3: RISK AND UNCERTAINTY

1 Introduction
1.1 Decision making involves making decisions now which will affect future outcomes and it is
unlikely that future cash flows will be known with certainty.

Risk
1.2 Risk exists where a decision maker has knowledge that several possible future outcomes
are possible, usually due to past experience. This past experience enables a decision
maker to estimate the probability of the likely occurrence of each potential future outcome.
Risk can be quantified.

Uncertainty
1.3 Uncertainty exists when the future is unknown and the decision maker has no past
experience on which to base predictions.
Uncertainty cannot be quantified but techniques can be adopted to reduce uncertainty.
These might include:
• Market research
Allowing for • Focus groups
uncertainty.

2 Risk preference
2.1 (a) Risk seeker – An optimist. A decision maker who is interested in the best outcomes
no matter how small a chance that they may occur.
(b) Risk neutral – A decision maker who is concerned with the most likely outcome.
(c) Risk averse – A pessimist. A decision maker who acts on the assumption that the
worst outcome might occur.

Lecture example 1 Idea Generation

Investment A B
Expected outcome $10,000 $10,000
Highest possible $25,000 $11,000
Lowest possible $(10,000) $9,000
Required
Which investment would be chosen by a decision maker who is:
(a) Risk seeking?
(b) Risk neutral?
(c) Risk averse?

1
 3: RISK AND UNCERTAINTY

Solution

3 Data tables
3.1 If there is one decision and one uncertain variable it is often easiest to display all options on
a data table, which may be generated easily by a spreadsheet.

Lecture example 2 Technique Demonstration

John must decide how best to use a monthly factory capacity of 1,200 units His demand from
regular customers is risky and as follows.
Monthly demand
(units) Probability

300 0.2
500 0.6
700 0.2
1.0
Regular customers generate contribution of $5 per unit. John has the opportunity to enter a special
contract which will generate contribution of only $3 per unit. For the special contract John must
enter a binding agreement now at a level of 900, 700 or 500 units.
Required
Display all possible contributions in a data table.

Solution
Workings:
Special contract (units)
Demand 900 700 500
(units)
300

500

700

2
3: RISK AND UNCERTAINTY

4 Expected values August 2021Q1C

Definition
4.1 Where there is uncertainty and a range of possible future outcomes has been quantified (for
example, best, worst and most likely) probabilities can be assigned to these outcomes and a
weighted average (expected value) of those outcomes calculated.
EV = Σpx
where p is the probability of the outcome occurring and x is the value of the outcome (profit
or cost).

4.2 When faced with a number of alternative decisions, the one with the highest EV should be
chosen.

Lecture example 3 Technique Demonstration

Required
Suppose we assume that in lecture example 2 John wants to maximise profits over the long term.
Find the optimal level of special contract to commit to every month, using expected values.

Solution
Special contract (units)
Demand P 900 700 500
(units)
300 0.2 4,200 3,600 3,000

500 0.6 4,200 4,600 4,000

700 0.2 4,200 4,600 5,000

EV

Workings:

3
 3: RISK AND UNCERTAINTY

Limitations of expected values

4.3 (a) EV is a long-term average, so that the EV will not be reached in the short term and
is therefore not suitable for one-off decisions.
(b) The results are dependent on the accuracy of the probability distribution. In particular,
it uses discrete variables rather than continuous variables (i.e variables are point
estimates rather than a continuous range). This may not accurately model the real
situation.
(c) EV takes no account of the risk associated with a decision.
(d) The EV itself may not represent a single possible outcome.

5 Decision methods December 2021Q3

Maximin decisions
5.1 Maximise the minimum return of each decision.
Risk averse decision-maker.

Lecture example 4 Preparation question

Risk averse
Required
Assuming a totally risk averse attitude, what would be taken using the data table in lecture
example 2?

Solution

Special contract (units)

Demand 900 700 500
(units)
300 4,200 3,600 3,000

500 4,200 4,600 4,000

700 4,200 4,600 5,000

4
3: RISK AND UNCERTAINTY

5.2 Criticisms of maximin

• Ignores the probability of each outcome occurring
• Is conservative (doesn’t try to maximise profit)

Maximax decisions
5.3 Aim for the best possible return.
Risk seeking decision maker.

Lecture example 5 Preparation question

Risk seeking
Required
Assuming a risk seeking attitude, what decision would be taken using the data table in lecture
example 2?

Solution
Special contract (units)
Demand 900 700 500
(units)
300 4,200 3,600 3,000

500 4,200 4,600 4,000

700 4,200 4,600 5,000

5.4 Criticisms of maximax

• Ignores the probability of each outcome occurring
• Is overly optimistic

5
 3: RISK AND UNCERTAINTY

Minimax regret decision rule August 2021Q1C

5.5 'Regret' means opportunity cost from making the wrong decision.
The decision rule chooses the option which minimises the maximum opportunity cost from
making the wrong decision.

Lecture example 6 Preparation question

Required
Using the minimax regret rule, what decision would be taken using the data table in lecture
example 2?

Solution
Special contract (units)
Demand 900 700 500
(units)
300 4,200 3,600 3,000

500 4,200 4,600 4,000

700 4,200 4,600 5,000

Opportunity Cost Table

Special contract (units)
Demand 900 700 500
(units)
300

500

700

Maximum
regret

6
3: RISK AND UNCERTAINTY

6 Perfect information
6.1 Information may be available about uncertain variables eg. market research.
6.2 If this information is guaranteed to predict the future with certainty it is defined as perfect
information.
6.3 Perfect information removes risk. It is therefore valuable.

6.4 Value of perfect information (VOPI)

EV (with perfect information) X

EV (no perfect information) (X)
VOPI X

Lecture example 7
John has been contacted by a market research company, which guarantees that the results of its
survey will be 100% correct.
These results will enable John to ascertain the demand from his regular customers every month, in
advance of accepting the special order.
Required
What is the maximum amount that John should pay for the survey?
Special contract (units)
Demand P 900 700 500
(units)
300 0.2 4,200 3,600 3,000
500 0.6 4,200 4,600 4,000
700 0.2 4,200 4,600 5,000

Solution

7
 3: RISK AND UNCERTAINTY

7 Joint probability tables

7.1 If there are two variables that are uncertain or risky it may be helpful to record the range of
possible outcomes in a joint probability table.

7.2 Analysis could take the form of expected values or the data table could be used to give
management an overview of the decision it is facing.

Lecture example 8 Technique demonstration

Brown Co has developed a new product.

The company is confident that demand for the product will be 30,000 units at a selling price of $25,
but both the variable cost per unit and the specific fixed costs associated with this product are
uncertain.
Brown Co believes that the following circumstances could occur.
VC Prob FC Prob
$ $
12 0.2 100,000 0.4
13 0.35 110,000 0.5
14 0.45 120,000 0.1
Required
(a) Construct a two-way data table for profit generated.
Fixed Costs
$100,000 $110,000 $120,000

$12
VC
$13

$14

8
3: RISK AND UNCERTAINTY

(b) Using the joint probabilities for each combination of fixed cost and variable cost, calculate
the expected value of Brown Co's profit.
Joint probability table
Fixed Costs
$100,000 $110,000 $120,000
Prob 0.4 0.5 0.1
$12 0.2
VC
$13 0.35

$14 0.45

Expected value of profit

Fixed Costs
$100,000 $110,000 $120,000
$12
VC
$13

$14

9
 3: RISK AND UNCERTAINTY

8 Decision trees
8.1 A decision tree is a pictorial method of showing a sequence of interrelated decisions and
their expected outcomes. Decision trees can incorporate both the probabilities of, and value
of, expected outcomes, and are used in decision-making.

8.2 Decision trees are most useful when there are several decisions and ranges of outcome.

Constructing a decision tree

8.3 Constructing a tree requires all the choices and outcomes to be drawn and the numbers
(probabilities, outcomes and EVs) to be entered.
The steps involved are:
(a) Plan the tree diagram and tick off all information given in the question as you use it in
the plan.
(b) Draw the tree from left to right, using a ruler, giving yourself as much space as
possible.
(c) Show a key in the answer detailing the different symbols for decisions and outcomes.

Evaluating a decision tree

8.4 Once drawn the optimal decision can be calculated using rollback analysis.
(a) Evaluate the tree from right to left.
– Calculate expected values at outcome points
– Take highest benefit at decision points
(b) Keep your workings on a different page.
(c) State clearly the initial decision to be made.

Lecture example 9 Technique demonstration

Captain Co runs its business through a number of centres. One of its centres is suffering from
declining sales and management has a range of options:
(a) To shut down the site and sell it for $5 million
(b) To undertake a major refurbishment
(c) To undertake a cheaper refurbishment
In the past 2/3 of such refurbishments have achieved good results, the other 1/3 being less
successful, achieving poor results.
The major refurbishment will cost $4,000,000 now. Estimates of the outcomes are as follows.
(1) Good results PV = $13,500,000
(2) Poor results PV = $6,500,000

10
3: RISK AND UNCERTAINTY

The cheaper refurbishment, costing $2,000,000 now would have the following outcomes:
(1) Good results PV = $8,500,000
(2) Poor results PV = $4,000,000
Required
(a) Construct a decision tree for Captain Co to show all possible decisions and outcomes.
(b) Evaluate the decision tree and recommend what action should be taken.

Solution
Workings:

9 Sensitivity analysis

9.1 Assessing probabilities of a range of variables may be difficult with certainty. Sensitivity
analysis permits an alternative way of assessing risk.
Decisions are assessed for their response to a change in a variable.
9.2 Approaches to sensitivity analysis are:
(a) Calculating the maximum percentage change in a variable before the decision would
change.
(b) Assessing if the decision would change if a variable changed by x% of estimate.
(c) Estimating by how much costs / revenues would need to change before the decision
maker would be indifferent between two options.

11
3: RISK AND UNCERTAINTY

9.3 Limitations of sensitivity analysis

• Only one variable can be tested at a time
• There is no decision rule
• It does not quantify the probability of the variable changing

Lecture example 10 Technique Demonstration

Company P is considering the launch of a new product.

Details are as follows:
$
Sales 10,000 units @ $10 100,000
Material costs $2 per unit 20,000
Labour cost $3 per unit 30,000
(50,000)
Fixed Overheads (30,000)
Profit 20,000
Required
Assess the sensitivity of the product to a change in
(a) Material cost
(b) Units sold

Solution

12
 3: RISK AND UNCERTAINTY

10 Simulation
10.1 Simulation models can be used to deal with decisions where there are a number of
Simulation uncertain variables.
models

10.2 Simulation models can be created using computers and random numbers. These numbers
are linked to probability distributions so that the number chosen occurs with the same
probability that the real life event would occur.

10.3 Simulation can be used for estimating queues in shops as this depends on two
uncertainties; arrival of customers at the shop and service time. Two sets of probabilities
and random numbers will be required.

11 Chapter summary
Section Topic Summary
1 Risk & Uncertainty Risk is where a decision-maker has past experience.
With uncertainty there is no past experience.
2 Risk preference There are three types of risk preference
• Risk seeker – optimist
• Risk averse – pessimist
• Risk neutral – uses expected values
3 Data tables Data tables are used to display all the possible outcomes
when there is one decision and one uncertain variable
4 Expected values Expected values are calculated as ∑ px
5 Decision methods Maximin – maximise the minimum return
Maximax – maximise the maximum return
Minimax regret – minimise the opportunity cost from
making the wrong decision
6 Perfect information Perfect information is guaranteed to predict the future
with 100% accuracy. Imperfect information is valuable
even though it may incorrectly predict future events.
The value of perfect information is calculated as:
$
EV with perfect information X
EV without perfect information (X)
Value of information X

13