EC3000 Advanced Microeconomics 2023-24

Seminar 2: Problem Set

Answers

Module leader: Guillaume Wilemme - [email protected]

Seminar Tutor: Mert Gumren - [email protected]

Seminars play a pivotal role in assimilating the content presented in lectures and in
achieving optimal readiness for both the midterm and final exams. Engaging in seminars
encompasses three essential actions for students:

• Preparing the problem set in the week before the seminar.

• Actively participating during the seminar sessions,

• Reviewing and repeating the seminars’ questions and exercises.

Each seminar is structured around a problem set that comprises short-essay questions,
requiring minimal mathematical manipulation, as well as numerical exercises. These
questions and exercises closely mirror the difficulty and subject matter of the midterm
and final exams. Questions marked with the symbol ♠ signify tasks that students may
choose to prioritise. Students are invited to think about the takeaway or morale of each
question and exercise.

Questions
Question 1. ♠ Steve has a lottery ticket that gives him £100 with probability 0.1. A
company proposes to buy his ticket for £6. Steve’s preferences are such that his risk
premium is £3. What is the certainty equivalent of this lottery ticket for Steve? You will
also explain carefully the concepts of certainty equivalent and risk premium. Does Steve
sell his ticket for £6? Why?

Answer — —Possible answer—

1
 The certainty equivalent of a risky asset for an individual is the fixed and deterministic
amount that provides the same expected utility to the individual. It is the minimum price
the individual would accept to sell this asset.
The risk premium of a given asset is the difference between the expected returns of this
asset and the certainty equivalent. The risk premium is the maximum price an individual
would pay to trade their risky asset for a riskless asset that provides the same expected
returns (not expected utility!).
To obtain the certainty equivalent, we use the equation that involves the risk premium.
The risk premium is £3 and the expected returns of the asset is 0.1 × 100 + 0.9 × 0 = £10.
The certainty equivalent is the difference: £7. It means that the lottery ticket provides
Steve with the same expected utility as a certain value of £7. Or, in other words, Steve
is indifferent between having his lottery ticket or having £7.
We can therefore answer the second part of the question. By definition of the certainty
equivalent, Steve would accept to trade his ticket for at least £7, his certainty equivalent.
As 6<7, Steve will not sell his ticket. If the company would have proposed him £8, Steve
would have accepted and would have been better off. —

Question 2. Many celebrities buy insurance for particular parts of their body. Singers
insure vocal chords, athletes insure arms or legs. Rumors say Heidi Klum has insured
her legs for $2 million each, Daniel Craig has insured his full body for $9.5 million and
Jennifer Lopez has insured her bottom for $27 million. Real Madrid insured Cristiano
Ronaldo’s legs for $144 million.
How can we make sense of these numbers? Use a maximum of 200 words.

Answer — —Possible answer. A good answer should explain that these purchases cor-
respond to a risk that involves very high costs.—
It is important to note that the insured parts of the body relate to the economic
activity of the person. Cristiano Ronaldo or Heidi Klum did not get insurance for their
vocal chords for instance. These famous persons face a risk that involves large costs.
If Daniel Craig has a harmful car crash then he may not play James Bond again. The
economic loss due to such an unfortunate accident is massive.
If these individuals are risk-averse, it makes sense for them to buy insurance. The
price of these insurance products are very high because the income loss to cover in case of
unfortunate events is very high as well. Note that this price must be much lower than the
compensation in case of unfortunate event. For example, we expect Cristiano Ronaldo
to receive way more than $144 million if he could not play football anymore due to an
injury. —

Question 3. In the theory we use, individual attitudes towards risk is a matter of pref-

2
erences. The theory can tell the best choice for an individual given their utility function.
What about groups of individuals? What should be the best choice under uncertainty for
a government? Arrow and Lind use the risk-spreading result to answer that question.

Arrow and Lind (1970) write:1

‘...when the risks associated with a public investment are publicly borne, the
total cost of risk-bearing is insignificant and, therefore, the government should
ignore uncertainty in evaluating public investments...This result is obtained
not because the government is able to pool investments but because the govern-
ment distributes the risk associated with any investment among a large number
of people. It is the risk-spreading aspect of government investment that is es-
sential to this result.’

Explain carefully the point of Arrow and Lind. Why can government ‘ignore uncertainty’ ?
Your answer must contain less than 200 words.

Answer — —Possible answer—

Governments, like many economic agents, must make decisions in an uncertain envi-
ronment. Arrow and Lind discuss how risks should be taken into account by public actors.
In lecture 1, we saw that risk affects decisions of private actors. In particular, risk-averse
individuals do not in general make the decisions that maximise expected returns.
Arrow and Lind suggest that decision-making for public actors is different because
of the risk-spreading result. A government is able to efficiently spread the risk of any
investment among the many taxpayers. Spreading the initial risk into many smaller
risks is efficient. In lecture 2, we saw that the risk disappears as the number of risk-
averse individuals sharing the risk increases. The group of many risk-averse individuals
is equivalent to a risk-neutral agent. The government can therefore takes decision as if it
was a risk-neutral agent. By definition, risk-neutral agents make decisions that maximise
expected returns, ignoring risks. —

Question 4. ♠ ‘Don’t put all your eggs in one basket.’ We could add at the end of the
sentence: ‘...even if it is the most solid basket you have.’
Comment this maxim in the light of portfolio diversification theory, using a maximum of
200 words.

Answer — —Possible answer. I first explain the meaning. I then state the diversification
result and highlight the connection with the maxim. Lastly, I comment on the maxim’s
extension.—
1
Arrow and Lind. ‘Uncertainty and the Evaluation of Public Investment Decisions.’ American Eco-
nomic Review, 1970.

3
 In this sentence, we can interpret baskets as risky assets and the eggs as capital to
invest. The maxim would thus mean not to invest entirely in one asset.
According to portfolio diversification theory (chapter 2 in coursebook), risks can be
reduced by mixing multiple assets. If there are two baskets to transport the eggs, then the
diversification result states that it is optimal to put some eggs in one basket and some in
the other instead of putting all the eggs in one basket. This is exaclty the recommendation
of the maxim.
The diversification result holds even if one asset is riskier. Therefore, even if one basket
is identified as the most solid and safe, it is still optimal to use the second, less safe, one.
—

Exercises

Exercise 1 (Risk spreading). ♠ This exercise illustrates risk spreading through an

example.

Daniel has an asset that yields X = £200 with probability 3/4 and X = £100 with
probability 1/4. His preferences are such that the certainty equivalent of this asset for
him is £168.

1. Why is Daniel risk-averse?

2. What is the maximum price a risk-neutral individual would pay to buy this asset?
Can Daniel pass a deal with a risk-neutral individual?

3. Daniel asks Ashley if she is interested in buying the asset. Ashley’s preferences are
captured by the utility function u(x) = (x + 100)1/4 . Show that Ashley would pay
at most £165 to buy this asset. Can Daniel sell his asset to Ashley?

4. Ashley has a twin sister, Mary-Kate, who has the same preferences over risk. Daniel
considers selling half of the asset, namely Y ≡ X/2, to each one. Show that Ashley
and Mary-Kate would each pay at most £85 to buy half of the asset.

5. Conclude and explain the gains from risk spreading.

Answer —

1. Usually, we would check that Daniel is risk-averse by showing that his utility function
is concave. Here, we need to find another strategy.
The expected returns of the asset are E[X] = 150+25 = 175, which is above Daniel’s

4
 certainty equivalent. This means that Daniel would be better off with having £175
instead of his asset. He is therefore risk-averse. Alternatively, we could have shown
that the risk premium is positive.

2. The preferences of risk-neutral individuals can be captured by linear utility func-

tions. For them, utility is maximised when expected returns are maximised. A
risk-neutral individual is therefore indifferent between having the asset and having
the expected returns £175 for sure. Thus, a risk-neutral individual would pay at
most £175 for the asset.
From the definition of certainty equivalent, Daniel would require at least £168 to
sell his asset. Since 168 < 175, a mutual agreement is possible.

3. Ashley would pay at most the price p that make her indifferent between buying the
asset and not doing anything. If she refuses the deal, she gets u(0) ≈ 3.16. If she
accepts the deal, she gets expected utility

3 1
E[u(X − p)] = u(200 − p) + u(100 − p)
4 4
3 1
= (300 − p) + (200 − p)1/4
1/4
4 4

We can check that E[u(X − 165)] ≈ 3.16. Ashley would therefore pay at most £165
for the asset.
A deal is not feasible because 165 < 168. Ashley does not want to pay enough, or,
equivalently, Daniel would not accept a sufficiently low price.

4. If Ashley refuses to buy half of the asset, she still gets u(0) ≈ 3.16. If she accepts
at a price p, she gets

3 1
E[u(Y − p)] = u(100 − p) + u(50 − p)
4 4
3 1
= (200 − p)1/4 + (150 − p)1/4
4 4

We can check that E[u(Y − 85)] ≈ 3.16. Ashley would therefore pay at most £85
for half of the asset. The same equations applies to Mary-Kate.

5. Together, Ashley and Mary-Kate would be able to pay up to 85 + 85 = £170 to buy

the entire asset. This is high enough for Daniel since 170 > 168.
By sharing the risks, several individuals can bear larger risks than if they were alone.
This is risk spreading.

5
Exercise 2 (Risk pooling with correlated risks). This exercise provides an example
of risk pooling with the risks can be negatively or positively correlated.

A winegrower has a vineyard in region A. It regularly happens that a bad weather

entirely destroys a vineyard. Denote XA the economic loss when the vineyard is de-
stroyed. XA = −100 when the vineyard is destroyed, and XA = 0 otherwise. Another
winegrower, with a vineyard in region B has the same problem with the same economic
risk: XB = −100 when the vineyard is destroyed, and XB = 0 otherwise.
Depending on the geographical proximity between the two regions, the risks can be
correlated or not. The probabilities of each possible state is given by the following table:

XA = 0 XA = −100
1+α 1−α
XB = 0 4 4
1−α 1+α
XB = −100 4 4

where −1 ≤ α ≤ 1. For instance, the probability that the weather is catastrophic in

region A and B at the same time is 1+α
4
(bottom-right). Denote u(x) the utility function
of each winegrower.

1. What happens when α = 1, 0 and −1? Explain what α captures.

2. What are the expectations of these variables, E[XA ] and E[XB ]? What is the ex-
pected utility of each winegrower in autarky, E[u(XA )] and E[u(XB )]?

3. Suppose they decide to pool risks together. Denote Y = XA +XB

2
. Do you think Y is
more or less risky than XA (or XB )?

4. What is the expected utility of each winegrower when they pool risks, E[u(Y )] as a
function of α?

5. Comment on the gains from pooling risks when α takes the value 1, 0 and -1.

Answer —

1. When α = 1, the table becomes

XA = 0 XA = −100
XB = 0 1/2 0
XB = −100 0 1/2

Either the weather is the same in the regions: either it is catastrophic for both,
either it is fine for both. In other words, the risks are 100% positively correlated.
When α = 0, the table becomes

6
 XA = 0 XA = −100
XB = 0 1/4 1/4
XB = −100 1/4 1/4

Each state has the same probability. In particular, the probability that the weather
is catastrophic in region A (resp. B) does not depend on the weather in region B
(resp. A). The weather is not correlated between the two regions, and so are the
risks.
When α = −1, the table becomes

XA = 0 XA = −100
XB = 0 0 1/2
XB = −100 1/2 0

The weather in A is catastrophic when it is fine in B, and vice versa. The risks are
100% negatively correlated.
According to these observations, it seems that α captures the degree of correlation
between the two events.

2. The probability that the weather is catastrophic in region A is the sum of two
probabilities: 1+α
4
+ 1−α
4
= 24 = 21 . The probability that the weather is fine is also
1/2. Note that the probabilities are the same for region B. The expectations are
therefore
1 1
E[XA ] = E[XB ] = × 0 + × (−100) = −50.
2 2
The expected utilities are

1 1
E[u(XA )] = E[u(XB )] = u(0) + u(−100).
2 2

3. The variable Y can take three values:

• Y = 0 with probability 1+α

4
(top-left case in the table)
• Y = −100 with probability 1+α
4
(bottom-right case in the table)
• Y = −100/2 = −50 with probability 1−α
4
+ 1−α
4
= 1−α
2
(the two remaining
cases in the table)

Notice that E[Y ] = E[XA ] = −50 (using the probabilities or the formula E XA +X ).
 B

2
Y appears to be less risky than XA because it yields the expectation value (-50)
more often, and the extreme values (0 and -100) less often.

4. The expected utility when pooling risks is

1+α 1−α 1+α

E[u(Y )] = u(0) + u(−50) + u(−100).
4 2 4

7
 5. When risks are 100% positively correlated,

1 1
E[u(Y )] = u(0) + u(−100) = E[u(XA )].
2 2

In that case, pooling risks yields the same expected utility as in autarky. Here,
pooling risks is not helpful. The idea of pooling risk is to compensate the loss of one
individual by the income of the other. However, in this scenario, when the weather
is catastrophic one winegrower, the other one systematically incurs a loss in this
scenario. When risks are not correlated,

1 1 1
E[u(Y )] = u(0) + u(−50) + u(−100).
4 2 4

Here, the two winegrowers simultaneously incurs a loss with probability 1/4. In the
other cases, they can help each others. When risks are 100% negatively correlated,

E[u(Y )] = u(−50).

This case is the best because there is no uncertainty. When one winegrower incurs a
loss, the income of the other compensate. Pooling risk in that case provides complete
insurance to both winegrowers.

Exercise 3 (Portfolio diversification). ♠ This exercise makes you prove the diversi-
fication result in an example.

Elliot has preferences that are captured by the utility function U (ω) = ω 2/3 . He has
to choose whether to invest his entire wealth in an asset and can choose between asset X
which pays £100 with a 70% chance and £0 with a 30% chance, or asset Y which pays
£100 with a probability of 40% and £0 with 60% probability. These assets have the same
price and are uncorrelated. We will round numbers at the second digit.

1. Without computing expected utilities, explain why Elliot strictly prefers asset X
over Y .

2. Show that Elliot is risk averse.

3. Compute the expected utility of each asset E[U (X)] and E[U (Y )].
Consider the returns Z(λ) of a portfolio that contains a share 1 − λ of X and a
share λ of Y .

8
 4. Show the four possible outcomes of the portfolio as a function of λ and the corre-
sponding probabilities.

5. What is the expected utility of the portfolio as a function of λ?

6. Compute the expected utility of the portfolio for the values λ = 0, 0.03, 0.06 and 1.
Conclude.

Answer —

1. X and Y have the same possible outcomes, 0 and 100, but X offers a higher chance
to get the best outcome and a lower chance to get the worse outcome. If Elliot is
rational, he would prefer X over Y whether he is risk-averse or not.

2. Take the utility function and differentiate it twice. We find u′ (ω) = 32 ω 2/3−1 =
1 −1/3
3
ω and u′′ (ω) = − 13 23 ω −1/3−1 = − 29 ω −4/3 . We consider positive values (ω ≥ 0)
so ω −4/3 is positive and the second derivative is negative. The utility function is
thus concave. Elliot is risk-averse because his utility function is concave.

3. We use the standard formulae for the expected utility:

E[U (X)] = 0.7U (100) + 0.3U (0) ≈ 15.08,

E[U (Y )] = 0.4U (100) + 0.6U (0) ≈ 8.62.

4. There are four possible cases for Z(λ), depending on the realisations of X and Y :

• with a 0.3 × 0.6 = 0.18 probability, both assets X and Y pay £0 and the
realization for the portfolio is Z(λ) = 0;
• with a 0.3 × 0.4 = 0.12 probability, the asset X pays £0, the asset Y pays
£100 and so the realization for the portfolio is Z(λ) = 100λ;
• with a 0.7 × 0.6 = 0.42 probability, the asset X pays £100 while the asset Y
pays £0 and the realization for the portfolio is Z(λ) = 100(1 − λ);
• with a 0.7 × 0.4 = 0.28 probability, both assets X and Y pay £100 and the
realization for the portfolio Z(λ) = 100(1 − λ) + 100λ = 100.

5. The portfolio yields the following expected utility to Elliot:

E[U (Z(λ))] = 0.18U (0) + 0.12U (100λ) + 0.42U (100(1 − λ)) + 0.28U (100)
= 0.12 × (100λ)2/3 + 0.42 × (100 − 100λ)2/3 + 0.28 × 1002/3

9
 6. First, notice that
E[U (Z(0))] = E[U (X)] ≈ 15.08

and
E[U (Z(1))] = E[U (Y )] ≈ 8.62.

We find E[U (Z(0.03))] ≈ 15.15 and E[U (Z(0.06))] ≈ 15.11. We can conclude that,
among the options λ = 0, 0.03, 0.06 and 1, the best one for Elliot is λ = 0.03 because
it yields the highest expected utility.

10