Semester II, 2022-2023 Jo Seung-gyu

NUS Business School

BMA5001 MANAGERIAL ECONOMICS

Lecture Notes 4

Economics of Uncertainty/Risk

1
Relevant Textbook Chapter
• Ch5 (Sections 5.1 – 5.3)

Readings:
• ‘How economics views uncertainty’
• ‘Risk Pools for the medically uninsured’
• ‘Marital Status/Gender and Risk Aversion’
• ‘Women Are Not Risk Averse – Society Teaches Them To Be
That Way, Study Shows.

2
Outline

1. Introduction
2. Uncertainty/Risk and Expected Value
3. Expected Utility
4. Preference toward Risk and Reducing Risk:
 An Insurance Problem
 Risk-sharing and Risk-pooling
5. Appendix

3
 Introduction and Learning Objectives

 We make decisions daily. Most require little effort and the choice is
obvious.
 What should you eat for lunch?

 Occasionally, however, we encounter decisions demanding more thought

because the situations involve a variety of uncertainty/risk and big stakes.
 Financial investment of your $100K
 Risking $40 million to acquire another firm

 How should we take these uncertainties/risks into account in our

decisions then? In this lecture notes, we will formalize how economics
views our decisions under uncertainty.

4
 How does economics view uncertainty and risk?
While closely related and occasionally (and erroneously) used
interchangeably, ‘uncertainty’ and ‘risk’ have important differences

 uncertainty is present when the likelihood

of future events is indefinite or incalculable
 risk is present when future events occur
with measurable probability

Examples:
o A die rolling or a roulette game — we have risk, not uncertainty.
o You could step in a mud puddle or you could cause a fire burning your
house – we do have uncertainty, not risk.

Here in our classroom economics, we only deal with risk.

5
 OUTLINE

1. Introduction

2. Uncertainty/Risk and Expected Value

3. Expected Utility
4. Preference toward Risk and Reducing Risk:
 An Insurance Problem
 Risk-sharing and Risk-pooling
5. Self Exercise

6
 Motivation: How would you rank the following options?

Lottery A: receive $10 with probability of 0.5 or pay $10 with probability of 0.5
Lottery B: receive $12 with probability of 0.5 or pay $10 with probability of 0.5
Lottery C: receive $8 with probability of 0.5 or pay $8 with probability of 0.5

 A vs B?
 A vs C?
 How about B vs C?

Under uncertain situation, we face a trade-off between expected Value vs

variability:
─ prefer a higher expected value and
─ generally prefer a lower variability ( variation or standard deviation),
if not always
─ Your detailed attitude or preference toward risk matters and we need
to formalize it. That is, the expected value approach does not explain
our behavior under uncertainty/risk too well.
7
 First Perception on Risk/Uncertainty: Expected Value (EV)?

Let us play a game:

 Roll a die. You win $6 if even number and lose $2 if odd.

 Ask yourself:
o Would you pay $1.50 to play the game? How about $2?
Anyone who is keen to pay $3 to play?

o What was the first thing you wanted to know about?

The Expected Value (i.e. expected amount to win) of the game :

EV = $6x(1/2) + (-$2)x(1/2) = $2.

8
 Formalization: Uncertainty/Risk and Expected Value (EV)

● A risky situation naturally involves

o Possible states:
i = 1, 2, … , n
o Quantified outcomes at each state (e.g. $):
X1, X2, … , Xn
o The probability or likelihood that each outcome will occur:
P1, P2, … , Pn

● Then the expected value (EV) of the outcomes is defined by:

EV = ∑𝑛𝑛𝑖𝑖=1 𝑃𝑃𝑖𝑖 𝑋𝑋𝑖𝑖

9
 How Valid is Expected Value concept?

‘St. Petersburg Paradox’

• I toss a coin until head appears.

• When the first head appears at the nth toss, you are paid $2n.
• Once the coin lands head, the game is over.

I am asking you to pay $X to join this game. How much would you pay?

$100? $50? $10?

10
 Expected Value of the St. Petersburg’s Coin Flip

n 1 2 3 …… n …...

Winning $ $2 $22 $23 …… $2n ……

prob. (1/2) (1/2)2 (1/2)3 …… (1/2)n ……

EV = $2(1/2) + $22(1/2)2 + $23(1/2)3 + … + $2n(1/2)n + …

=∞
 Finding:
Expected Value concept is flawed; it does not explain our actual behavior
under uncertainty too well.

 Challenge:
What is missing here? How can we generalize one’s decision under
uncertainty?
11
.
 OUTLINE

1. Introduction
2. Uncertainty/Risk and Expected Value
3. Expected Utility
4. Preference toward Risk and Reducing Risk:
 An Insurance Problem
 Risk-sharing and Risk-pooling
5. Appendix

12
 What Is Missing in EV Approach?

Experiment: Which prospect would you prefer?

Prospect A: Sure $100  EV = $100

Prospect B: ($300, -$100) with (50%, 50%)  EV = $100

 Some would prefer A while others would prefer B:

 Our decision is not solely based on the expected value. It matters how
we view uncertainty/risk.
 We need a classifying scheme that helps us factor a risk into decisions.

13
 Alternative Scheme: Expected Utility (EU) Approach

 Possible states: i = 1, 2, … , n
 Quantified outcomes at each state (e.g. $): X1, X2, … , Xn
 The probability or likelihood that each outcome will occur: P1, P2, … , Pn
 Suppose the person’s personal utility from Xi is given by U(Xi).
Given the above, the expected utility (EU) is defined as the weighted
average of the utilities U(Xi) that the individual derives from each
consequences Xi, with the probability Pi of that consequence occurring
used as the weight:

EU = ∑𝒏𝒏𝒊𝒊=𝟏𝟏 𝑷𝑷𝒊𝒊 𝑼𝑼(𝑿𝑿𝒊𝒊 ) EV = ∑𝒏𝒏𝒊𝒊=𝟏𝟏 𝑷𝑷𝒊𝒊 𝑿𝑿𝒊𝒊

Examples:
• Risky $2 vs Sure $2 – which one would you prefer?
• The satisfaction we would get from $3 million isn’t necessarily three times of the
14
satisfaction we’d get from $1 million.
 OUTLINE

1. Introduction
2. Uncertainty/Risk and Expected Value
3. Expected Utility
4. Preference toward Risk and
Reducing Risk
 An Insurance Problem
 Risk-sharing and Risk-pooling
5. Self Exercise

15
 An Illustration: Consider the following situation:

(a) Your initial wealth: $100

(b) You are offered a gamble: (Win $50, Lose $50) with probability of (1/2, 1/2)

● Expected Value Approach:

(i) Expected Wealth (No Gamble) = $100
(ii) Expected Wealth (Gamble) = 0.5 ($100+$50) + 0.5 ($100-$50) = $100
But a guaranteed wealth of $100 as in (i) and an uncertain wealth of $100 as
in (ii) are not viewed the same way to you. Thus, expected value does not
explain our various decisions: Your preference toward risk matters.

● Expected Utility Approach:

Let us suppose your utility function over wealth is U($).
(i) Expected Utility (No Gamble) = U($100) ---------------------------------- (a)
(ii) Expected Utility (Gamble) = 0.5U($100+$50) + 0.5U($100-$50) ---- (b)

(a) > (b)  you are ‘risk averse’ (you prefer certainty)
(a) < (b)  you are ‘risk loving’ (you prefer uncertainty/risk)
(a) = (b)  you are ‘risk neutral’ (you are indifferent)
16
 Three Attitudes Towards Risks

Different people have different views, perspectives, and preferences about risk.
Some people prefer to avoid risk (risk averse), others enjoy engaging in risk (risk
loving), and some others are indifferent (risk neutral).
.
 A person is said to be risk averse if she prefers a certain level of wealth to a
risky ( or uncertain) wealth with the same expected value.
─ Such preference can be characterized by a concave utility over monetary
outcome.
─ Examples: buying insurance, avoid gambling

 A person is said to be risk loving if she prefers a risky ( or uncertain) wealth over
a certain level of wealth when the expected value remains the same.
─ Such preference can characterized by a convex utility over monetary
outcome.
─ Examples: gambling, criminal activity, living without insurance

 A person is said to be risk neutral if she is indifferent between a certain level of

wealth and an uncertain one with the same expected value.
─ Such preference can be characterized by a linear utility over monetary
outcome.
17
 The fact is:

‘Most of us are risk averse’ in general, if not all.

 If you are risk averse, you would prefer a certain (risk-free) wealth $X to
an uncertain (risky) expected wealth of $X.

 Thus, we are interested in:

• How to represent such a risk-averse (and risk-loving/neutral as well)

preferences?
• How to quantify (or visualize) the maximum amount one is willing to
pay to avoid the risks? – Risk Premium

● Let’s investigate through an insurance problem below.

18
  Example: Fire Insurance

 Your current home value (X) is $225.

─ $200 in house and
─ $25 in land value

 You have got a risk factor, though:

Each year, there is a probability of p = 20% that the house will burn
to the ground in a fire and a probability of 1 – P = 80% that it will not.
(If fire, you lose the whole value of the house structure $200
– that is, you are left with $25 only from the land.)

19
 Your current home value (X) is $225.
 Viewed from Expected Value (EV)?. • $200 in house and
• $25 in land value
• P=20% to have a fire
The expected loss from fire is = $200x0.2 = $40

● (If no insurance, i.e. self-insured) The expected value of your home is

EV = $225 – $40 = $185 or uncertain $185 --- (b)
EV = 0.2x($225-$200)+ 0.8x$225 = $185

● (If buy a full-coverage insurance at the ‘actuarially fair premium’ of $40)

Then, the expected value of your home is
guaranteed $185 --- (a)
EV = $225 – $40 = $185 or
EV = 0.2x ($225 – $40 – $200 + $200) + 0.8x($225 – $40) = $185

Questions:
 Would you buy the insurance at $40? How about at $41? $45?
 Up to how much would you pay for the insurance?
 Note that the Expected Value (EV) approach does not help us here .

20
 Viewed from Expected Utility (EU) Approach: Now, let’s introduce U scheme.

 If you are risk averse: At the insurance premium of $40

EU(with insurance) > EU*(no insurance)

= U(guaranteed $185) = 0.2U($25) + 0.8U($225)

 It implies that you are willing to pay more than $40 to avoid the risk.
─ What is the maximum amount $X you would pay for the insurance? The
condition is as follows:

U($225 – $X) = 0.2U($25) + 0.8U($225)

─ If you are risk averse, $X would be greaer than $40. The ‘extra $’ you are
willing to pay above $40 is called a risk premium.

 Thus, your risk premium depends on what your U looks like.

 Next, we represent preferences towards risk first, and then find the risk
premium.
21
 Various Preferences Toward Risk

 If no insurance: Your expected utility is EU*(no insurance) = 0.2U(25) + 0.8U(225).

 If insured at the insurance premium of $40: Your guaranteed utility is U($185).

Utility (U)
U($185)?
If insured at the insurance premium of $40
U($225) ●
●
If no insurance ● ● If U($185) here, then risk-averse
EU*(no insurance)
= 0.2U($25) + 0.8U($225) ● If U($185) here, then risk-neutral

If U($185) here, then risk-loving

U($25) ● ●

A = $40
(=expected loss = fair premium)

● ● ● Wealth ($X)
$25 $185 $225 22
(if fire: 20%) (EV=0.2x25+0.8x225) (if no fire: 80%)
 Visualization of Three Risk Attitudes through Utility Curves:
(Risk Averse, Risk Loving, Risk Neutral Preferences)

Utility (U)
Risk-loving Risk-neutral

U(225) Risk-averse
Examples
EU*(no insurance)
e) Functional
Risk Attitude
Example
Risk Averse U = X0.5
Expected Loss
U(25) Risk Loving U = X2
= $40
Risk Neutral U = aX (a>0)

Wealth (X)
25 $185 225
(if fire: 20%) (if no fire: 80%)

23
Recall:
A risk-averse person would be willing to pay a risk premium ( = extra $ above
the expected loss) to be risk-free.

How big is your risk premium then?

─ That is, to be risk-free, up to how much would you be
willing to pay on top of the expected loss of $40?)

To find it, you first want to solve for the maximum price ($X or Pmax) you
want to pay for the insurance:

U($225 – $X) = 0.2U($25) + 0.8U($225)

Let’s solve for it below by visualizing.

24
  Risk Premium & Indifference Premium

U($) U($225 – Pmax)  Pmax: The maximum insurance premium you

= EU*(self-insured or no insurance)
are willing to pay
U(225) U = W 0.5 (called the ‘indifference premium’)
U(185)  A = fair premium (=expected loss) = $40
EU*(no insurance)
EU*(no
EU (no insurance)
insurance) =
● ●  B = risk premium = Pmax – A (i.e. Pmax – 40)
=
U(insured at Pmax)
0.2U($25) + 0.8U($225)

U(25) B A Pmax must be such that:

U(225 – Pmax) = 0.2U(25) + 0.8U(225)

Pmax

● Wealth W ($) (Example) If U = W 0.5 :

$25 $185 $225
(if fire: 20%) (if no fire: 80%)
(225 – Pmax)0.5 = 0.2(25)0.5 + 0.8(225)0.5
$(225 – Pmax)
Then, Pmax = $56, A = $40, B = $16
(called ‘Certainty Equivalent’)

(See the Appendix at the end for the details.)

25
 Reducing Risk

 Insurance
o As we have seen in the previous slides.

 Diversification
o Invest in mutual fund, than individual financial asset

 Risk-pooling/Risk-sharing
o You are risk averse. How about the risk to the insurance companies?
̶ ‘Law of large numbers’

(From the readings)

• ‘Risk pools for the medically uninsurable aid those turned down
for health insurance’
• ‘Marital Status, Gender and Risk Aversion’
26
 Reality Check:

(from the readings)

‘Marital Status and Gender Differences in Risk Aversion’
‘Women Are Not Risk Averse – Society Teaches Them To Be That way’

 Marital Status?

 Gender Differences?

 Testosterone?

27
APPENDIX

(How to Calculate Risk Premium from the Fire Insurance Example)

28
 Appendix: Calculating Risk Premium

Suppose your preference toward risk is represented by a utility

function of U(W) = W0.5 . Consider the previous fire insurance
example from the lecture notes.

(1) How big would your risk premium be?

(1) What is the maximum price – i.e. the indifference premium

– you are willing to pay for the full-coverage insurance?

29
Solution:

The situation can be viewed as below:

Utility

U = W0.5
U(225) = 15
EU (no insurance)
no insurance)
= 13 = 13

Indifference Premium (PM)

Indifference Premium (P
= 56
U(25) = 5

Risk premium = 16

Expected Loss = 40

Wealth ($)
30
25 169 185 225
The expected utility from no insurance is

EU (no insurance) = pUIf Fire + (1 – p)UIf No Fire

0.5 0.5
= 0.2x(25) + 0.8x(225) = 0.2x5 + 0.8x15= 1 + 12 = 13.
Thus, the maximum insurance premium (Pmax) you are willing to pay would be such
that you are indifferent between having your home insured at the price of Pmax and no
insurance. That is,

[ U(insurance at Pmax) = ]
0.2xU(225 – Pmax – 200 + 200) + 0.8U(225-Pmax) = U(225-Pmax)
= 13 [ = EU (no insurance)]

0.5
Thus, (225 – Pmax) = 13 or Pmax = 56

And the risk premium is $16 ( = 56-40)

31
Wrap Up:

Indifference Premium:

The indifference premium = 56 > 40 = expected loss. You are

willing to pay more than the expected loss since you are risk
averse as supported by the concave utility function.

Risk premium:

You are willing to spend $16 ( = 56-40) on top of the expected

loss to hedge yourself against the risk (i.e. uncertainty) involved
in the situation.

32
End of Lecture Notes 4

33