Chapter 11

Risk Analysis

1
 Outline
   Sources, measures, and perspectives on risk
   Sensitivity analysis
   Scenario analysis
   Break-even analysis
   Hillier model
   Simulation analysis
   Decision tree analysis
   Managing risk
   Project selection under risk
   Risk analysis in practice
   How financial institutions analyse risk

2
 Techniques for Risk Analysis

Techniques for Risk

Analysis

Analysis of Stand- Analysis of

Alone Risk Contextual Risk

Sensitivity Scenario Corporate Market Risk

Analysis Analysis Risk Analysis Analysis

Break-even Hillier
Analysis Model

Simulation Decision tree

Analysis Analysis
3
 Sources and Perspective of Risk
Sources of Risk

• Project-specific risk
• Competitive risk
• Industry-specific risk
• Market risk
• International risk

Perspectives on Risk

• Standalone risk
• Firm risk / Corporate Risk
• Market risk / Systematic Risk
4
 Measures of Risk

Risk refers to variability. It is a complex and multi-faceted

phenomenon. A variety of measures have been used to capture
different facets of risk. The more important ones are:
• Range
• Standard deviation
• Coefficient of variation
• Semi - variance

5
 Sensitivity Analysis
(‘000)
YEAR 0 YEARS 1 - 10
1. INVESTMENT (20,000)
2. SALES 18,000
3. VARIABLE COSTS (66 2/3 % OF SALES) 12,000
4. FIXED COSTS 1,000
5. DEPRECIATION 2,000
6. PRE-TAX PROFIT 3,000
7. TAXES 1,000
8. PROFIT AFTER TAXES 2,000
9. CASH FLOW FROM OPERATION 4,000
10. NET CASH FLOW (20,000) 4,000

NPV = -20,000,000 + 4,000,000 (5.650) = 2,600,000

RS. IN MILLION
RANGE NPV
KEY VARIABLE PESSIMISTIC EXPECTED OPTIMISTIC PESSIMISTIC EXPECTED OPTIMISTIC
INVESTMENT (RS. IN MILLION) 24 20 18 -0.65 2.60 4.22
SALES (RS. IN MILLION) 15 18 21 -1.17 2.60 6.40
VARIABLE COSTS AS A 70 66.66 65 0.34 2.60 3.73
PERCENT OF SALES

FIXED COSTS 1.3 1.0 0.8 1.47 2.60 6 3.33

Merits:

1.1. It shows how robust / vulnerable a project is to changes in

values of underlying variables
2.2. Indicates where further work may be done – explore how the
variability of critical factor be contained
3.3. Appealing – articulates the concerns that project evaluator
have

Demerits:

1.Doesn't show how likely is the change in variable, which affects

NPV
2.Change in only 1 variable is considered, practically impossible
3. Interpretation of results is subjective

7
 Scenario Analysis
• Zen Enterprises is evaluating a Project. Depending on
the response of the market-the factor which is the
largest source of uncertainty for the success of the
project-the management of the firm has identified
three scenarios:
• -The product will have the moderate appeal to a large
segment of customers at a modest price.
• -The product will have strong appeal to a large segment
of market which is high price-sensitive.
• -The product will appeal to a large segment of the
market which will be willing to pay a high price.
• Calculate the NPV, under each scenario.

8
 Scenario Analysis
Procedure
1. Select the factor around which scenarios will be built.
2. Estimate values of each of the variables for each Scenario
3. Calculate NPV / IRR under each scenario
NET PRESENT VALUE FOR THREE SCENARIOS
(RS. IN MILLION)
SCENARIO 1 SCENARIO 2 SCENARIO 3
INITIAL INVESTMENT 200 200 200
UNIT SELLING PRICE (IN RUPEES) 25 15 40
DEMAND (IN UNITS) 20 40 10
REVENUES 500 600 400
VARIABLE COSTS 240 480 120
FIXED COSTS 50 50 50
DEPRECIATION 20 20 20
PRE-TAX PROFIT 190 50 210
TAX @ 50% 95 25 105
PROFIT AFTER TAX 95 25 105
ANNUAL CASH FLOW 115 45 125
PROJECT LIFE 10 YEARS 10 YEARS 10 YEARS
SALVAGE VALUE 0 0 0
NET PRESENT VALUE (AT A DISCOUNT 377.2 25.9 427.4
RATE OF 15 PERCENT) 9
Limitation:

1.Based on the assumption that there are few well delineated scenarios
2.Scenario analysis expands the concept of estimating the expected
value

10
 Break-Even Analysis
• Accounting Break –Even Analysis
Fixed Costs + Depreciation 1+2
= = Rs. 9 million
Contribution margin ratio 0.333

Cash flow forecast for Naveen’s flour mill project

(‘000)
Year 0 Year 1 - 10
1. Investment (20,000)
2. Sales 18,000
3. Variable costs (66 2/3% of sales) 12,000
4. Fixed costs 1,000
5. Depreciation 2,000
6. Pre-tax profit 3,000
7. Taxes 1,000
8. Profit after taxes 2,000
9. Cash flow from operation 4,000
10. Net cash flow (20,000) 4,000

• Cash Break-Even Analysis 11

 Hillier Model
Uncorrelated Cash Flows
n Ct
NPV =  –I
t = 1 (1 + i)t
n t2 ½
 (NPV) = 
t = 1 (1 + i)2t
Perfectly Correlated Cash Flows
n Ct
NPV =  –I
t=1 (1 + i) t
n t
 (NPV) = 
12
t=1 (1 + i)t
 Simulation Analysis
Procedure
1. Choose variables whose expected values will be
replaced with distributions
2. Specify the probability distributions of these
variables
3. Draw values at random and calculate NPV
4. Repeat 3 many times and plot distribution
5. Evaluate the results

13
Obtaining Probability Distributions of Basic Variables

Defining the probability distributions of basic variables is

an important step in simulation. Two approaches may be
used for obtaining probability distributions:
• Portrait approach
• Building block approach

14
 Portrait Approach.
The portrait approach is similar to the portrait method used for
identifying suspects. According to this approach a standard
probability distribution (normal, beta, chi-square, poisson,
uniform, exponential, or any other) is drawn up, usually by a
statistician, on the basis of the judgment expressed by the
expert (informant). This is shown to the expert for his
comments. The expert may suggest changes if the distribution
does not conform with his judgment. For example, he may
suggest that the probabilities at the tails should be greater or
the probability of the modal value should be higher. The
statistician modifies the earlier distribution to incorporate the
changes suggested by the expert, till he is satisfied that the
probability distribution represents his judgment well.
15
 Building Block Approach.
In the second approach, the ‘building block’ approach the
probability distribution is defined by the expert. He attempts to
quantify his judgment by a procedure which is as follows: (i) he
chooses the range encompassing possible values; (ii) he divides
the range into intervals which he thinks have different
probabilities associated with them; (iii) he assigns probabilities to
these intervals such that  pi = 1; (iv) he may divide intervals into
sub-intervals if he feels that the probabilities within an interval
are different; and (v) he continues this process till he arrives at a
distribution which represents his judgment well.
This process often leads to a step rectangular distribution
and has the following advantages (i) the expert has complete
freedom in expressing his judgment; and (ii) it squares well with
the principle of using all available information, no more no less.
16
 Some Probability Distributions

(a) Uniform Distribution (b) Trapezoidal Distribution

(c) Step Rectangular Distribution (d) Normal Distribution

17
 Problem of Correlation and the Level of
Disaggregation
• In practice, correlations may exist among the distribution of
several factors. When such a dependency exists the factors which
are correlated should be considered together. For this purpose,
the joint probability distribution of correlated factors have to be
developed. This adds immensely to the problem of estimation.
• In this context we must consider the choice relating to the level of
disaggregation. Now the problem is, to which level of detail
should we go?
• The choice of the level of aggregation or disaggregation would be
finally based on the trade-off between the advantages of clarity of
judgment and the complexities of disaggregated analysis. Since
the influence of correlations is more significant than that of the
shape of any particular distribution, it may be preferable to limit
18
disaggregation.
 Issues in Applying Simulation

• What should the output be ?

• Is project variability enough?

• How should the extreme values be used ?

• How should the results of simulation be used?

19
 World Bank’s Experience
It may be instructive here to review the experience of World
Bank. The following are its summary remarks on simulation.
• Simulation is a powerful technique which permits use of
a great deal of information which would otherwise be lost.
• It is a highly efficient medium of communication
• It is not a technique which replaces skilled judgment. On
the contrary, it often requires the use of far more judgment
than the traditional analysis.
• Despite the method’s value, the treatment of correlations
between variables remains a major problem. It is clear that
results can be completely misleading if correlations are not
handled properly. 20
 Evaluation
An increasingly popular tool of risk analysis, simulation offers
certain advantages:
• Its principal strength lies in its versatility. It can handle
problems characterised by (a) numerous exogenous variables
following any kind of distribution, and (b) complex
interrelationships among parameters, exogenous variables,
and endogenous variables. Such problems often defy the
capabilities of analytical methods.
• It compels the decision maker to explicitly consider the
interdependencies and uncertainties characterising the
project.
Simulation , however, is a controversial tool which suffers from
several shortcomings.
• It is difficult to model the project and specify the probability
distributions of exogenous variables 21
• Simulation is inherently imprecise. It provides a rough
approximation of the probability distribution of net present
value ( or any other criterion of merit). Due to its imprecision,
the simulated probability distribution may be misleading when
a
tail of the distribution is critical.
• A realistic simulation model, likely to be complex, would most
probably be constructed by a management scientist, not the
decision maker. The decision maker, lacking understanding of
the model, may not use it.
• To determine the net present value in a simulation run the risk-
free discount rate is used. This is done to avoid prejudging risk
which is supposed to be reflected in the dispersion of the
distribution of net present value. Thus the measure of net
present value takes a meaning, very different from its usual 22
one,
 Decision Tree Analysis
• Decision tree analysis is a tool for analysing situations where
sequential decision making in face of risk is involved.
• The key steps in decision tree analysis are:
1.    Identifying the problem and alternatives
2.    Delineating the decision tree
3. Specifying probabilities and monetary
outcomes
4. Evaluating various decision alternatives

23
 Decision Tree
The decision tree, exhibiting the anatomy of the decision situation,
shows :
  The decision points (also called decision forks) and the alternative
 options available for experimentation and action at these decision
 points.
  The chance points (also called chance forks) where outcomes are
dependent on a chance process and the likely outcomes at these
points.
The decision tree reflects in a diagrammatic form the nature of the
decision situation in terms of alternative courses of action and chance
outcomes which have been identified in the first step of the analysis.
A decision tree can easily become very complex and cumbersome if
an attempt is made to consider the myriad possible future events and
decisions. Such a decision tree, however, is not likely to be a very useful
tool of analysis. Over-elaborate, it may obfuscate the critical issues.
Hence an effort should be made to keep the decision tree somewhat
simple so that the decision makers can focus their attention on major
future alternatives without being drowned in a mass of trivia 24
 Specification of Probabilities and Monetary Value
of Outcomes
Once the decision tree is delineated, the following data have to be
gathered :
  Probabilities associated with each of the possible outcomes at
various chance forks, and
  Monetary value of each combination of decision alternative and
chance outcome.
The probabilities of various outcomes may sometimes be defined
objectively. For example, the probability of a good monsoon may
be
based on objective, historical data. More often, however, the
possible
outcomes encountered in real life are such that objective
probabilities for them cannot be obtained. How can you, for
example, 25
define objectively the probability that a new product like an
 Evaluation of Alternatives
Once the decision tree is delineated and data about probabilities and monetary
values gathered, decision alternatives may be evaluated as follows :
1. Start at the right-hand end of the tree and calculate the expected
monetary value at various chance points that come first as we proceed
leftward.
 
2. Given the expected monetary values of chance points in step 1, evaluate
the alternatives at the final stage decision points in terms of their
expected monetary values.
3.  At each of the final stage decision points, select the alternative which has
the highest expected monetary value and truncate the other alternatives.
Each decision point is assigned a value equal to the expected monetary
value of the alternative selected at that decision point.
4. Proceed backward (leftward) in the same manner, calculating the
expected monetary value at chance points, selecting the decision
alternative which has the highest expected monetary value at various
decision points, truncating inferior decision alternatives, and assigning
values to decision points, till the first decision point is reached. 26
 Vigyanik case
The scientists at Vigyanik have come up with an electric
moped. The firm is ready for pilot production and test
marketing. This will cost Rs.20 million and take six months.
Management believes that there is a 70 percent chance that the
pilot production and test marketing will be successful. In case
of success, Vigyanik can build a plant costing Rs.150 million.
The plant will generate an annual cash inflow of Rs.30 million
for 20 years if the demand is high or an annual cash inflow of
Rs.20 million if the demand is moderate. High demand has a
probability of 0.6; Moderate demand has a probability of 0.4.
To analyse such situations where sequential decision making is
involved decision tree analysis is helpful.

27
 Vigyanik Case

C21 : High
demand Annual
cash flow
Probability 30 million
: 0.6
D21:Invest
c2
-Rs 150
million C22 : Moderate Annual
C11 : Success demand cash flow
D2 Probability
Probability 20 million
D11: Carry out pilot
: 0.4
production and : 0.7 D22: Stop
market test c1
-Rs 20
million
C12 : Failure D31: Stop
D1 D3
Probability : 0.3

D12:Do nothing

28
 Vigyanik Case
The alternatives in the decision tree shown are evaluated as follows:
1. Start at the right-hand end of the tree and calculate the EMV at chance point C 2 that
comes first as we proceed leftward.
EMV(C2) = 0.6 [30xPVIFA (20, 12%)] + 0.4 [20 x PVIFA (20, 12%)]
= Rs.194.2 million
2. Evaluate the EMV of the decision alternatives at D 2 the last stage decision point.
Alternative EMV
D21 (Invest Rs.150 million)Rs.44.2 million
D22 (Stop) 0
3. Select D21 and truncate D22 as EMV(D21) > EMV(D22).
4. Calculate the EMV at chance point C1 that comes next as we roll backwards.
EMV (C1) = 0.7 [44.2] + 0.3 [0] = Rs.30.9 million
5. Evaluate the EMV of the decision alternatives at D 1 the first stage decision point :
Alternative EMV
D11 (Carry out pilot production and
market test at a cost of Rs.20 million) Rs.10.9 million
D12 (Do nothing) 0
Based on the above evaluation, we find that the optimal decision strategy is as follows : Choose D 11
(carry out pilot production and market test) at the decision point D 1 and wait for the outcome at
29
the chance point C1. If the outcome at C1 is C11 (success), invest Rs.150 million; if the outcome at
 Airways Limited Case
Airways Limited has been set up to run an air taxi service in western India. The
company is debating whether it should buy a turboprop aircraft or a piston engine
aircraft. The turboprop aircraft costs 3500 and has a larger capacity. It will serve
if the demand turns out to be high. The piston engine aircraft costs 1800 and has a
smaller capacity. It will serve if the demand is low, but it will not suffice if the
demand is high.
The company believes that the chances of demand being high and low in
year 1 are 0.6 and 0.4. If the demand is high in year 1, there is an 80 percent
chance that it will be high in subsequent years (year 2 onward) and a 20 percent
chance that it will be low in subsequent years.
The technical director of Airways Limited thinks that if the company
buys a piston engine aircraft now and the demand turns out to be high the
company can buy a second-hand piston engine aircraft for 1400 at the end of year
1. This would double its capacity and enable it to cope reasonably well with high
demand from year 2 onwards.
The payoffs associated with high and low demand for various decision
alternatives are shown in Exhibit 1.1.The payoffs shown for year 1 are the payoffs
occurring at the end of year 1 and the payoffs shown for year 2 are the payoffs for
year 2 and the subsequent years, evaluated as of year 2, using a discount rate of 12
percent which is the weighted average cost of capital for Airways Limited.
30
Exhibit 1.1 Decision Tree Year 1
Year 2
High demand
(0.8) 7000
High demand (0.6)
C2 Low demand
1000 (0.2)
1000

C1
High demand
Turboprop
(0.4)
- 4000
7000
Low demand (0.4)
C3 Low demand
200 (0.6)
600
High demand
(0.8)
6000
Low demand
D1 C5
(0.2)
High demand Expand
- 1400 600
(0.6)
D2 High demand
500 (0.8)
Do not
expand 2500
Piston engine
C6 Low demand
- 1800 (0.2)
800
C High demand
4 (0.4)
2500
Low demand (0.4)
C7 Low demand
300 (0.6)
800
31
• If Airways Ltd. buys the turboprop aircraft, there are no
further decisions to be made. The NPV will be
• = -4000 + 0.6(1000) + 0.4(200) +
• 1.12
• +0.6{0.8(7000)+0.2(1000)}+0.4 {0.4(7000)+0.6(600)}
• (1.12) (1.12)
• =389
• If Airways Ltd. buys Piston Engine Aircraft and demand in
year 1 turns out to be high, a further decision has to be
made with respect to capacity expansion. To evaluate the
Piston engine Aircraft, proceed as follows:
• First calculate the NPV of the two options, viz. expand and
do not expand at Decision Point D2

32
• Expand NPV= 0.8(6000)+0.2(600) - 1400 = 2923
• 1.12
• Do Not Expand
• NPV= 0.8(2500) +0.2(800) = 1929
• 1.12
• Truncate Do not expand as it is inferior. This means NPV
at D2 will be 2923
• Third calculate the NPV of the Piston Engine Aircraft at
D1
• NPV= -1800 + 0.6(500+2923) + 0.4(300)
1.12
+ 0.4{0.4(2500)+0.6(800)}
(1.12) (1.12)
= 613
33
• Since the NPV of the Piston Engine Aircraft (613) is
greater than the NPV of the turboprop aircraft(389), the
former is better bet. So the recommended strategy for
Airways Ltd. is to invest in the Piston Engine Aircraft at
Decision Point D1, and if the Demand in year 1 turns out
to be high, expand capacity by buying another Piston
Engine Aircraft.
• OPTION TO EXPAND Note if Airways Ltd. does not
have the option of expanding at the end of year 1, the NPV
of the Piston Engine Aircraft, would be =
• -1800+ 0.6(500) + 0.4(300) +
• 1.12
• +0.6{0.8(2500)+0.2(800)}+0.4{0.4(2500)+0.6(800)}
• (1.12) (1.12)
• = 80 Thus the 0ption to expand has value 613-80=533
34
Given the decision tree with abandonment possibilities, let us calculate the NPV of
the turboprop aircraft and the piston engine aircraft.

0.6 [1000 + {0.8(7000) + 0.2 (1000)}/(1.12)] + 0.4 (200 + 3600)

NPV (Turboprop) = -4000 +
(1.12)
= 667

0.6 (500 + 2923) + 0.4 (300 + 1400)

NPV (Piston engine) = -1800 + = 641
(1.12)
Note that the possibility of abandonment increases the NPV of the Turboprop
aircraft from 389 to 667. This means that the value of the option to abandon is:

Value of abandonment option = NPV with abandonment - NPV without abandonment

= 667 - 389 = 278

For the piston engine aircraft the possibility of abandonment increases the NPV
from 613 to 641. Hence the value of the abandonment option is 28.

35
 Corporate Risk Analysis

• A project’s corporate risk is its contribution to the overall

risk of the firm

• On a stand-alone basis a project may be very risky but if its

returns are not highly correlated – or, even better, negatively

correlated — with the returns on the other projects of the

firm, its corporate risk tends to be low

36
 Managing Risk

• Fixed and variable cost

• Pricing strategy
• Sequential investment
• Improving information
• Financial leverage
• Insurance
• Long-term arrangements
• Strategic alliance
• Derivatives

37
 Project Selection Under Risk

• Judgmental Evaluation

• Payback Period Requirement

• Risk Adjusted Discount Rate

• Certainty Equivalent Method

38
 Risk Analysis in Practice

• Conservative Estimation of Revenues

• Safety Margin in Cost Figures

• Flexible Investment Yardsticks

• Acceptable Overall Certainty Index

• Judgment on three Point Estimates

39
 Relative Importance of Various Methods of
Assessing Project Risk
A survey of corporate finance practices in India found the
relative importance of various methods of assessing project risk
to be as follows:
% of companies rating it as
very important or important
• Sensitivity analysis 90.10
• Scenario analysis 61.60
• Risk-adjusted discount rate 31.70
• Decision tree analysis 12.20
• Monte Carlo simulation 8.20
Source : Manoj Anand “Corporate Finance Practices in India: A
Survey”. Vikalpa, October –December 2000. 40
 How Financial Institutions Analyse Risk
To evaluate the risk dimensions of a project financial
institutions calculate several indicators, the most important
ones being the break even point, the debt service coverage
ratio, and the fixed assets coverage ratio. In addition, they
carry out sensitivity analysis.
The break-even point for a project is calculated with
reference to the year when the project is expected to reach its
target(or expected) level of capacity utilisation, which is usually
the third or the fourth operating year. Further, it is calculated
in terms of capacity utilisation. So it is called break-even point
capacity utilisation (BEPCU).

41
 Summary
  Risk is inherent in almost every business decision. More so in capital
budgeting decisions as they involve costs and benefits extending
over
a long period of time during which many things can change in
unanticipated ways.
 Investment proposals, however, differ in risk. A research and
development project is typically more risky than an expansion
project and the latter tends to be more risky than replacement
project.
  The variety of techniques developed to handle risk in capital
budgeting fall into two broad categories : (i) approaches that
consider the stand-alone risk of a project (sensitivity analysis,
scenario analysis, break-even analysis, Hillier model, simulation
analysis, and decision tree analysis). (ii) approaches that consider
the
contextual risk of a project (corporate risk analysis and market risk
analysis).
 Sensitivity analysis or “what if” analysis answers questions like42 :
“What happens to NPV or IRR if sales decline by 5 per cent or 10
  In sensitivity analysis, typically one variable is varied at a time. If variables
are inter-related, as they are most likely to be, it is helpful to look at
implications of some plausible scenarios, each scenario representing a
consistent combination of variables. Firms often do another kind of
scenario
analysis called the best case and worst case analysis.
 As a financial manager, you would be interested in knowing how much
should be produced and sold at a minimum to ensure that the project does
not ‘lose money’. Such an exercise is called break-even analysis and the
minimum quantity at which loss is avoided is called the break-even point.
The beak-even point may be defined in accounting terms or financial
terms.
  Under certain circumstances, the expected NPV and the standard deviation
of NPV may be obtained through analytical derivation; as proposed by
H.S.
Hillier.
  Sensitivity analysis indicates the sensitivity of the criterion of merit (NPV,
IRR or any other) to variations in basic factors. Though useful, such
information may not be adequate for decision making. The decision maker
would also like to know the likelihood of such occurrences.43 This
information
  Decision tree analysis is a useful tool for analysing sequential decisions in
the face of risk. The key steps in decision tree analysis are : (i)
identification
of the problem and alternatives. (ii) delineation of the decision tree. (iii)
specification of probabilities and monetary outcomes. (iv) evaluation of
various decision alternatives
  A project’s corporate risk is its contribution to the overall risk of the firm.
Put differently, it reflects the impact of the project on the risk profile of
the
firm’s total cash flows. On a stand-alone basis a project may be very risky
but if its returns are not highly correlated – or, even better, negatively
correlated – with the returns on the other projects of the firm, its
corporate
risk tends to be low.
 Aware of the benefits of portfolio diversification, many firms consciously
pursue a strategy of diversification. The logic of corporate diversification
for 44
reducing risk, however, has been questioned. Why should a firm diversify
  Once information about expected return (measured as NPV or IRR or
some other criterion of merit) and variability of return (measured in
terms
of range or standard deviation or some other risk index) has been
gathered,
the next question is, should the project be accepted or rejected. There are
several ways of incorporating risk in the decision process : judgmental
evaluation, payback period requirement, risk-adjusted discount rate
method, and the certainty equivalent method.
  Often managers look at risk and return characteristics of a project and
decide judgmentally whether the project should be accepted or rejected.
Although judgmental decision making may appear highly subjective or
haphazard, this is how most of us make important decisions in our
personal
life.
 In many situations companies use NPV or IRR as the principal selection
criterion, but apply a payback period requirement to control for risk. If
an
investment is considered more risky, a shorter payback period is
required. 45
  Under the risk profile method, the probability distribution of NPV, an
  The risk-adjusted discount rate method calls for adjusting the discount
rate to reflect project risk. If the project risk is same as the risk of the
existing investments of the firm, the discount rate used is the WACC of
the firm; if the project risk is greater (lesser) than the existing
investments of the firm, the discount rate used is higher (lower) than the
WACC of the firm.
  Under the certainty equivalent method, the expected cash flows of the
project are converted into their certainty equivalents by applying
suitable
certainty equivalent coefficients. Then, the risk-free rate is applied for
discounting purposes.
 The methods of risk analysis commonly used in practice are : (i)
conservative estimation of revenues, (ii) safety margin in cost figures,
(iii)
flexible investment yardsticks, (iv) acceptable overall certainty index,
and
(iv) judgment on three-point estimates. 46
  The analysis of risk factor in practice can be improved if the probability