOO
t*o.
D) Failure with an abandonment cost of $40 at'year one.
Chapter 6: Unce.tainty and Risk Ahalysis
For Case A, the chance factor is; (0.6)(0.S)r

\
\
1) t-100- e00(PlF zo,t) + 450(P/A 20 eXp/F 20 il fo.eoy= +1e8.48
e00(P/F zo',t) + 300(p/A 2['eylerr zo. jlj tt'.z1)= +33.13
!) i-t00 - e00(Pilzo',1) -
9) f-100 - .250(piF i6',ill (c.obi" = -60.87
D) [-'100 - 540(P tF zo',t)] (0 4) = -219.99
=-
l'roject Expected Net Present Value (ENPV)
= _49.26
Risk Adjusting the Cash Flows
-500(1.0) 450(0.6x0.5) 450(0.6X0.s)

-400(0.6) 300(0.6x0.35) 300(0.6)(0.35)
CF(Prob.) -100(1.0) -40(0.4) 250(0.6X0.1s)
Year0lzg_10
Risk Adj CF -100 -7s6 zzo.s 198
ENPV = -100 -756(PlF20,1)+ 220.5(plF29,2)

+ 1 98(p/A2 g,g) (p /F 2g,2)
ENPV = 49.26
Expected Net Present Value (ENPV) at20o/o

4.031 4.03.1 .8333
{1450(P I A2O, 9X. 5) + 300( P/A20, g) (. 35) + ZSo (p I F (. 1 5) - 4001 (. 6)
20, 1)
.8333
- 40(.4) - s00)(p/F2},i - 100 = _4s.26
This result is only slighfly less than zero compared to the total pro-
. costs of million,
ject $1 therefore, slightly unsatisfactory or break_
even economics are indicated.
Expected Present Value Ratio (EPVR)

EPVR = 49.26 / 100+[500 + a00(.6) + 0(.a)](p /F20,1)
= -0.07
The small negative EPVR result indicates the same slighily unsat-
isfactory or break-even economics shown earlier with ENpV analysis.
EXAMPLE 6-9 Expected NPV Anatysis

Calculate the,,Expected NPV of a project which will cost $270,000
at time zero. This investment has a 60.0% probability of generating
downstream development and equipment costs-at tne=end-of yeai
one estimated to total $500,000. lf the secOnd expenditure at the end
of year one is successful, there is a g0.0% probability it would lead to
the generation of net cash flows totaling $400,000 per year at the
end of each of years two through ten. Failure from the time zero
investment would result in no additional cost or benefit to the
investor. However, should the project fail after the year one cost, a
net cost of $250,000 would be realized at the end of year two from
dismantlement costs and salvage of equipment. Should this project
be accepted? Use a minimum rate of return of 1Z.Ooh.
Solution: All Values in Thousands.
.-$270 P=(0.60) -$500 P=(o.eo) $400. ..$4oo

,......-* (n)
P=(0.4) P=(0.10)
(c) (B)
$o -$2s0
Expected Value
There are three (3) possible outcomes in this solution:
A) Successful Development, Chance Factor,

P=(0.6)(0.e) =(0.54)
p = (0.6X0.1) = (0.06)
B) Failure at Year 2, Chance Factor,
c) Failure at Time Zero, Chance Factor, p = (0.40)
EVA) [-270-500(P/F1 2,1) + 400(P/\2.9XP/F p,i)(0.s4) = +640.71

EVs) [-270 - 500(P/F1 2,i - 250(PlF p,2)] (0.00) = -54.94
EV6) [-270] (0.40) = -108.00
Project Expected Net Present Value (ENPV) = +477.77
Chaoter 6: Uncertainty and Risk Analysis 309
Risk Adjusted Cash Flows

Year 0. - 1 .,, 2 3-10
400(0.6x0.e)
CF(Prob.) -270fi.q -500(0.6) -2s0(0.6x0.1) 400(0.6x0.e)
Risk Adj CF -270 -300 201 216
ENPV = -270 - 300(P/F1 2,1) + 201(PlFp,2)
+ 216(PlA12,g(PlFe,Z)
= +477.77 > 0, accept
Expected Net Present Value
ENpv @ 12.0o/o = +0o17;ffiX0.e)3/8Fe122:?,,0 u,

0.79719 0.89286
- 250(P/F1 2,2)(0.1X0.6) - 500(P/F12,1X0.6)
- 270 = +477.77 > 0, ok
Example 6-94
Utilizing the data from Example 6-9, what additional cost could be
incurred at time zero tor either research or geological/geophysical
data and give the investor the same risk adjusted NPV of $477.77?
Assume the additional time zero cost will increase the probability of
success in year '1 from 0.6 to 0.8. Therefore, the probability of failure
in the same period will be reduced from 0.4 to 0.2.
Solution:
Let X equal the additional cost to be incurred at time zero.
-x ^-270 P=(0.8) -500 400 400

(A)
0
P=(0.2) P=(0.1)
0 (c) -250 (B)
,!
310
Economic Evaluation and lnvestment
Decision Methods
ENPV Approach:
Ecohomicary equivarent, mutuary_excrusive
always have equar net preseirt arternatives wi,
at equating the current project
r"ir".. Therefore, this sorution rooks
Npr;ith the Npv br tnlr"uired prob_
abilities based on the new invest*.ni,
x at time zero as forows:
477 .77 = [400(p/A
eX,g)(0.9) _ 500J(p/ F ex,i@.8)
- 250(p/F12/",2)(0.1)(0.S) _ z7O _ X
477.77 = [400(s.s2825X0.9) _
so0](0.89286)(o.B)
_ 2s0(0.7e71exo.1x0.8) _270
_X
477.77 = 1012.98_ 15.94 _270_X
X = 249.27
Risk Adjusted Cash Flow Approach,
ENpV:
-x -270 P=(0.8) _5W P=(0.9) 4OO 400. .400
;; . -, (n)
P=(0.2 P=(0.1)
0 (c) -250 (B)
Risk Adjusted Cash Flow Catcutations:
400(0.72)
-270(1.0) -500(o.s) -250(0.08) 4oo((0.72)
01 0.......... 10
Risk Adjusted Cash Ftows:
-.270 _4oo
268 288 . . .288
3..........10
ENpv = - o-
27 400(p?F81Yji"?, + z6s@ /{1\ir,
4.9676 o.7g71g
+ 2Eg(p / A1
2"7",g) (p / F 1 2y",2)
ENPV = 727.02
,11
The difference between the New ENpv of 227.oz

and the originar
ENPV ot 472.27 is 249.25 which represents the
additional cost that
could be incurred at time zero and ailow the invest,oi
to obtain the
desired ENPV of 477.27.
It was emphasized earlier in this section that expected value
represents
the average gain or loss per investment that
an investor would realize over
many repeated investments of the type being analyzed.
whether we work
with expected vaiue, expected Npv, expected pvR
or expected RoR, the
average meaning of results is similar. A common
misconception that some
people have about expected value analysis is
that it oft;n is not valid
because investors serdom repeat the same type
investments over and over.
These people have missed the basic .*p".t.d
value analysis premise that
even though each specific investment <tecision
relates to u urilr"ty different
investment, if we consistently select investment
altematives having positive
expected value, ENPV or FPVR, (or having
expected ROR greater than the
minimum R.oR), over the'long run our average
rate of return on invested
capital will be greater than the minimum RoR.
Similar to material pre_
sented earlier in chapter 4 for risk free analysis,
to rank non-mutually
exciusive alternatives using risk adjusted reiults,
you must use either
expected PVR or expected Growth ROR resultr.
fo evaluate mut,ally
exclusi'e alternatives with risk adjusted results
based on any valid anarysis
technique, irlcremental analysis is the key to
correct economic decision
making. These ruies hold true even though each
investment decision rerates
to a different investment prospect with different
probabilities of success and
failure at different stages ofeich project.
6.7 Protrability of Survival (Financial Risk Analysis)

Probability of survival refers to the probability
that you will not go
bankrupt with a given amount of capitar to invest in-projecis
with estimated
probabilities of succes.s. This concept is
a financial .irt unutyris considera_
tion rather than specificaily relating to
risk analysis. A project
with a large positive expected value and ".ono*ic
a smail probability of success may
economically look better than ail other investment
opportunities under con_
sideration. However, if.failure of the project
would lead to bankruptcy, the
project most likely will be considered financially
unacceptable due to the
financial risks. The itaricized statement preceding
Example 6-6 says ..a pos-
itive expected value is a necessary but not a
sufficient condition for a satis-
factory investment." This meanr ihut u positive
expected value indicates an
economically satisfactory investment but not necessarily a financially satis-

factory investment..small investors in the oil and gas drilling, business,usu.
ally attempt to reduce rheir financiar risk of total failure liankruptcy; by
-intei-
taking a small interest in a large number of projects rather than large
ests in a few projects. This diversified investment portfolio upp.oo.h giu..
the probability of success a better chance to be realized over time as the fol-
lowing example illustrates.
EXAMPLE 6-10 Probability of survival Applied to Exploration

consider an exploration manager who is faced with the task of
determining whether to invest 91,000,000 of a small company's
money in 10 independent exploration projects which will each cost
$100,000 with a 10% probabitity of achieving a gs,000,000 profit
above cost; lf the company will go bankrupt if not successful on at
l"3pt ] exploration project (or if the exploration manager will lose his
job), discuss their probability of survival.
Solution:
Expected Value = 5,000,000(.10) - .tOO,OOO(.9) = +$+t0,000
The positive expected value is economically satisfactory and is a
necessary but not a sufficient condition for this investment to be
satisfactory. The question to be answered now is what is the proba-
bility of survival, or stated another way, what is the probability of get-
ting at least 1 success out of 10 tries. This relates to the financial
acceptability of the project.
Probability of at Least one SucceSS = 'l probability of Zero Successes
-
Probability of Zero Successes = (.g)10 = .34g5
Therefore, probability of survival = 1-.84g5 = .6515 or 65.15%
which is certainly considerably less than a sure thing (100% proba-
bility of success). Note that ten projects each having a 10% chance
of success do not give a 1oo% probability of overall success.
companies often get together in joint ventures on large exploration
projects to have more capital available so that enough exploration projects
can be made to make the probability of survival higher than it would be if
they operated alone. If in Example 6-r0 a sum of $2,000,000 is available to
invest in 20 projects, the probability of survival increases to l-.1214 .g7g6
=
C,rapter 6: Uncertainty and Risk Analysis 313
and if $3,000,000 is available the probability of survival is .95i7. Joint ven-

tui'es can increase survival probabilities to very tolerable levels when explo-
ration work is being done in areas where enough geological and geophysi-
cal work has been done to predict reasonable probabilities of success on any
gir cn tr).
The same type of reasoning can be applied to research and development
projects in all types of industries, but it is very difficult to come up with
truly meaningful probabilities of success. However, if these probabilities are
not estimated explicitly, managers will base a decision impiicitly on "gut
feel" for data, and these authors feel it is better to explicitly state the bases
upon which decisions are made.
The statistical basis of the probability. of survival calculations just pre-
sented is the binomial distribution for mutually exclusive altematives. The
general binomial distribution equation is:
Probability of Exactly "r" successes from "n" tries = glpr11-pyn-r
where C? = Combination of "n" things taken "r" at a time.
nl n factorial
r!(n-r)l r factorial times (n-r) factorial
nl = n(n - lXn - 2) ---- (3) (.2) (t)

0! =lbydefinition
P = Probability of success on a given try.
I -P = Probability of failure on a given try.
Note that for Example 6-10, the probability of zero success from l0 tries is:
c1o0r.rlor.o)'0 = .ffi (1x.9)10 = (.p110 = 0.3485
6.8 Risk Due to Natural Disaster

Another type of risk to be considered is that due to natural disaster. The
following example illustrates the evaluation of data using probabilistic
expected value concepts for this type of problem.
EXAMPLE 6-11 optimum lnvestment to Minimize Flood Damage

Costs
A manufacturing company plans to build a plant on low land near a
river that floods occasionally. lt is considered necessary to build a
levee around its facilities to reduce potential flood damage. Four dif-
ferent sizes of levees that give different levels of protection are being
considered and the plant manager wants to know which size levee
will minimize total expected annual cost to the company from the
sum of (1) amortization of the levee cost over 20 years for a before
tax "i " of 157": plus (2) maintenance costs, plus (3) expected dam-
age to the plant and levee if the levee is not high enough or strong
enough to hold back flood water. Analyze the following oita to deter-
mine the optimum levee size.
LeveeLevee Probability of a Expected Damage Annual

size cost Flood Exceeding Levee if Flood Exceeds Maintenance
Size During year Levee Size
1 $120,000 .20 $ 8o,ooo $4,000

2 140,000 .10 110,000 5,000
3 160,000 .05 125,000 6,000
4 200,000 .025 150,000 7,000
Solution:
Expected Total
Levee Damage Annual Expected
Size Annual Levee Cost + PerYear + Maint. = Annual Cost
0.1598
1 120,O00(A/P15,20)=19,180 16,000 4,000 $39,'180
2 '140,000(0.1598)
= 22,370 11,000 5,000 38,370
3 160,000(0.1598) = 25,570 6,250 6,000 37,820
4 200,000(0.1598) = 31,960 3,750 7,000 42,710
Levee size number 3 is the optimum selection to minimize

expected equivalent annual cost.
Cilapter 6: Uncertainty
and Risk Analysis
J,3
{i.9 Option pricing Theory Related

to ENpV
k. Patent rights to new technorogy,
Iand and minerar rights
erries ofren se, at prices to minerar prop_
signifi-c'#rrv rrigrr", ;il;;il'*"ff
erty' Those caicurations
ur. uur.J-Jn tf," *ort .-r..,.0 of the said prop_
production rar.es,
il:ff'J#n:::l;:f*T":nf.r:' -
'lr''
p.q..ti* ;#" or the ."u,on
pri ci n g erre* rnay,
u" r"r u t" J io ;;;,i:,
me)' occur if actual production il ?:,H#ffi::"ffi;, T:#Hl
rates or product prices
than the mosr expecred
productio;;;". and prices turn out to be higher
costs tum out to be lowei ro. ir," a*ual operating
than trr. rnort expected
may also rerate ro the decision
,o J.ruy
^p."ar.ii* grt option pricing
,utu"r),
greater procuct nrice
in the future, o, ,.urir" g."rr..'.*#es iiLy ura caprure a
than what is
rhe t i ai ;;;;. u.
[',XJ,x,JI'J:H i ll:t {',:.Xl' }'i:' io ten or dec re a s e

"
pricine
because
"r ,;;i #;:rt"#.:Hption
i,r Ji'"., rerationshif
"n""i
options are referred to as_either
j
puts. or cars. A generar
,lffi review of each
i
i, :tffi:rH::
-io, x;3;*ti*1*
u-ni
",
;.:;
; *" d in ch ap,er
opti on p.i g,
"i,
th ar, reader .r; ;#:;,[HJ ::,"]X#t :tii,f iilf
;:lli:f ;i&lliT |i,:lll' uno' oir,"i Derivatiu;,;; ;;k-schores and
a carl option gives rhe buver

rhe right to buy
'd:l'Jii*^:'Ji:rt::l's' a
ouncesorgora,RounilTf
as a srike price) for
._Tff :lT#:Iff X,!?:S;L:i::,?#*,::iil*;
a specifrec p"rioo or tir.
rs generallv limited to rn9'.u,i,i;;;":r* an opdon
a specifiLa p".iro-."r"rred
dare when a ca, option ro by;; ;;;., expirarion
d,k";;;;;',11, g" il ;#;;#;t
asset, rhe calt is said
to be .,in_the ,"rr.l"; price or the
value of rhe asser The differen;;;;;;;,
the marker
r,ro. i,n* is often referred io ..inrrinsic
value'" when a call"i1.1T r:;;
option strike price ir'g..ur". as
of an asser. the varue.of ,rr. than the current
,il il ..our_of_the_rnoney.,market value
be
no intrinsic varue. Ther"j".", "pir",i'i.
,t"^";d;, based solely on specuration since it has
future price movement. of a
The ou..rrirur* of any
option is iefened to as a
fl:tfr,T*JlJ,J:*"ms mav "*''i'l*"componenrs known as inrrinsic
Typically, investor
caroptioni";;;;;'J:1JxiTil:ffJ;Jr:",TTfi
,?:,",ffi ::Tfl ffi I
price and make a profit, if the stock price moves above the strike price.
Based on either today's prices or the most expected future prices, a patent or
mineral property may have much smaller Npv value compared to the value
based on higher potential future prices and or lower future production costs.
This potential additional value can be captured with either of two
approaches. First, a mathematical formula calted the Black scholes equa-
tion usually is used to value common stock options. Modification of the
Black-Scholes model can make it applicable to valuing patents and mineral
rights. Most people consider this approach to be very mathematically
sophisticated, and therefore, often difficult to explain and sell to manage-
ment. The second approach to capturing the call option value of a patentor
mineral right type of asset is to use expected net present value (ENpv)
analysis. we have seen in previous examples that ENpv can measure both
uncertainty concerning the probability of failure and uncertainty concerning
the magnitude of a parameter and the timing when that parameter might bi
incurred.
Advocates of the option pricing analysis (real-options analysis) technique

often state several perceived disadvantages associated with discounted cash
flow (DCF) valuation of investments using Npv. Four of these commonly
perceived disadvantages fol low:
l. DCF analysis tends to under value investments because use of average

production and product prices and costs may not capture value associ-
ated with possible higher future production, prices and productivity or
lower future costs.
2. Use of today's expectation of investment future cash flow is not
indicative of the true value of an investment and prohibits accounting
for the embedded investment opportunities in long-lived assets.
3. Traditional DCF analysis fixes the knowledge base today for deter-
mining future cash flow over the life of a project, which prohibits
accounting for large investments occurring in stages.
4. DCF analysis constrains investment timing by assuming that invest-
ments cannot be deferred, so investment decisions must be based on
today's information.
The authors of this text disagree with these perceived DCF analysis dis-
advantages and feel they represent misperceptions of the assumptions and
calculations upon which proper DCF analysis should be based. If you based
Chapter 6: Uncertainty and Risk Analysis 317
DCF analysis on the assumptions given in the four DCF disadvantage srate-
ments, then DCF analysis would be improper and disadvantageous. Follow-
ing are our rebuttals to the perceired disudvanrages:
1. Economic analyses of all types require projecting the future. Assum-

ing today's production rates, prices and costs will stay the same over a
project life typicaliy is a very weak economic analysis assumption,
You must project the actual future production rates, product prices,
and costs by projecting escalation rates for the parameters realized
over the project life. If different future cash flow scenarios are felt to
be possible, probabilities of occurence must be projected for the dif-
ferent scenarios so that the cumulative probabilities equal 1.0
2. Investors dc not need to use today's dollars and probably should not
use today's dollars as the basis for most DCF analyses. Nerv project
and development or expansion options can be built into any year of an
evaluation. There is no reason to feel that DCF anaiysis precludes you
from accounting for possible future changes in production, product
prices, costs, productivity and other evaluation parameters. You must
project the future possibilities at the time you do the analysis for either'
DCF analysis or option pricing analysis. And the future assumptions
that you make will have a very significant impact on your analysis
results with either DCF analysis or the option pricing methodology.
3. DCF analysis does not have to assume that project development will
start today. With ENPV analysis you can build development and expan-
sion into analyses over unlimited years, taking advantage of new tech-
nology and product pricing information available at different stages.
4. Investments can be deferred to future times vrith DCF analysis the

same as with options-pricing analysis. Evaluation results with any
method of analysis reflect the input data and the assumptions applied
to the input data such as timing, escalation rates and tax effects in
after-tax analyses. ENPV analysis can be done today for a project
expected to be developed several years in the future. In Example 6-12,
the development occurs today at time zero, but development could
easily be deferred several years with similar analysis procedures.
ur8 Economic Evaluation and lnvestment Decision Methods
Example 6-12 ENPV Analysis to Capture Call Option Value

Consider that you are interested in acquiring the rights to develop
a project. To minimize calculations needed, assume the project has a
four-year life with expected production, prices per unit, operating
costs per unit and capital costs shown in the following table. Calcu-
late the project NPV for minimum discount rates of 10"/" and20%.
Atl values in 000's except selling price
Assuming a1O0o/" Probability of Occurrence, Most Expected!
Year 0 1234
Production, Units 100 100 100 100
Selling Price, $/Unit 20 22 24 26
Gross Revenue, $ 20oo 2200 2400 2600
Operating Costs, $ -1200 -1400 -1600 -1800
Capital Costs, $ -2400
BTCF -2400 800 800 800 800
NPV @ 10% =+136
NPV @ 20/"=-329
Based on these valuations, an investor satisfied with a 10% rate of
return on invested dollars could pay $136 at time zero to acquire the
project rights and an investor desiring a 2O'/" rate of return on
invested dollars would need to be paid $329 at year zerc to take the
project rights and just gel a 20"/" ROR.
Next, instead of using average or most expected values, assume
there is a 6O"k probability of the scenario just described actually occur-
ring. Further, there is a 10% probability of a worse case scenario and
a3O"/" probability of a better scenario as summarized below:
Assuming a6O/o Probability of Most Expected Values
Year 0 1234
Gross Revenue, $ 2000 2200 2400 2600
Operating Costs, $ -1200 -1400 -1600 -1800
Capital Costs, $ -2400
BTCF -2400 800 800 800 800
NPV @ 10% = +136
NPV @ 20/" = -329
Chapter 6: Uncertainty and Risk Anatysis
3't9
Assuming alao/o probabirity of worst case occurrence

Year 0 1234
Gi-oss Revenue, $ 2000 1800 1800 1800
Operating Costs, g
g -12A0 -1S00 -1800 -210O
Capital Costs, -Z4OO
BTCF _2400 800 300 -300
NPV @ 10o/o=-1750
NPV @ 2O/"=-1770
Assuming a3oo/o probability of Best Case Occurrence
Year 1234
Production, Units 100 130 .30
19016C
Seliing Price, $/Unit 22 25 35
Gross Revenue, $ 220A 3250 4800 6650
Operating Costs, g
-1200 -1300 -1400 _1500
Capital Costs, $ -Z3OO
BTCF -2300 1000 1950 3400 5150
NPV @ 10"/"= +6293
NPV @ 2Oo/"= +433g
Combining these results with the probability of occurrence
yields
the project ENpV as follows:
Expected Value al1O"/. Expected Value @ 20"/o
Expected 136(0.6) = 82 Expected -3rr(0^6)-1r?
Worst -1730(0.1)=-178 Worst -1770(0.1) = -1Zl
Best 6293(0.3) = 1969 Best 4309(0.3) = 1302
Expected Value = 1797 Expected Value = g2g
The ENPV ar 1o"k indicates an investor courd pay thirteen
times as
much ($1797 vs $136), for the project as was det,ermined
by the initial
expected value analysis based on average expected values.
Ar a 2o"k
discount rate, the ENPV is $geg versus _$SZg with the
ur"r"g" or most
expected value analysis. These differences reflect the idea
in *option
pricing effects" which were discussed earlier.
The ENpv captures this
concept, but the average or most expected value analysis
alone did not.
The '100% probability of occurrence NPV results based on average

or most expected values may be best for making a "develop" versus
"do nothing" decision. However,'the ENPV results capture the "option
pricing value" by accounting for the 30% probability of realizing
higher production and prices and lower costs. An investor may not
want to base a develop decision based on 30% probability of occur-
rence data, but the same investor may be willing to pay a higher
price to acquire and hoid the project to realize the opportunity to
profit from the best case scenario if it occurs.
6.10 Summary
Sensitivity analyses are a means of identifying those critical variables
that if changed, could considerably impact the profitability measure such as
rate of return or net present value. Three types of sensitivity were addressed
including:
1. Single Variable
2. Multi-Variable or Best Case - Woist Case
3. Probabilistic Sensitivity (Monte Carlo)
Two software models that address the probabilistic sensitivity (Monte
Carlo) are known as @Risk and Crystal Ball. These template programs
interact with most spreadsheets providing users with a low cost methodol-
ogy to these approaches.
Risk analyses identify the likelihood of project failure and the subsequent
cost to the investor. In this text the subject of project risk was addressed uti-
lizing the decision theory approach based on expected value, (EV) which is
defined as:
EV = Expected Profit - Expected Cost
Examples then illustrated several approaches that could be used with equal
validity for the level of complexity presented in the textbook examples and
problems. These approaches included:
1. Chance Factor, (Expected Value)
2. Expected NPV, (Incorporated the probabilities of success and failure
directly into present worth equation)
3. Risk Adjusted Cash Flows, (By applying probabilities of occurrence
directly to the expected values each compounding period, one risk
adjusted cash flow results which can then be manipulated in a manner
consistent with examples and problems prior to Chapter 6.
L
321
of these methods, the Chance Factor, Expected

value methodorogy
broadest range of acceptabilrty for thi more
statisrically comprlx
lr lT[1j*
An alternative method sometimes used to measure
financial or political
risk involves ac1-iusting the ,,1i.scouut rite, i'i lor
greater chances of farlure.
\vhile common in inclustry practice, it is not
a prefen-ed approach. This is
due to the ditficurties that a'se whe;; comparing
prdects ti:at mightbe
competing for limited capital. Discussion in
tt. expectei rate of return soru_
tion to Exarnple 6-6 provided further amplificatio..on
tt is topi,
For more information on these and rerated topics,
see; ,.Decision Anary-
sis.for?etroleum Exploration," by paul D. Newendorp,
or, ,,Mineral Explo-
ration Decisions," by Deverle p. Harris.
Economic Evaluation and lnvestment Decision Methods
PROBLBMS
6-l A roulette wheel has 38 different stopping slots numbered from I to

36 plus 0 and 00. Eighteen numbers are re4 18 are black. with 0 and
00 green. calculate the expected value for the following situations
where the term "payofF' means in addition to return of the bet or
"profit above all costs."
A) The payoff is $35 for a $1 bet on any number.
B) The payoff is $17 for a $1 bet on any line between two numbers
(meaning the bettor wins if either number hits).
c) The payoff is $8 for a $1 bet on any corner between four numbers.

D) The payoff is $1 for a $1 bet on red or black or odd or even.
6-2 Due to uncertainty in development costs, the cost of a new manufac-

turing process is considered to have the following possibilities:
Cost($) Probability of Occurrence
5,000 0.10
8,000 0.30
10,000 0.40
14,000 0.20
What is the expected cost of the new process?
6-3 An electronics manufacturer is considering entering into a research

and development venture. The research and development investment
requires $100,000 at time 0 and, if successful, generates $g0,000 in
profit each year for 5 years. Salvage value is estimated to be zero.
Experience suggests that such projects have a 40vo probability of suc-
cess. If the before-tax minimum rate of retunl is 2ovo, should the man-
ufacturer undertake the project? Use expected NpV and expected ROR
Analysis.
6-4 If you bet $5 on the outcome of 3 football games, you can win $25
total (not above the bet cost) if you correctly pick all 3 winners. If the
odds are considered even in each game or if point spreads are speci-
fied that are considered to make each game an even odds bet, what is
the expected value of your bet if you neglect tie game situations?
l.
Chapter 6: Uncertainty and Flisk Analysis 323
6-5 An initial investment of S50,00i1 is projected to generate cash flows

over a three year project life as follows:
Probabilit-v of
Year Cash Flow Occurrence
1 $25,000 .40
18,000 .60
30,000 .50
20,000 .s0
J 35,000 .70
25,000 .30
Evaluate the expected NPV of this investment for a minimum rate of
return of l5Va.
6-6 A research and development manager associates a probability of suc-

cess of 9.6) with a research investment at time zero being suc-
60Vo
cessful and generating the need for an additional $300,000 develop-
ment investment at the end of year one which is estimated to have a
probability of 907o (0.9) of successfully generaring profits of $200,000
per year for year two through 10, assuming a washout of escalation of
operating costs and sales revenue. If failure occurs after the time zero
research investment a reclamation cost of $100.000 rvill be realized at
the end of year one. If failure occurs after the year one investment, the
salvage value will be $250,000 at the end of year two for equipment
salvage. To achieve a before-tax expected ROR of 25Vo in this invest-
ment, use expected NPV Analysis to determine how much money can
be spent on research at time zero assuming the year l0 salvage value is
zero? What is the risk free project NPV valuation?
6-7 Calculate the expecte,j net present value of a project which will cost
$70,000 at time zero, considering there is a 5OVo chance that the invest-
ment at time zero will be successful, which will require an additional
investment of $120,000 at year one. There is a70Vo chance of success
of the year one investment yielding profits over a six year period (years
2 to 7) equal to $125,000 per year. Failure at time zero will result in an
abandonment cost of $10,000 atyeat one and failure at year one will
result in a salvage value of $50,000 at year two. Should this project be
considered from an economic viewpoint if the minimum rate of return
is 20Vo? Compare your expected NPV result with risk-free NpV.
324 Economic Evaluation and Investment Decision Methods
What additional cost could be incurred at time zero and still give the
investor an ENPV of 2.0 (or $2,000) at time zero? Assume the addi-
tional investment in R&D or geological or geophysical work would
improve the initial probability of success.from 507o to 8-07o..reducing
the initial risk of failure at year 1 to 20Vo.
6-8 A project that would have a time zero cost of $170,000 is estimated to
have a 407a probability of generating :ret income of $60,000 per year
for each of years one through 10 with a zero salvage value, a 307o
probability of generating net income of $50,000 per year for each of
years one through 10 with a zero salvage value, a 207o probability of
generating net incomes of $40,000 per year for each of years one
through 10 with a zero salvage value and a lU%o probability of failing
and generating a $20,000 salvage value at year one. For a minimum
rate of return of 20Vo calculate the project expected NPV. What is the
project expected ROR?
6-9 Two years ago, a petroleum company acquired the mineral rights to a
' property for which an offer of $1 million cash has been received now
at year 0. Development of the property is projected to generate esca-
lated dollar cash flow in millions of dollars of -1.5 in time zero, and
+1.0, +1.8, +1.2, +0.8 and +0.4 in years one through five respectively.
If the minimum DCFROR is 207o, should the company keep and
develop the property or sell if there is considered to be a 607o proba-
bility of development generating the year one through five positive
cash flow and 407a probability of failure generating zero cash flow in
years one through five?
6-10 Two alternatives are being considered for the development of an invest-
ment project. Alternative 'A' would start development now with esti-
mated development and equipment costs of $10 million at time zero
and $20 million at yea-r one to generate net revenues of $6 million in
year one and $12 million per year uniformly at years two through l0
with a zero salvage value. Alternative 'B' would start development ttvo
years from now for estimated development and equipment costs of $15
million at year two and $30 million at year three to generate net rev-
enues of $9 million in year three and $ 18 million per year uniformly at
years four through 12 with a zero salvage value. Use NPV Analysis for
a minimum ROR of 207o to determine the economically better alterna-
tive and verify the NPV results with PVR Analysis assuming:
Chapter 6: Uncertainty and Risk Analysis 325
A) rclEc probahility of success is associated with all investments.
B) Alternative 'A' has a 60Vo probability of success associated with

year 0 cost and 1007o probabiiity of success associated with the
year 1 cost. with zero net salvage r.alue to be realized if failur-e
occurs at year 0. Alternative'B' has an 807o probability of success
associated rvith the year 2 cost and a l00Vo probability of success
associated u,,ith the year 3 cost, with a zeto net salvage value to be
realized if failure occurs at vear 3.
6-t 1 An investor has paid $100,000 for a machine that is estimated to have
a 70Vo probability of successfully producing 5000 product units per
year for each of the next 3 years when the machine is estimated to be
obsolete with a zero salvage value. The product price is the unknown
to be calculated, so it is estimated to be $X per unit in year I escalated
dollars and to increase lla/o in year 2 and 6Vo in year 3. Total operat-
ing costs are estimated to be $8,000 in year I escalated dollars and to
increase l5%o in year 2 and 87a in year 3. The annual inflation rate is
estimated to be 7Vo. What must be the year 1, 2 and 3 escalated dollar
product selling price if the investor is to receive a 127a annually com-
pounded constant dollar expected DCFROR on invested dollars? Con-
sider zero cash flow to be realized the 307a of the time the project
fails. This assumes that equipment dismantlement costs will exactly
offset any salvage value benefits.
6-12 A plant operation is scheduled to be developed for a time zero capital

cost of $400 million with year I through l0 revenues of $200 million
per year less operating costs of $100 million per year with a zero sal-
vage value. Assume a washout of escalation of operating costs and
revenue each year. "Washout" means any operating cost escalation is
offset by the same dollar escalation of revenue (not the same percent
escalation) so profit rernains uniform at the today's dollar value profit.
A) Evaluate the project ROR and analyze the sensitivity of the result
to changing project life to 5 years or l5 years.
B) Evaluate the sensitivity of project ROR to increasing the time zero
capital cost to $600 million and $800 million for the l0 year proj-
ect life.

6 Uncertainty and Risk Analysis

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

6 Uncertainty and Risk Analysis

Uploaded by

Copyright:

Available Formats

CHAPTER 6

UNCERTAINTY AND RISK ANALYSIS

. titative analysis approach is based on analyzing the effects that risk

6.2 Sensitivity Analysis to Analyze Effects of Uncertainty

As described in the previous section, sensitivity analysis refers to analysis

used to determine how much change in a variable would be necessary to

EXAMPLE 6-1 Single Variable Sensitivity Anatysis

C=$2a0,000 $67,000 $67,00Q $67,000 $67,000 $O7,OOO

A) lnitial lnvestment Sensitivity Analysis

B) Project Life Sensitivity Analysis

C) Annual Profit Sensitivity Analysis

Percent variations in ROR due to changes in annual profit are very

D) Salvage Value Sensitivity Anaiysis

Salvage Change in Percent Change in 18.0%

42,004 -40 15.9 -11.9

Sensitivity analysis shows that accuracy of salvage value is the

6.3 The Range Approach to Sensitivity Analysis

EXAMPLE 6-2 Range Approach Sensitivity Analysis

Use the range approach to evaluate the investment described in

Best Case Expected Case Worst Case

6.4 Probabilistic Sensitivity Analysis

permits "probabilistic analysis" by the Monte Carlo simulation technique. A

In principle, a relative frequency distribution graph is converted to the

has a much higher percentage of the parameter values in a given

Figure 6-2 Cumulative Frequency Distribution

Figure 6-3 Frequency Distribution of RoR for Varying simulations

EXAMPLE 6-9 A Simple probabilistic

6.5 Expected Value Analysis (Economic Risk Analysis)

EXAMPLE 6-4 Expected Value Analysis of a Gambling Game

EXAMPLE 6-5 Expected Value Analysis of a Simplistic

Certainly if you have enough investments of this type in which to invest

Let P = probability of success, I -p = probability of failure

6.6 Expected NPV, Expected pVR, and Expected ROR Analysis

EXAMPLE 6-6 Expected Vatue Apptied to ROR,

should the project be done? compare expected varue,

Solution, Values in Dollars:

C = $90,000 P=0.4 | =$50,000 . . . . I =$50,000

A) Expected vatue Anarysis rncruding Time varue of Money

B) Expected Net present Value

For these statistically simplistic models, another approach

ENpv = -e0,000 + 20,000(priliijr, *14,184

Risk Free NpV = (50,000)(pr"^133^,r,- (e0,000) +$ee,540 >

by trial and error, f i" = ROR = 47.60/o> i* = 10% so accept

D) Expected Present Value Flatio Analysis

Risk Free PVR = +99,540/90,000 = +1..l1

Expected value anarysis in general involves

EXAMPLE 6-7 Expected varue Anarysis

Solution, in Thousands of Doltars:

. *Determine the possibte different outcomes and the subsequent

Expected Net present Value (ENPV) at 2}o/o

Alternate Form of (ENpV) Equation

-500(1.0) 450(0.6X0.5) 450(0.6X0.s)

EXAMPLE 6-8 Expected value Economics of a New process

time zero, and, it is considered 100% certain.that experimental

..letermine the possibre different outcomes and the subsequent

For Case A, the chance factor is; (0.6)(0.S)r

-500(1.0) 450(0.6x0.5) 450(0.6X0.s)

ENPV = -100 -756(PlF20,1)+ 220.5(plF29,2)

Expected Net Present Value (ENPV) at20o/o

Expected Present Value Ratio (EPVR)

EXAMPLE 6-9 Expected NPV Anatysis

Solution: All Values in Thousands.

.-$270 P=(0.60) -$500 P=(o.eo) $400. ..$4oo

A) Successful Development, Chance Factor,