Risk Preferences
Individuals are described by their utility functions. These utility functions describe our
preferences for different outcomes. For a given level of wealth W, an individual may have the
utility of wealth as U(W). For wealth, it is reasonable to assume 𝑈𝑈 ′ (𝑊𝑊) > 0, i.e., more is always
preferred to less or “non-satiation.” This means that an increase in wealth will always lead to
an increase in utility, no matter how small. With this, individuals are classified as (1) Risk
neutral; (2) Risk-averse; (3) Risk preferring.
Risk Neutral: If the individual is indifferent in receiving the expected value of gamble (simple
average or probabilistic average in case different outcomes have different probabilities) or the
gamble itself. For example, if you make a bet and the expected value from the bet is Rs 100; If
you are indifferent in directly getting Rs 100 sure shot (without playing this gamble) vs taking
this bet/gamble (which includes different probabilities for different outcomes, including
losses). This also means that for this kind of investor, expected wealth is important for the
investor, not the risk (or variance in wealth, i.e., multiple possible wealth outcomes). Therefore,
for this investor, utility is a linear function of wealth: 𝑈𝑈 ′′ (𝑊𝑊) = 0. This also means that
changing the risk of outcome has no effect on the utility (well-being for a given level of wealth):
U[E(W)] = E[U(W)].
Risk Averse: Utility function of a risk-averse individual is concave, that is 𝑈𝑈 ′′ (𝑊𝑊) < 0. This
also means that the individual prefers a certain (ensured) amount over the bet (gamble) with
the same expected value, i.e., U[E(W)] > E[U(W)]. This also means that risk-averse individuals
prefer less risk to more and demand additional risk-premia to take on the extra-risk. [Think of
two investments, between SBI FD with assured returns of 7% vs Nifty-50 expected returns of
7%, which one would you choose.]
Assume a gamble with two possible outcomes for wealth: 𝑊𝑊1 and 𝑊𝑊2 (shown in Fig. 2), each
with 50% probabilities. Then the expected wealth from this gamble is E(W)=0.5*𝑊𝑊1 +0.5*𝑊𝑊2 .
This will lie midway on the straight-line joining the points 𝑊𝑊1 and 𝑊𝑊2 on the curve. If one
computes the expected utility of wealth E[U(W)], the same will fall midway between U(𝑊𝑊1 )
and U(𝑊𝑊2 ). It is very easy to observe here that, for this kind of utility function: U[E(W)]>
E[U(W)]. Also, as can be seen from the graph, to generate the same level of utility E[U(W)], a
lower level of certain (ensured) wealth CE is required. The more concave this function
(downward slopping), the higher the difference between U[E(W)] and E[U(W)], i.e., the
individual will be more risk-averse. There exists a sure payment [sure payment is called
“Certainty Equivalent” (CE) here] that makes this individual derive the same utility as the
gamble. As can be seen, this certainty equivalent is less than E(W), and this difference
represents the risk-premia of this risk-averse individual. That is, this individual is willing to
take the extra-risk, that is, derive the same expected utility from a risky outcome if the
expectation of outcome is more than CE [which is E(W)-CE]. Or in other words, this risk-
averse individual is willing to settle for a lower wealth outcome with certainty [that is derive
the same utility].
Risk Preferring: A risk-preferring individual would prefer the riskier situation (more
uncertainty). His utility function would be convex, i.e., 𝑈𝑈 ′′ (𝑊𝑊) > 0. This also means that
U[E(W)] < E[U(W)]. However, commonly it is assumed that individuals are rational and risk-
averse.
Figure 1: Different type of utility functions
Fig 2: Risk Aversion and certainty equivalent
Implication for risk (Standard deviation, σ) and Return (R)
For a risk-averse investor, the relationship between return and utility will look something like
as shown in Fig. 3. More return will always have more utility (Non-satiation). However, for a
risk-averse investor, the relationship exhibits diminishing marginal utility. For example, an
increase of Rs 1 Lakhs, change from Rs 1 Lakh to Rs 2 Lakh results in 10 units of additional
utility. However, the next 2 Lakhs offer only five units of additional utility. So, the expected
wealth have to be increased further to generate the same amount of additional utility.
Fig. 3: Return (wealth) utility graph for diminishing marginal utility
Fig. 4: Indifference curves on expected Return-Risk (Std. Dev.)
In the Fig. 4, u1, u2, u3 are indifference curves for an individual. For each of these, u3>u2>u1.
This means, moving up on the indifference curve (i.e., increasing the level of risk) would
require increasing the level of expected returns to maintain the same level of utility. This person
is risk-averse. However, u4 is horizontal. For this individual, the utility remains the same
despite the increase in risk; this means, the person is risk-neutral and is only concerned with
the expected outcome. That is, the marginal utility does not diminish with increasing risk.
Fig. 5: Return (wealth) utility graph for increasing marginal utility
Here, the Return (wealth) utility graphs are shown for people with convex marginal utility. For
this kind of person, the expected return and risk (SD) would be as shown in Fig. 6. For this
person, an increase in risk results in a lower level of return expectations. It means that he enjoys
risk (gambler, speculator).
Fig. 6: Indifference curves on expected Return-Risk (Std. Dev.)
