Foundations in Calculus

and Probability
YU SHENG LOH
Utility Functions
The St. Petersburg Paradox
The St. Petersburg game is played by flipping a Expected
fair coin until it comes up tails, and the total 𝑛 𝑃(𝑛) Prize
Payoff
number of flips, 𝑛, determines the prize, which 1 1⁄2 2 $1
equals $2!
2 1⁄4 4 $1
If you were a “rational” gambler, how much
3 1⁄8 8 $1
would you pay to participate in this game?
4 1⁄16 16 $1
5 1⁄32 32 $1
6 1⁄64 64 $1
7 1⁄128 128 $1
8 1⁄256 256 $1
9 1⁄512 512 $1
10 1⁄1024 1024 $1
Utility
Utility
◦ The satisfaction that an individual obtains from a particular course of action
◦ E.g. the happiness obtained from eating a scoop of ice cream

Risk (uncertainty): a situation in which the probabilities of the different possible outcomes are
known, but it is not known

50% $200

$125

50% $50
Utility Function
Utility can be assigned to each possible value of the investor’s wealth by utility
◦ Quantifying utility

Types of utility functions

◦ Log utility function: 𝑈 𝑤 = log 𝑤 , for 𝑤 > 0
"
◦ Quadratic utility function: 𝑈 𝑤 = 𝑤 + 𝑑𝑤 ! , for − ∞ < 𝑤 < − !#
$ ! %"
◦ Power utility function: 𝑈 𝑤 = &
, for 𝑤 > 0
%'$
◦ Exponential utility function: 𝑈 𝑤 =𝑒
Expected Utility Theorem
A utility function 𝑈 𝑤 can be constructed as representing an investor’s utility of wealth 𝑤 at
some future time
Decisions are made in a manner to maximise the expected value of utility given the investor’s
particular beliefs about the probability of different outcomes
Example: An investor who has the log utility function faces the uncertainty described in the
previous slide. What is the expected utility of wealth?
◦ Expected utility 𝐸[𝑈(𝑤)] = 0.5 log(200) + 0.5 log(50) = 4.6052

What is the utility of expected wealth?

◦ Expected wealth 𝐸 𝑤 = 0.5 200 + 0.5 50 = $125
◦ Utility of expected wealth 𝑈 𝐸 𝑤 = log 125 = 4.828
Expected Utility Theorem
𝑈 𝐸𝑤 > 𝐸[𝑈(𝑤)]
Non-satiation
◦ 𝑈( 𝑤 > 0
◦ Investor prefers more wealth to less
◦ Positive marginal utility of wealth

Risk aversion
◦ 𝑈 (( 𝑤 < 0
◦ Risk-averse investor values an increase in wealth less than a
decrease
◦ Will reject a fair gamble
Certainty Equivalent
There will exist a certain guaranteed outcome, at which the investor will be indifferent between
the risky investment, and the guaranteed outcome
From the previous example, 𝐸[𝑈(𝑤)] = 4.6052
𝑈 𝑤 ∗ = 𝐸[𝑈(𝑤)] = 4.6052
log 𝑤 ∗ = 4.6052
𝑤 ∗ = 𝑒 #.%&'( = 100
The investor is in fact willing to give up $25 in order to not be faced with risk
Optimal Investment
When making an investment decision, investors wish to maximise their utility:
An investor is considering investing $1,000 for a period of one year and is faced with the
following two options:
◦ Asset A, a risk-free asset with a return of 8.5%
◦ Asset B, a risky asset with will either return 15% with a probability of 0.4, or 4% with a probability of 0.6

The investor is determining the proportions 𝑥) and (1 − 𝑥) ) to invest in Asset A and Asset B,
respectively
Assume the investor has a quadratic utility function 𝑈 𝑤 = 10*# 𝑤 ( + 10*( 𝑤, where 𝑤 > 0
Optimal Investment
1. Set up equation for expected utility, 𝐸[𝑈 𝑤 ], in terms of 𝑥)

2. Differentiate 𝐸[𝑈 𝑤 ] with respect to 𝑥)

+,[. / ]
3. Set =0
+1!

4. Solve for 𝑥)
Optimal Investment
𝐸 𝑈 𝑤

= 0.4𝑈 1085𝑥) + 1150 1 − 𝑥) + 0.6𝑈(1085𝑥) + 1040 1 − 𝑥) )

= 0.4𝑈 1150 − 65𝑥) + 0.6𝑈 1040 + 45𝑥)
""*+%,*-" # ""*+%,*-" "+.+/.*-" # "+.+/.*-"
= 0.4 + + 0.6 +
"++++ "++ "++++ "++
# #
""*+#%! ,* ""*+ -"/,*#-" ""*+%,*-" "+.+#/! .* "+.+ -"/.*#-" "+.+/.*-"
= 0.4 + + 0.6 +
"++++ "++ "++++ "++

= 0.4 132.25 − 14.95𝑥) + 0.4225𝑥)! + 11.5 − 0.65𝑥) + 0.6 108.16 + 9.36𝑥) + 0.2025𝑥)! + 10.4 + 0.45𝑥)
= 52.9 − 5.98𝑥) + 0.169𝑥)! + 4.6 − 0.26𝑥) + 64.896 + 5.616𝑥) + 0.1215𝑥)! + 6.24 + 0.27𝑥)
= 128.636 − 0.354𝑥) + 0.2905𝑥)!
Optimal Investment
𝐸𝑈 𝑤 = 128.636 − 0.354𝑥) + 0.2905𝑥)(
#0[2($)]
= 0.581𝑥) − 0.354
#-"

Set 0.581𝑥) − 0.354 = 0

0.581𝑥) = 0.354
+.7*.
𝑥) =
+.*8"

𝑥) = 0.6093
60.93% of assets to be invested in risk-free asset Asset A
Dominance
What makes one asset better than another?
If return of Asset A is higher than Asset B under every possible state of outcome

Poor Average Good

Asset A -2% 5% 12%
Asset B -4% 2% 8%
Absolute dominance
◦ This rarely occurs
◦ E.g. expected returns of risk asset is higher than risk-free asset. However, in a bad state, returns of risk
asset is lower
Stochastic Dominance
Probabilistic way of comparison
Does not require specification of an individual’s utility
function
But, instead, able to rank outcomes and preferences
based on cumulative distribution function (CDF)
Factors in assumptions of non-satiation and risk
aversion
◦ Higher returns are better than lower returns
◦ Less risk is preferable to more risk

Consider two investment portfolios, 𝐴 and 𝐵, and their

respective CDF of returns, 𝐹) and 𝐹2
First-Order Stochastic Dominance
Using assumptions of non-satiation
◦ Higher returns are better than lower returns

A will dominate B (investor prefers portfolio A to

portfolio B) if
◦ 𝐹) ≤ 𝐹9 for all 𝑥, and
◦ 𝐹) < 𝐹9 for some value(s) of 𝑥

𝐹) = 𝑃(𝑥 < 𝑋)
For a given level of return, you want a smaller
probability that this can be achieved
◦ This implies that the probability of achieving a return
higher than the given level of return is higher
◦ 1 − 𝐹) ≥ (1 − 𝐹9 )
First-Order Stochastic Dominance
Example: Assume Asset X and Asset Z offer the following possible outcomes:
Cumulative Probability
Asset X Asset Z Return Asset X Asset Z
Return Probability Return Probability 5% 1⁄2 0
7% 1⁄2 8% 1⁄2 6% 1⁄2 1⁄2
7% 1⁄2 6% 1⁄2 7% 1 1⁄2
8% 1 1

The cumulative probability shows that

◦ 𝐹: ≤ 𝐹; for 𝑥 = 5, 6, 7, 8
◦ 𝐹: < 𝐹; for 𝑥 = 5, 7

Asset Z first-order stochastically dominates Asset X -> investor should choose Asset Z
Second-Order Stochastic Dominance
Using assumptions of non-satiation and risk-aversion
◦ Higher returns are better than lower returns
◦ Less risk is preferable to more risk

A will dominate B (investor prefers portfolio A to

portfolio B) if
1 1
B 𝐹) 𝑦 𝑑𝑦 ≤ B 𝐹2 𝑦 𝑑𝑦 ∀𝑥
*3 *3
𝐹9 𝐹)
1
This also implies ∫*3 𝐹2 𝑦 − 𝐹) 𝑦 𝑑𝑦 ≥ 0 ∀𝑥
For instance, if two normal distributions have the same
means but different variances, the one with the lower
variance second-order stochastically dominates the
other
Second-Order Stochastic Dominance
Example: Assume Asset U and Asset V offer the following possible outcomes:
Cumulative Probability 𝚺 Cumulative Probability
Return Asset U Asset V Return Asset U Asset V
5% 1⁄4 0 5% 1⁄4 0
6% 1⁄2 3⁄4 6% 3⁄4 3⁄4
7% 3⁄4 3⁄4 7% 6⁄4 6⁄4
8% 1 1 8% 10⁄4 10⁄4
The sum of cumulative probability shows that
◦ ∫ 𝐹< ≤ ∫ 𝐹2 for 𝑥 = 5, 6, 7, 8
◦ ∫ 𝐹< < ∫ 𝐹2 for 𝑥 = 5

Asset V second-order stochastically dominates Asset U

◦ Risk-averse investors should choose Asset V, which offers the same expected return as Asset U but with
a lower variance
Stochastic Dominance
You are evaluating two investment portfolios, A and B, with your cumulative distribution
" "
functions of wealth, 𝑤: 𝐹) 𝑤 = 𝑤 , 𝐹2 𝑤 = 𝑤 for 0 ≤ 𝑤 ≤ 1
# $

a. Show that Asset A first-order stochastically dominates Asset B

b. Show that Asset A second-order stochastically dominates Asset B

Stochastic Dominance
a.
" "
For 0 ≤ 𝑤 ≤ 1, 𝑤 ≤ 𝑤
# $

" "
For 0 < 𝑤 < 1, 𝑤 < 𝑤
# $

Therefore, the two conditions for first-order stochastic dominance hold and portfolio A first-
order stochastically dominates portfolio B
Stochastic Dominance
b.
1 / " 1 / "
∫4 𝐹) 𝑦 𝑑𝑦 = ∫& 𝑦 # 𝑑𝑦 ∫4 𝐹2 𝑦 𝑑𝑦 = ∫& 𝑦 $ 𝑑𝑦
/ /
# % # % % & % &
= 𝑦# = 𝑤# = 𝑦$ = 𝑤$
' & ' 5 & 5

%
#
∫ 7! (/) %
/# (: %*& (: ()*+, (: +
(: "
= & = 𝑤# $ = 𝑤 +# = 𝑤 +# = 𝑤 "+
∫ 7' (/) $
/ $
;& ;& ;& ;&
&
Stochastic Dominance
b.
"
(:
< 1 and 𝑤 "+ <1
;&
"
(:
Therefore, 𝑤 "+ <1
;&
This implies that:
◦ ∫ 𝐹) ≤ ∫ 𝐹9 for 0 ≤ 𝑤 ≤ 1
◦ ∫ 𝐹) < ∫ 𝐹9 for 0 < 𝑤 < 1

Therefore, the two conditions for second-order stochastic dominance hold and portfolio A
second-order stochastically dominates portfolio B
Practice Questions
1. An investor has a quadratic utility function: 𝑈 𝑤 = 𝑤 − 10*' 𝑤 ( . Her income is uncertain
and will be either $40,000 with 75% probability, or $30,000 with 25% probability.
◦ Find the income range that is appropriate for this utility
◦ Determine the minimum level of guaranteed income that makes her indifferent from the uncertain
income

2. You are evaluating two investment portfolios, A and B, with your cumulative distribution
"
functions of wealth, 𝑤: 𝐹) 𝑤 = 𝑤, 𝐹2 𝑤 = 𝑤 # for 0 ≤ 𝑤 ≤ 1
◦ Determine which asset first-order stochastically dominates the other
◦ Determine which asset second-order stochastically dominates the other