EC3000 Advanced Microeconomics 2023-24

Seminar 3: Problem Set

Answers

Module leader: Guillaume Wilemme - guillaume.wilemme@leicester.ac.uk

Seminar Tutor: Mert Gumren - mg534@leicester.ac.uk

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. Explain the idea behind collateralised debt obligations in less than 100
words.

Answer — —Possible answer—

Collateralised debt obligations (CDOs) are financial products that are built from a
pool of assets. Instead of dividing the pool of assets into products of similar risk, the
pool is split in different tranches of ordered risk. The returns of the pool are allocated
according to a decreasing order: the top tranches first, and the bottom tranches last. The

1
top tranches are, therefore, the least risky product, less risky than the assets in the pool.
The bottom tranches instead bear the highest risks. —

Question 2. ♠ Read the section ‘The Market for Insurance and The Poor’ (page 157) in
the article written by Banerjee and Duflo (2007).1 Comment on the following extract:
‘Poor households also bear most health-care risks (both expenditures and foregone earnings)
directly. For example, Gertler and Gruber (2002) find that in Indonesia a decline in the
health index of the head of the household is associated with a decline in non-medical
expenditures.’
Explain carefully, in less than 200 words, how we can conclude from the result of Gertler
and Gruber that households in Indonesia are not fully insured.

Answer — —Note. You are asked in this question to interpret the second part of the
quote. You must understand and interpret the connection between: i) ‘a decline in the
health index of the head of the household’, ii) ‘a decline in non-medical expenditures.’—
—Possible answer—
Gertler and Gruber seem to observe a relationship between two economic outcomes: a
measure of health in the household and consumption (non-medical goods excluded). We
can suppose these authors want to test whether household consumption is affected by a
change in health outcomes. Empirically, they find a relationship: Households consume
less, once we abstract from medical expenditures, when the head of household has health
issues.
This relationship is not compatible with full insurance. If the households were fully
insured against health risks, then consumption would not be affected. A full insurance
coverage would provide additional income to cover for the medical spendings, and so
households would consume as normal. The response of household consumption to the
health shocks is an empirical evidence of partial or no insurance. —

Question 3. ♠ Dorothy is thinking about creating a business. Her returns if she creates
a business are X = 1000 with probability 0.5, or X = −100 with probability 0.5. Alter-
natively, Dorothy can be hired and get Y = 200 for sure. Her certainty equivalent of X
is 100. Is it optimal for her to create a business or not?
A bank is willing to help her. With a loan from the bank, she would not loose income
if the project fails but she would get less income if it succeeds. Her returns with the loan
are Z = 800 with probability 0.5 and Z = 0 with probability 0.5. Her certainty equivalent
of Z is 300. What is now the best option for her?
Use this example to illustrate the social benefits of insurance for production decisions.
Your answer must contain less than 300 words.
1
Banerjee and Duflo. ‘The Economic Lives of the Poor,’ Journal of Economic Perspectives, 2007.

2
Answer — —Possible answer—
Between X and Y , Dorothy optimally chooses not to create a business because the
certainty equivalent of X is lower than Y = 200. For the same reasons, Dorothy optimally
chooses to create a business with a loan because the certainty equivalent of Z is higher
than Y = 200.
In this example, the loan provides some insurance to Dorothy on the one hand. There is
an income transfer from the good state (success) to the bad state (failure). The difference
between the two payoffs is lower with the loan, hence uncertainty is reduced. This is a
reason to prefer Z over X. On the other hand, the bank is charging the good state for
the risk so that the expected returns are reduced, E[X] = 450 > E[Z] = 400. This is a
reason to prefer X over Z. In that situation, we need more information about Dorothy’
s preferences. We have this information with the certainty equivalents. Comparing the
certainty equivalents, we can conclude that Dorothy prefers Z to X.
Thanks to this loan, Dorothy now makes the choice that maximised expected total
surplus, which is creating a business. By creating a business, Dorothy generates the asset
X that she shares with the bank: she gets Z and the bank gets X − Z. The following
table describes the three situations:

No business Business without loan Business with loan

Dorothy gets Y X Z
Bank gets 0 0 X −Z
Total surplus Y X X
Expected total surplus E[Y ] = 200 E[X] = 450 E[X] = 450
We know, from the theory of the producer under uncertainty, that a risk-averse producer
can make socially inefficient decisions when they are not insured. With an access to
insurance (even partial insurance as in the example), this problem is mitigated as the
risk-averse producer tends to behave like a risk-neutral producer, maximising expected
profits. —

Exercises
Exercise 1 (Demand for insurance). This exercise introduces you to the optimal de-
cision regarding insurance. Depending on the price of insurance, an individual decides the
type of coverage (partial or full).

Jennifer’s preferences over wealth can be described by the utility function

√
U (x) = x.

3
Suppose that her entire wealth is provided by a property worth £1million. However this
property has a 10% chance of being completely destroyed by a fire. Denote X the random
value of the property.

1. Show that Jennifer is risk averse.

2. Compute the utility of the expected value and the expected utility of the property
for Jennifer. Why do they differ?

3. Suppose that a fire insurance contract is available that offers full coverage for Jen-
nifer’s property in case of accident. The insurance product costs C in total. Find the
expected utility of Jennifer when she opts for this insurance product, as a function
of C.

4. What is the maximum amount C̄ Jennifer would pay for this full insurance?

5. Suppose now that Jennifer can choose the level of coverage for her property. Care-
fully explaining your analysis, show that, if the insurance price is 15p per pound,
Jennifer would buy approximately £436,000 of coverage (£435,897). Explain why,
if the price were of 10p per pound, she would insure her property for its full value.

Answer —

1. You must compute the derivative, and then the second derivative of the utility
function. You must show that this second derivative is negative to prove that the
utility function is concave. If the utility fnction is concave, then Jennifer is risk-
averse. We have computed this second derivative in previous exercises.

2. The expected value is the weighted average of the payoffs in each of the possible
states of the world weighted for the probability of each state. The utility of expected
value is the utility of this expected value and in the present example is the following:
√
U (E[X]) = 0.9 × 1000000 + 0.1 × 0 ≈ 948.68

The expected utility E[U (X)] is the weighted average of the utilities in each of the
possible states of the world weighted for the probability of each state. It is the
standard tool used to analyse decision making under uncertainty. In the present
case the expected utility can be computed as follows:
√ √
E[U (X)] = 0.9 1000000 + 0.1 0 = 900

They differ, with U (E[X]) > E[U (X)], because Jennifer is risk-averse. Jennifer
would prefer trading the uncertain asset X, which yields expected utility E[U (X)],

4
 for a riskless asset that provides the same expected returns, say X0 = E[X], which
yields E[U (X0 )] = U (X0 ) = U (E[X]). This is the complete insurance theorem.

3. Jennifer’s wealth under full insurance will be denoted Y . We are looking for a
definition of E[U (Y )]. If there is no fire, Jennifer receives the value of the house,
1 million, but she pays C to the insurer. If there is a fire, the house is no more
but Jennifer receives 1 million from the insurance. In both cases, she receives
1000000 − C. The expected utility if she opts for the insurance is

E[U (Y )] = 0.9U (1000000 − C) + 0.1U (1000000 − C) = U (1000000 − C)

√
= 1000000 − C

4. The maximum price Jennifer is willing to pay is such that she is indifferent between
insurance and no-insurance, between Y and X. If the cost is higher than the thresh-
old cost, she would prefer not being insured. We solve E[U (Y )] = E[U (X)] to find
C̄:
p
1000000 − C̄ = 900,
C̄ = 1000000 − 810000 = 190000.

The maximum price is therefore £190,000.

5. When Jennifer can choose the level of coverage q her asset can be described by a
random variable Z (that implicitly depends on q). Her expected utility E[U (Z)] of
buying £q of coverage when the price is 15p per pound is:
1 1
E[U (Z)] = 0.9 (1000000 − 0.15q) 2 + 0.1 (q − 0.15q) 2
1 1
= 0.9 (1000000 − 0.15q) 2 + 0.1 (0.85q) 2

In order to find the level of q that maximises the utility we compute the first-order
condition (FOC) with respect to q:

∂E[U (Z)]
=0
∂q
1 1 1 1
0.9 × (−0.15) × (1000000 − 0.15q) 2 −1 + 0.1 × 0.85 × (0.85q) 2 −1 = 0
2 2
− 21 − 12
− 0.0675 (1000000 − 0.15q) + 0.0425 (0.85q) = 0
1 1
675 (1000000 − 0.15q)− 2 = 425 (0.85q)− 2

We could go on and solve for q (try to do it!), but here we are given the an-
1
swer. With q = 436000, we can check that 675 (1000000 − 0.15q)− 2 ≈ 0.698 and

5
 1
425 (0.85q)− 2 ≈ 0.698. This shows that q = 436000 solves the first-order condition.
At 15p per pound, Jennifer would buy 436,000 units of coverage (for a total cost
close to £65,000).
If the price is equal to 10p per pound, then the expected utility is
1 1
E[U (Z)] = 0.9 (1000000 − 0.1q) 2 + 0.1 (0.9q) 2 ,

and the FOC becomes

∂E[U (Z)] 1 1 1 1
= 0.9 × (−0.1) × (1000000 − 0.1q)− 2 + 0.1 × 0.9 × (0.9q)− 2 = 0
∂q 2 2
1 1
(1000000 − 0.1q)− 2 = (0.9q)− 2
1000000 − 0.1q = 0.9q
q = 1000000

For that price, Jennifer buys full insurance and the 1 million value of the house is
covered, for a total cost of 0.1 × 1000000 = £100, 000. The price 10p is actuarially
fair because the probability of loss is 10%. It is not surprising that Jennifer buys
full insurance in that case.

Exercise 2 (Production under uncertainty). ♠ Wallace owns a firm that sells frozen
fries. He must choose the quantity of potatoes x to buy to produce a quantity y =
√
f (x) = 2 x of fries. Wallace receives the profits π as income, which generates utility
U (π) = 100 ln(π/90). The price of potatoes is normalised to 1. The price of french fries,
however, is uncertain. With probability q = 0.25, the price will be p1 = 7 (state 1). With
probability 1 − q, the price will be p2 = 11 (state 2).

1. Show that Wallace is risk-averse. Write Wallace’s expected profits and Wallace’s
expected utility as functions of x.

2. What is the optimal quantity of potatoes Wallace would buy if he were risk-neutral,
or equivalently, if he were maximising expected profits? Show that this quantity is
xA = 100.

3. Consider another option xB = 80. Which option does Wallace prefer between xA
and xB under the utility function U ? Explain.

From now suppose Wallace can sell french fries on a futures market. The price on
this market is pF .

6
 4. What is the price pF when the futures market is competitive?

5. Wallace decides to sell all his french fries on the futures market. Does he prefer to
buy a quantity xA or xB ? Conclude.

Answer —

1. To check that Wallace is risk-averse, we differentiate U twice:

1 π
U 0 (π) = 100 × × ln0 .
90 90

The first derivative of ln(x) is 1/x and so

1 1 1 90 100
U 0 (π) = 100 × × π = 100 × × = .
90 90
90 π π

The second derivative is

100
U 00 (π) = − .
π2

The second derivative is negative when π is positive, thus the utility function is
concave and Wallace is risk-averse.

Expected profits are

q[p1 f (x) − x] + (1 − q)[p2 f (x) − x] = (qp1 + (1 − q)p2 )f (x) − x

= (0.25 × 7 + 0.75 × 11) f (x) − x
√
= 20 x − x

Expected utility is NOT the utility of expected profits. Expected utility is expressed
as
 √   √ 
14 x − x 22 x − x
qU [p1 f (x) − x] + (1 − q)U [p2 f (x) − x] = 25 ln + 75 ln .
90 90

2. If Wallace were risk-neutral, he would maximise expected profits (because his utility
function would be linear). We write the first-order condition when differentiating
expected profits:

1
20 x−1/2 − 1 = 0,
2
10 = x1/2 ,
x = 102 = 100.

7
 Wallace would choose xA = 100 as quantity of inputs.

3. We compute the expected utility for xA and xB . Using the formula from the first
question, we find expected utility at 1.3 with xA , and at 2.3 with xB .
Thisproves that, owning to risk aversion, Wallace buys less inputs and produces less
than what would maximise expected profits.

4. The price pF is obtained from a no-arbitrage condition. Suppose a financier is

selling z units on the futures market, thus getting a revenue pF z. To provide z
units to their buyer, they must buy z units on the spot market. With probability
q, the financier must spend p1 z. With probability 1 − q, they must spend p2 z. This
financial operation should yield no profits:

7 33
pF z = qp1 z + (1 − q)p2 z = z + z = 10z.
4 4

The price on the futures market is therefore pF = 10.

5. When Wallace sells everything on the futures market, he does not face any uncer-
tainty. His expected utility is simply the utility:
 √ 
20 x − x
U (pF f (x) − x) = 100 ln .
90

We find the utility is 10.5 with xA and 9.4 with xB . We can conclude that Wallace
chooses xA , which is the profit-maximising quantity of potatoes.

Here, the first welfare theorem applies because there is a market to insure Wallace
(completeness) and it is competitive. Wallace buys full insurance when selling on
the futures market.