You are on page 1of 5

Choice Under Uncertainty II

Contingent Consumption
Preferences Under Uncertainty
Expected utility using states of nature outside individuals control (been using)
Actions can be taken to alter the consumption bundles(/wealth) available in different states of nature
o Ex. if we back up computer files to external discs, we reduce our loss in the event of a hard
disc crash, but at a cost (in time, effort, and the external medium itself)
o Ex. fire an theft insurance has a cost (premium), but changes the consumption bundle
available in the event of burglary or a fire
Consumer can choose substitute consumption in one state of nature (no fire) for consumption in
another (house burns down), to increase expected utility
State-contingent consumption plans that give = expected utility are equally preferred
State-Continent Budget Constraints
Ex. accident insurance with variable face value (insurance buyer can choose coverage amount)
o Each \$1 of accident insurance costs
o Consumer has \$m of wealth
o Cna = consumption value in no-accident state
o Ca = consumption value in accident state
o

Without insurance
Ca = m L
Cna = m where L = \$amount of loss
If L is larger relative to m consumption may be very different in the 2 states

Buy \$K of accident insurance at a premium rate of per dollar of face value (coverage)
o Cna = m K
o Ca = m L K + K = m L + (1 )K
Consumption in event of insured loss is:
o Base income m less
o Loss plus
o The net proceeds of insurance (claim less premium)
Consumption is there is no loss is:
o m less insurance premium paid
Buy \$K of accident insurance at rate
o Cna = m [ins. Prem.] = m - K
o Ca = m L K + K = m L + (1 )K
o Solving for K
K = (Ca m + L) /(1 )
o Substituting
Ca = m (Ca m + L) /(1 )

Numerical Example
Initial wealth = \$50,000
Amount of loss = \$40,000
Amounts of consumption in accident and no-accident states when different
amounts of insurance are bought are:
o CNA = 50,000 (0.1)(Face)
o CA = 50,000 40,000 + (0.9)(Face)

face

Preferences Under Uncertainty

What is the MRS of an indifference curve?
Get consumption C1 w/ prob 1 and C2 with prob 2
(1 + 2) = 1
EU = 1U(c1) + 2U(c2)
For constant EU, dEU = 0

Choice Under Uncertainty

How is rational choices made under uncertainty?
o Choose the most preferred affordable state-contingent consumption plan
State-Contingent Budget Constraints

Competitive Insurance
Suppose entry to insurance industry is free
Expected economic profit = 0
K - aK (1 a)0 = ( - a)K = 0
free entry = a
if price f \$1 insurance = accident probability then insurance = fair
(insurers premium revenue just =s expected claims payments)
when insurance is fair (premium per \$ = probability of claim), rational insurance choices satisfy

Marginal utility of income must be the same in both states

How much fair insurance does a risk-averse consumer buy?
o Risk-aversion MU(c) decrease and c increases
o Risk averse consumer buys full insurance

Unfair Insurance
Suppose insurers make +ve expected economic profit
K - aK (1 a)0 = ( - a)K > 0
Then > a

rational choice

requires:

since

o
for risk-averter

MU(ca) = MU(cna)

a risk-averter buys less than full unfair insurance

Example
Jimmy owns racing car enters in competitions
Utility of wealth function U(W) = (W)0.5
If car crashes = destroyed 0 wealth
Not in a crash 252,000 wealth (value of car); he owns no other wealth
Due to nature of race-car driving and the skill of he drivers he hires prob of his car being destroyed
in a year = 1/6
Can buy \$K of insurance on his car ad rate of (\$2/7) for every dollars worth of insurance he buys
How much insurance will Jimmy buy, if he maximizes expected utility over the states (crash,no crash)?
Budget Constraint
NO insurance
o Wealth in state N (no crash) = 252,000
o Wealth in state C (crash) = 0
FULL insurance bought
o Pays premium (2/7)(252,00) = \$72000
o Wealth at state N 252,000 72000 = \$180,000
o Wealth in state C 252,000 (ins. Claim) 72,000 (prem) = \$180,000

Form:
PNACNA + PACA < B
Can determine B when we choose PNA
When CNA = 252,000 CA = 0 (no insurance
When CNA = 180,000 CA = \$180,000 (full insurance)
PA = 2/7 (ins prem rate per \$1 face)
o Choose PNA = (1 _ PA) = 5/7
o Then B 180,000 = (5/7)252,000 + (2/7)0
o And B 180,000 = (5/7)180,000 = (2/7)180,000
Budget Constraint
o (5/7)CNA + (2/7)CA < 180,000
each \$1 unit of CN costs \$(5/7) [pay \$5, get \$5 + \$2 save premium]
each \$1 unit of CC costs \$(2/7) [pay \$2 premium, get \$7 of claim]
budgetconstraint across 2 states:
o (5/7)CN + (2/7)CC = 180,000 for Cc < 180,000
o he cant buy more than 100% insurance and will buy less than 100% since prem is unfair
prob of loss = 1/6 (16.67%, prem = 2/7 = 28%
tangency condition:

Tangency combined with Budget Constraint

sub optimal amount fo state C consumption into budget constraint
(5/7)CN + (2/7)CC = 180,000
(5/7)[4 CC] + (2/7)CC = 180,000
(20/7)CC + (2/7)CC = (22/7) CC = 180,000
CC = [180,000](7/22) = \$57,272
CN = \$229,091
(4CC)
o 252,000 229,092 = \$22,909 is spent on prem, to buy (7/22)22,909 = \$80,181 of face value

in state C, Cc = \$80,181 (claim) - \$22,909 (premium) = \$57,272

Uncertainty is pervasive
what are rational responses to uncertainty?
o A portfolio of contingent consumption goods
Diversification
Two firms A and B
Shares cost \$10
Prob As profit = 100
Prob As profit = 20
You have 100 to invest.

Bs profit = \$20
Bs profit = 100
How?

\$100/10 = 10 shares
you earn \$1000 with prob and \$200 with prob
expected earning: \$500 + \$100 = \$600
\$100/10 = 10 shares
you earn \$1000 with prob and \$200 with prob
expected earning: \$500 + \$100 = \$600
o buy 5 shares each firm
earn \$600 for sure
diversification has maintained expected earnings and lowered risk
typically, diversification lowers expected earnings in exchange for lowered risk
o

Example
buy shares in: Shell Oil, Day & Ross Trucking
state 1: Low Oil price
o trucking profits high but oil profits low
state 2: High Oil Price
o trucking profits low (cost of fuel high), but oil profits high
Risk Eliminates (Generally Reduced) Through Risk Pooling
principle is that if we have enough investment, the returns from which are not perfectly positively
correlated
o not all of them lose at the same time
the pool of risks has a smaller risk (technically measured by variance) than does any single asset
(share)
100 risk-averse persons each independently risk a \$10,000 loss
loss prob = 0.01 (1%)
initial wealth is \$40,000
no insurance: expected wealth
o 0.99 x \$40,000 + 0.01(40,000 10,000) = \$39,900
o expected utility < utility \$39,900 with certainty (by risk aversion)
Mutual Insurance
each of the 100 persons pays \$100 into a mutual insurance fund
total premium = \$100*\$100 = \$10,000
premium collections = to loss to one (unlucky) contributor
on average 1/100 experiences a \$10,000 loss each period is compensated fully funds paid in
certain wealth for each of 99 w/o loss = \$39,900
loss of \$10,000 fully compensated, loser has wealth \$39,900 - \$10,000 = \$39,900
iff each consumer = risk-averse, utility of \$39,900 with certainty > expected utility when
expected wealth = \$39,900

risk-spreading implies every contributor has higher utility than with no mutual insurance
Risk Pooling
many independent gambles, some win, some lose
all gains and losses are shared
if gambles perfectly vely correlated, risk eliminated (there is no loss to anyone)
o imperfectly correlated risk reduced