Chapter 12 Decision-Making under Conditions of Risk and Uncertainty

1 Risk and Uncertainty

A distinction is often drawn by decision theorists between risk and uncertainty
 Risk: a term applied to a situation where there are several possible outcomes and there is
relevant past experience to enable statistical evidence to be produced for predicting the possible
outcomes
 Uncertainty: a term applied to a situation where there are several possible outcomes but there is
little previous statistical evidence to enable probabilities to be attached to possible outcomes

The approaches that rely on specifying the probabilities for alternative courses of action represent
attempts to measure the risk of the alternative course of action.

Because decision problems exist in an uncertain environment, it is necessary to consider those

uncontrollable factors that are outside the decision-maker’s control and that may occur for
alternative courses of actions.

Events / states of nature: in the context of risk and uncertainty, factors that are outside the decision-
maker’s control

Probability: in the context of risk, the likelihood that an event of state of nature will occur, normally
expressed in decimal form with a value between zero and one

A value of zero denotes a nil likelihood of occurrence, whereas a value of one signifies absolute
certainty. The total of the probabilities for events that can possible occur must sum to one.

Probability distribution: list of possible outcomes for an event and the probability that each will
occur

Objective probabilities: probabilities that can be established mathematically or compiled from

historical data
Subjective probabilities: probabilities that are based on an individual’s expert knowledge, past
experience and on observations of current variables which are likely to affect future events

The advantage of the subjective probability approach is that it provides more meaningful information
than stating the most likely outcome.

2 Probability distributions and expected value

The presentation of a probability distribution for each alternative course of action can provide useful
additional information to management, since the distribution indicates the degree of uncertainty
that exists for each alternative course of action. Probability distributions enable management to
consider not only the possible profits for each alternative course of action, but also the amount of
uncertainty that applies to each alternative.

Expected value: calculated by weighting each of the possible outcomes by its associated probability

The sum of the weighted amounts is called the expected value of the probability distribution. In
other words, the expected value is the weighted arithmetic mean of the possible outcomes.

Single most likely estimate: the outcome with the highest probability attached to it

The expected value calculation takes into account the possibility that a range of different profits are
possible and weights these profits by the probability of their occurrence.
Expected value of a decision represents the long-run average outcome that is expected to occur if a
particular course of action is undertaken many times

The expected values are the averages of the possible outcomes based on management estimates.
There is no guarantee that the actual outcome will equal the expected value.

3 Measuring the amount of risk

In addition to the expected values of the profits for the various alternatives, management is also
interested in the degree of uncertainty of the expected future profits.

Standard deviation: the square root of the mean of the squared deviations from the expected value

√∑
n
σ= ¿¿¿
x=1

Coefficient of variation:

Measures such as expected values, standard deviations or coefficient or variations are used to
summarize the characteristics of alternative courses of action, but they are poor substitutes for
representing the probability distributions, since they do not provide the decision-maker with all the
relevant information.

There is an argument for presenting the entire probability distribution directly to the decision-maker.
Such an approach is appropriate when management must select one from a small number of
alternatives, but in situations where many alternatives need to be considered, the examination of
many probability distributions is likely to be difficult and time consuming. In such situations,
management may have no alternative but to compare the expected values and coefficient of
variation.

4 Attitudes to risk by individuals

Risk seekers: an individual who, given a choice between more or less risky alternatives with identical
expected values, prefers the riskier option
Risk averter: an individual who, given a choice between more or less risky alternatives with identical
expected values, prefers the less risky option
Risk neutral: an individual who, given a choice between more or less risky alternatives with
identical expected values, would be indifferent to both alternatives because they have the same
expected values

A risk seeker would generally choose the alternative that has the highest expected value and is
unlikely to be put off by the low probability of any of the potential adverse outcomes.

Expected values represent a long-run average solution, but decisions should not be made on the
basis of expected values alone, since they do not enable the decision-maker’s attitude towards risk to
be taken into account.

As most business managers are unlikely to be neutral towards risk, and business decisions are rarely
repeated, it is unwise for decisions to be made solely on the basis of expected values. At the very
least, expected values should be supplemented with measures of dispersion, and where possible,
decisions should be made after comparing the probability of distributions of the various alternative
courses of action.
5 Decision tree analysis
In practice, more than one variable may be uncertain, and also de value of some variables may be
dependent on the values of other variables. Many outcomes may therefore be possible, and some
outcomes may be dependent on previous outcomes.

Decision tree: a diagram showing several possible courses of action and possible events and the
potential outcomes for each of them

Each alternative course of action or event is represented by a branch, which leads to subsidiary
branches for further courses of action or possible events. Decision trees are designed to illustrate the
full range of alternatives and events that can occur, under all envisaged conditions. The value of a
decision tree is that its logical analysis of a problem enables a complete strategy to drawn up to
cover all eventualities before a firm becomes committed to a scheme.

The box indicates the point at which decisions have to be taken and the branches emanating from it
indicate the available alternative courses of action. The circles indicate the points at which there are
environmental changes that affect the consequences of prior decisions. The branches from these
points indicate the possible types of environment (states of nature) that may occur.

Note that the joint probability of two events occurring together is the probability of one event times
the probability of the other event.

The decision tree provides a convenient means of identifying all the possible alternative courses of
action and their interdependencies. This approach is particularly useful for assisting in the
construction of probability distributions when many combinations of events are possible.

6 Buying perfect and imperfect information

Expected value of perfect information: the maximum amount it is worth paying for additional
information in an uncertain situation, calculated by comparing the expected value of a decision if the
information is acquired against the expected value in the absence of information

In practice, it is unlikely that perfect information is obtainable, but imperfect information may still be
worth obtaining. However, the value of imperfect information will always be less than the value of
perfect information except when both equal zero. This would occur where the additional information
would not change the decision.

7 Maximin, maximan and regret criteria

In situations where uncertainty exists it is not possible to assign meaningful estimates of probabilities
to possible outcomes. Where this situation occurs managers might use any of the following criteria to
make decisions: maximin, maximan or the criterion of regret.

Maximin criterion: a decision rule based on the assumption that the worst possible outcome will
always occur, and the decision-maker should therefore select the largest payoff under this
assumption
Maximax criterion: a decision rule based on the assumption that the best possible outcome will
always occur, and the decision-maker should therefore select the largest payoff
Regret criterion: a decision rule based on the fact that if a decision-maker selects an alternative that
does not turn out to be the best, he or she will experience regret and therefore decisions should be
made that will minimize the maximum possible regret
8 Risk reduction and diversification
It is unwise for a firm to invest all its funds in a single project, since an unfavorable event may occur
that will affect this project and have a dramatic effect on the firm’s total financial position.

Diversification strategy: a strategy of investing in a range of different projects in order to minimize

risk

If the diversification strategy is followed, an unfavorable event that affects one project may have a
relatively lesser effect on the remaining projects and have a small impact on the firm’s overall
position.

The objective in pursuing a diversification strategy is to achieve certain desirable characteristics

regarding risk and expected return.

Risk is eliminated completely when perfect negative correlation exists between the cash flows of the
proposed projects and the cash flows of the existing activities. When the cash flows are perfectly
positively correlation, risk reduction cannot be achieved when the projects are combined. For all
other correlation values, risk reduction advantages can be obtained by investing in projects that are
not perfectly correlated with existing activities.