Topic 2A: The Standard Model: Expected Utility

Econ 4660: Behavioral Economics

Professor Ted ODonoghue
Spring 2014

()

Expected Utility

Spring 2014

1 / 28

Topic 2: Choice Under Uncertainty

Topic 2A: The Standard Model: Expected Utility

()

Expected Utility

Spring 2014

2 / 28

Dening the Environment

Denition: A lottery (or gamble or risky prospect) is a set of possible
outcomes and a probability of each outcome occurring.

Example A: If you go to a roulette table and place $10 on BLACK, you

have just purchased a lottery. With probability 18/38 you will receive $20,
and with probability 20/38 you will receive $0.
Example B: When you buy a new car, you are buying a lottery. For
instance, it might be that with probability 1/2 it will be a car that you
love (high value), with probability 2/5 it will be a car that only adequately
serves your needs (low value), and with probability 1/10 it will be a
lemon (worthless).
Example C: More abstractly, you might get x1 with probability p1 , x2 with
probability p2 , and x3 with probability p3 (where p1 + p2 + p3 = 1).
()

Expected Utility

Spring 2014

3 / 28

Dening the Environment

Ways to write lotteries:
As a probability tree.
As a vector of outcome-probability pairs:
Example A: Payo = ( $20, 18/38 ; $0, 20/38 ).
Example B: Car Value = ( high, 1/2 ; low, 2/5 ; worthless, 1/10).
As a type of equation:
Example A:
Payo =

$20
$0

with prob 18/38

with prob 20/38

Example B:

()

8
< high value
low value
Car Value =
:
worthless
Expected Utility

with prob 1/2

with prob 2/5
with prob 1/10
Spring 2014

4 / 28

Dening the Environment

In the realm of choice under uncertainty, lotteries are the objects of choice.
Next step: Develop models of behavior.
Suppose you face a choice between two lotteries. How do you decide?

()

Expected Utility

Spring 2014

5 / 28

Model #1: Expected Value Theory (EV)

A possible model of behavior: People choose the option with the largest
expected value.
Denition: The expected value of a lottery x = ( x1 , p1 ; ... ; xn , pn ) is
n

EV (x)

pi xi .

i =1

()

Expected Utility

Spring 2014

6 / 28

A Problem for EV
St. Petersburg Paradox
Consider the following bet: Im going to ip a coin, and Im going to keep
on ipping it until I ip a HEADS. Then youll be paid as a function of
how many times we ip. Specically:
If I immediately ip a HEADS, you get $2.
If I ip one TAILS and then a HEADS, you get $4.
If I ip two TAILS and then a HEADS, you get $8.
If I ip three TAILS and then a HEADS, you get $16.
And so forth....
How much are you willing to pay for this bet?
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
The paradox:
The EV of this bet is , but people are unwilling to pay much for it.
Hence, EV is not a good description of peoples choices.
()

Expected Utility

Spring 2014

7 / 28

Whats Missing from EV? Risk Aversion

Denition: A person is risk-averse if for any (risky) lottery x she prefers

to have EV (x) for sure instead of lottery x.
Denition: A person is risk-neutral if for any (risky) lottery x she is
indierent between EV (x) for sure vs. lottery x.
Denition: A person is risk-loving if for any (risky) lottery x she prefers to
have lottery x instead of EV (x) for sure.

=) Evidence suggests people are risk-averse.

()

Expected Utility

Spring 2014

8 / 28

Model #2: Expected Utility Theory (EU)

A second model of behavior: People choose the option with the largest
expected utility.
Denition: The expected utility of a lottery x = (x1 , p1 ; ...; xn , pn ) is
n

EU (x)

pi u (xi ).

i =1

()

Expected Utility

Spring 2014

9 / 28

Basic Implications of EU

Could an expected-utility maximizer (EU maximizer) choose

(100, 12 ; 0, 12 ) over (200, 12 ; 0, 12 )?
.

Point: If we put no restrictions on the utility function u (x ), then

expected-utility theory can explain any individual choice.

()

Expected Utility

Spring 2014

10 / 28

Basic Implications of EU
Three Comments:
(1) We say that expected-utility theory can explain a choice if there
exists a utility function u (x ) such that an EU maximizer with utility
function u (x ) would make that choice.
(2) Even when we put no restrictions on the utility function u (x ),
although an individual choice cannot violate EU, combinations of choices
can violate EU.
E.g.,
AND

(100, 21 ; 0, 12 )

(200, 21 ; 0, 12 )

(100, 13 ; 500, 23 )

(200, 31 ; 500, 23 )

(3) When we put restrictions on the utility function u (x ), then even an

individual choice could violate EU.
()

Expected Utility

Spring 2014

11 / 28

Basic Implications of EU
A natural restriction: We virtually always assume that more is better
that is, u is increasing in x.
Denition: Lottery x dominates lottery y (or equivalently, lottery y is
dominated by lottery x) if for any amount z the probability of getting
at least z in lottery x is larger than the probability of getting at least z in
lottery y.
Examples:
(200, 12 ; 0, 12 )

vs.

(100, 21 ; 0, 21 )

(200, 12 ; 0, 12 )

vs.

(150, 12 ; 50, 12 )

(200, 13 ; 150, 31 , 75, 13 )

vs.

(150, 12 ; 75, 12 )

Result: Under expected utility with more is better, a person will never
choose a dominated lottery.
()

Expected Utility

Spring 2014

12 / 28

Basic Implications of EU
To incorporate risk-aversion (or risk-seeking), we must assume more about
the utility function.
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Result: Under expected-utility theory, a person is risk-averse if and only if
her utility function is concave (u 00 (x ) < 0).
Result: Under expected-utility theory, a person is risk-neutral if and only if
her utility function is linear (u 00 (x ) = 0).
Result: Under expected-utility theory, a person is risk-loving if and only if
her utility function is convex (u 00 (x ) > 0).
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Note: In this class, whenever we say that someone is a risk-averse (or
risk-neutral or risk-loving) expected-utility maximizer, well implicitly
assume more is better as well.
()

Expected Utility

Spring 2014

13 / 28

5 Features/Issues of Expected Utility

(1) The utility function is unique up to a positive a ne transformation.
Denition: The function u (x ) is a positive a ne transformation of the
function u (x ) if u (x ) = a u (x ) + b for some a > 0 and any b.
Result: If u (x ) is a positive a ne transformation of u (x ), then they
reect the same preferences that is, for any pair of lotteries x and y,
EU (x) > EU (y) if and only if E U (x) > E U (y).

(2) Can u (x ) be negative? Can EU (x) be negative?

Answer: sure.
Point: The level of utility doesnt matter, only the comparisons.
()

Expected Utility

Spring 2014

14 / 28

5 Features/Issues of Expected Utility

(3) Integration: EU operates on nal wealth states.
Proper way to use expected-utility theory:
Let w be a persons wealth.
Let x = (x1 , p1 ; ...; xn , pn ) be a lottery (or gamble or risky prospect).
Note: x yields income xi with probability pi .
EU theory says evaluate prospect x according to:
n

EU (x; w )

pi u (w + xi ).

i =1

()

Expected Utility

Spring 2014

15 / 28

5 Features/Issues of Expected Utility

(4) Closely related: EU operates on reduced lotteries.
Example: If you face the possibility of both lottery
x = (100, 1/2; 100, 1/2) and y = (90, 1/2; 90, 1/2), you must rst
combine these into a reduced lottery before applying EU.
If they are independent, the reduced lottery is

(190, 1/4; 10, 1/4; 10, 1/4; 190, 1/4).

If they have perfect positive correlation, the reduced lottery is

(190, 1/2; 190, 1/2).

If they have perfect negative correlation, the reduced lottery is

(10, 1/2; 10, 1/2).

()

Expected Utility

Spring 2014

16 / 28

5 Features/Issues of Expected Utility

(5) We can measure the degree of risk aversion.
Goal: Wed like to make statements of the form, If person A is more
risk-averse than person B, then A is less likely to accept a particular
gamble.
Question: How to formalize more risk averse?
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Denition: The coe cient of risk aversion at wealth w is
r (w ) =

u 00 (w )
.
u 0 (w )

^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Suppose wA is person As wealth and wB is person Bs wealth. If
r (wA ) > r (wB ) then person A is less likely than person B to accept a
risky gamble.
()

Expected Utility

Spring 2014

17 / 28

5 Features/Issues of Expected Utility

Note: A persons risk aversion depends on both her utility function and
her wealth.
u 00 (w )
r (w ) = 0
.
u (w )
How is ones willingness to accept a gamble likely to depend on my wealth?
Economists often assume risk aversion decreases with wealth:

" w =)# r (w ).

()

Expected Utility

Spring 2014

18 / 28

An Important Functional Form

Denition: The CRRA utility function is
u (z ) =

(z )1
1

for 2 [0, ).

Some features:
For = 0, equivalent to u (x ) = x, or risk neutrality.
For = 1, equivalent to u (x ) = ln x.
Note that r (w ) = /w , hence " =)" r (w ) and " w =) # r (w ).
Person feels the same about the following gambles:
gamble ($11, 1/2;

$10, 1/2) from wealth $1, 000

gamble ($110, 1/2;

$100, 1/2) from wealth $10, 000

gamble ($1100, 1/2;

$1000, 1/2) from wealth $100, 000

Empirical studies suggest in the 1-5 range.

()

Expected Utility

Spring 2014

19 / 28

Application: Demand for Insurance

Suppose you have wealth $1000.
But there is a 10% chance that you will suer a loss of $250.
Hence, you face the lottery ( 1000, .9 ; 750, .1 ).
Insurance agent: I will fully insure you if you pay me p.
Implication: If buy insurance, you face the lottery ( 1000

p , 1 ).

What is your willingness to pay for full insurance?

Answer: Its the p such that

( 1000

p ,1 )

( 1000, .9 ; 750, .1 )

or
u (1000
()

p ) = (.9) u (1000) + (.1) u (750).

Expected Utility

Spring 2014

20 / 28

Application: Demand for Insurance

Suppose instead:
To insure against the entire $250 loss, it costs p.
To insure against proportion of the $250 loss, it costs p.
What is the optimal ?

Because is continuous, use calculus:

Step 1: Derive nal lottery as a function of .
Step 2: Calculate expected utility as a function of .
Step 3: Use calculus to solve for the optimal .

()

Expected Utility

Spring 2014

21 / 28

Application: Demand for Insurance

As a function of the you choose, nal lottery is:

x() = ( 1000

p , .9 ; 750 + (250

p ) , .1 ).

As a function of the you choose, expected utility is:

EU (x()) = (.9) u (1000

()

p ) + (.1) u (750 + (250

Expected Utility

p ) ).

Spring 2014

22 / 28

Application: Demand for Insurance

^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Denition: Insurance is actuarially fair if the premium equals the expected
payout.
(Insurance is actuarially unfair if the premium is larger than the expected
payout, and insurance is actuarially overfair if the premium is smaller than
the expected payout.)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Lets assume insurance is actuarially fair.
Here, actuarially fair insurance means p = (.10)(250) = $25.

()

Expected Utility

Spring 2014

23 / 28

Application: Demand for Insurance

Substituting p = $25 into EU (x()):

EU (x()) = (.9)u (1000

25) + (.1)u (750 + 225).

^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Conclusion: If insurance is actuarially fair, and if you are risk averse, then
you will choose to fully insure.

()

Expected Utility

Spring 2014

24 / 28

Application: Demand for Financial Assets

Suppose you have wealth 1000, and that you plan to invest it for one year
and then consume it. Two assets in which to invest:
A risk-free asset that yields a certain return of 0%.
A risky asset that yields a return of 50% with probability 1/2 and a
return of 40% with probability 1/2.
Let be the amount you invest in the risky asset, so 1000 is the
amount you invest in the risk-free asset. (Assume 0 1000.)
If do well, nal wealth =
(1000 )(1.00) + (1.50) = 1000 + (.50).
If do poorly, nal wealth =
(1000 )(1.00) + (0.60) = 1000
()

Expected Utility

(.40).
Spring 2014

25 / 28

Application: Demand for Financial Assets

As a function of the you choose, you face lottery:

x() = ( 1000 + (.50) ,

1
; 1000
2

1
).
2

(.40) ,

As a function of the you choose, your expected utility is:

EU (x()) =

()

1
1
u (1000 + (.50)) +
u (1000
2
2

Expected Utility

(.40)).

Spring 2014

26 / 28

Expected Utility and Revealed Preferences

Recall that a theoretical goal is to prove theorems of the form: A persons
behavior can be described by a model if and only if the persons choices
satisfy some set of axioms.
Axioms (properties about a persons choices over lotteries):
(1) Preferences are complete and transitive.
(2) Preferences are continuous.
(3) Independence axiom: If x % y, then for any lottery z,
(1 )x + z % (1 )y + z for any 0 < < 1.
Theorem: A persons behavior can be described by expected utility theory
if and only if her choices over lotteries are complete, transitive, continuous,
and satisfy the independence axiom.
()

Expected Utility

Spring 2014

27 / 28

Expected Utility A Final Thought

Many people view EU to be a good normative theory.

EU seems to fail as a positive (descriptive) theory. Thats where well go
next....

()

Expected Utility

Spring 2014

28 / 28