Roadmap

• Introductions, course outline, and administration

QRAM 8600 • Arbitrage
Theory of Risk Sharing • Expected Utility

Lecture 1

Stephen H. Shore
Fall 2022

1 2

Stephen Shore: Introduction Get to know each other

• Raised in San Francisco Bay Area
• Princeton undergraduate 1998
• Harvard Ph.D. 2003 (studied finance)
• Taught at Wharton 2003-2007
• Taught at Johns Hopkins 2007-2011
• Deputy Assistant Secretary of the Treasury
2011-2012
• At GSU since 2012; now RMI Department Chair.
• In my research, I look at risks faced by
households, income risk and consumer credit.

3 4

Contact Information Course Resources

• Office hours: Wednesdays 11am-noon am or by • Course website on iCollege
appointment
• Textbook: Eeckhoudt, L., C. Gollier, and H.
• Location: My office on 11th floor of 35 Broad St.
Schlesinger, Economic and Financial
• Email [email protected] Decisions under Risk, Princeton University
Press, 2005.
• Handouts

• iCollege Tour/Content

5 6

1
 Course Assignments
• Go to iCollege Monday after 4:30pm to find
assignments for the week. (Practice problems,
videos, class instructions, etc.)
• Do these assignments in advance of Thursday What Are Capital Markets
4:30pm-7:00pm class. and What Are They For?
• Come to class, engage with the material, and
interrupt me! (1/12 class participation grade)
• 3 problem sets (1/12 each homework grade)
• 2 midterms, 1/6 each midterm grades
• Final exam, 1/3 of the grade

7 8

What Are Capital Markets? Classifying Financial Assets

• Real assets (investments) require an input of • Fixed-income securities promise to make fixed
resources today and deliver an output of payments in the future.
resources later. (example: farm) – bills (short-term), bonds (longer-term)
• Financial assets (investments) are claims to the • Equities (stocks) are claims to a share of the
output produced by real assets. profit of a corporation.
• Capital markets (financial markets) are the • Derivative securities (derivatives) make
markets in which financial assets are traded. payments that depend on the prices of other
• Time and uncertainty are crucial! financial assets.
– futures, swaps, options

9 10

Classifying Financial Assets What Are Capital Markets For?

• Shift the timing of consumption so it need not
Uncertainty of payoff
coincide with the timing of income
Stocks • Allocate risk (e.g. share it among many people,
Options
or concentrate it on those who are most willing
to bear it)
• Allocate resources to the most productive real
investments (e.g. by separating the ownership
Bills Bonds and management of companies)
(money market) • Aggregate the information of market participants,
thereby revealing it to others.
1 year Time to payoff

11 12

2
 Sketch of the Financial System

The Institutional Perspective

and the Functional Perspective

13 14

Perspectives on the System Institutional Sketch

Institutional perspective:
• Example: what do banks do?
– Meeting place for borrowers and lenders
– Pool resources of lenders to make large loans
– Diversify by lending to many borrowers
– Build expertise through volume of lending
• Takes existence of banks as given

15 16

Perspectives on the System Functional Sketch

Functional perspective:
• Example: what needs to be done?
– Workers need to accumulate funds for retirement
– A large factory needs to be financed
– Homeowners need insurance against fires and
hurricanes
• Takes clients’ needs as given

17 18

3
 Our Perspective
• In this course, we will emphasize functions of the
financial system
• This functional approach is a great strength of
the academic field of finance
Basic Concepts of Finance

19 20

Basic Concepts of Finance

• Arbitrage
– What do we know about asset prices by just
assuming that there are no risk-free profit
opportunities?
The Discrete-State Model
• Optimality
– What is the best combination of risky financial assets,
given their returns and an individual investor’s
preferences?
• Equilibrium
– What is the nature of asset prices and allocations
when all investors have optimized and the markets
clear?

21 22

A Simple Framework A Simple Framework

We will explore these concepts within a simple • Assets are defined by their payoffs. Asset i pays
framework: off Xij in state j.
• Two periods, “now” and “the future” • The price of asset i is written pi.
• A finite number S of states may occur in the • The N securities are jointly defined by an N-by-S
future. The states are numbered j=1…S. matrix X whose i,j element is Xij. (each security
• A finite number of assets are available. The is a row; each state is a column)
assets are numbered i=1…N. • The prices of the N securities can be stacked
into a N-by-1 vector p.

23 24

4
 Portfolio Construction Example
• We can form a portfolio by buying wi shares of States:
asset i, i=1…N. If we stack the number of Farm: [3 2 1] (j= rain, small drought, large drought)
shares into a N-by-1 vector w, then
– the payoff on the portfolio is W=w’X Assets:
((1xN)x(NxS)=1xS, payoff in each of S state) • [3 2 1] (i=1 whole farm, p1=2)
and
• [2 2 1] (i=2, first 2 dollars produced, p2=1.75)
– the price of the portfolio is w’p. (1xN)x(Nx1)=1
• [1 1 1] (i=3 first dollar produced, p3=1)
• The portfolio is just a new asset that we have
constructed using the old ones.

25 26

Example continued portfolios

States are columns, assets are rows w (3x1) = (for example)
X (3x3) = p (3x1)=

1
3 2 1 2 -1
2 2 1 1.75 0
1 1 1 1

27 28

Comprehension Questions
• Consider a portfolio of 1 unit of asset 1
and -1 units of asset 2.
Complete Markets and Arrow
• What is the cost of buying this portfolio of
securities? Debreu Securities
• What are the payoffs of this portfolio in
the three states?

29 30

5
 Complete Markets Complete Markets
• Suppose we can construct a portfolio that pays
• Markets are complete if S securities exist with
$1 in state j and $0 in all other states. This linearly independent payoffs.
asset is called an Arrow-Debreu security for • In this case the matrix X that we start with is S-
state j. by-S and has full rank. Thus its inverse exists.
• We write its price as qj and call it the state price • To construct S Arrow-Debreu securities, define
for state j. an S-by-S matrix of portfolio weights X-1.
• The payoff on the set of portfolios is X-1X=I.
• If there are S Arrow-Debreu securities, one for
• The cost of the set of portfolios is X-1p=q.
each state, then the payoff matrix X = I, the
• To construct any other vector of payoffs x, buy
identity matrix. x’X-1 shares of the securities at cost
• In this case we say that markets are complete. x’X-1p=x’q.
• We write the set of state prices in a vector q.

31 32

A-D security construction A-D construction, continued

w1 = w2= w3 W=inverse(X)= [w1 w2 w3]

0 1 -1 0
1 -1
-1 -1 2 -1
-1 2
2 0 -1 2
0 -1

33 34

A-D security prices Riskless rate

q1 = w1 ‘ * p = 0.25 • Sum of all A-D security prices is cost of
getting $1 for sure.
q ( 3 x 1) = w ‘ * p = • Riskless rate is just:
0.25
– 1/(sum of all A-D security prices)-1
0.5

0.25

35 36

6
 Incomplete markets Complete Markets
Imagine asset 1 didn’t exist and someone tried to • The Law of One Price is most helpful when
introduce it; what do we know about its price? markets are complete
We know all A-D securities must have non-negative • Then any asset can be priced by breaking it into
prices primitive state-contingent payoffs, and adding up
p= q = W’p = the prices of each piece
p1 q1=p1-1.75>0
1.75 q2=-p1+2*1.75 -1 >0
1 q3=-1.75+2=0.25
p1>=1.75 or q1<0; p1<=2.5 or q2<0

37 38

Concept Check Lottery Tickets and rf

• The numbers lottery is a game where
– You buy one ticket with three-digit number for $1 (for
one particular integer from 1 to 1,000)
– A number is drawn randomly
– If your number comes up you make $900, otherwise 0
• There are 1000 states, and 1000 securities
• Question: If you buy a specific numbered ticket,
how many units of a specific Arrow-Debreu
security are you buying?
• Question: What is the riskfree rate of interest in
the numbers lottery?
(Go into breakout rooms and work it out.)

39 40

Are Markets Complete? Coin Toss Example

• At first sight, this seems most unlikely • Consider tossing a coin twice and counting the
• There are thousands of traded financial assets, number of heads
but surely there are far more states of the world • There may be 0, 1, or 2 heads so there are three
than assets final states
• Not so fast! • We can represent uncertainty by an event tree.
– We can ignore irrelevant dimensions of uncertainty,
i.e. those that do not affect asset payoffs
– Frequent trading in long-lived assets can sometimes
reduce the number of assets we need to get complete
markets

41 42

7
 Coin Toss Example Coin Toss Example
2 ¼
• For each coin toss, suppose there are two
H½ assets:
H½ T½ – one “head” asset that costs $0.50 and pays $1 if the
½
toss is a head, $0 otherwise
1
H½ – one “tail” asset that costs $0.50 and pays $1 if the
T½
toss is a tail, $0 otherwise
T½
0
• To get $1 only if there are two heads, invest
¼
$0.25 at the first toss in half a unit of the head
asset. If the first toss is a head, reinvest your
winnings in one unit of the head asset.

43 44

Coin Toss Example Coin Toss Example

• The same approach works to get $1 only if there • The lesson: Dynamic trading in two assets can
are zero heads (two tails). achieve complete markets for more than two
• Question: How should you invest in order to get final states.
$1 only if there is one head? • In fact, you only need as many assets as there
are branches leaving each node of the event
• (Costs $0.5. Buy half of each asset at start $0.25 each. One pays off for $0.5. Then buy one of the asset that didn’t pay off last time with the $0.5.)

tree.
• It is common to value options by modelling
uncertainty with two-branch (binomial) or three-
branch (trinomial) event trees.

45 46

So Are Markets Complete?

• Even with dynamic trading, we must recognize
the possibility of events that are not captured in
a simple model: extra branches on the event
tree Arbitrage in Theory
• Nonetheless, complete markets models are
extremely useful devices for thinking about
valuation of complex securities.

47 48

8
 Arbitrage Arbitrage
An arbitrage opportunity is an investment strategy Financial theory assumes there are no arbitrage
that opportunities
• never requires a cash outflow now or in the • Any opportunities that may arise are exploited so
future quickly that we never observe them
• This theory should not be taken literally, but as a
• generates a cash inflow either benchmark that helps us to identify, profit from,
– in one or more states in the future but not now (type 1 and rapidly eliminate arbitrage opportunities
arbitrage opportunity), or
– now and possibly in one or more states in the future
too (type 2 arbitrage opportunity)

49 50

Implications of No-Arbitrage Arbitrage and q>0

• Law of One Price. Two assets or portfolios with Imagine asset 1 existed with a price above 2.5 or
the same payoffs in every state of the world below 1.75. How could you make money?
must have the same price.
– If not, short the expensive one and buy the cheap
We know all A-D securities must have non-negative
one! Cash inflow today with no costs in any future prices
state of the world. (Type 2 arbitrage.) p= q = W’p =
• Arrow-Debreu securities must have positive p1 q1=p1-1.75>0
state prices.
– If one has a negative price, buy it! Cash inflow today 1.75
and in one future state of the world, with no costs in q2=-p1+2*1.75 -1 >0
any other state. (Type 2 arbitrage.)
– If one has a zero price, buy it! Cash inflow in one 1 q3=-1.75+2=0.25
future state of the world, with no costs in any other p1>=1.75 or q1<0; p1<=2.5 or q2<0
state. (Type 1 arbitrage.)

51 52

Implications of No-Arbitrage Stochastic Discount Factor

• If markets are complete, these restrictions allow • Absence of arbitrage means there exists an SDF
us to price all possible assets. (Stochastic Discount Factor):
– q(j)=pi(j) * m(j) (m(j) is the SDF for state j, q is the
• We already have the state price vector q.
price of the A-D security, pi is the probability)
• To price any other asset i with payoff vector xi, – SDF indicates how much more or less the A-D
just use pi = xi’q. security for a state costs than its probability.
• The “Happy Meal Theory”: the price of a Happy – The prices of existing assets can be priced as if there
Meal is the sum of the prices of small fries, is are A-D securities.
cheeseburger, drinks, and toy. – With complete markets, A-D security prices are
unique. When incomplete, may be multiple sets of A-
D security prices that rationalize asset prices.

53 54

9
 Expected Utility Expected Values?
• Up to now, we only required that agents • Why not choose the portfolio with the
are not satiated, that is, they like more to highest expected value?
less - the implication is the absence of • Portfolio w (Nx1 shares in N assets) has
arbitrage opportunities, and the payoff w’X (1xN)x(NxS)=(1xS) payoffs in
consequences discussed. each of S states.
• How do agents choose among risky • Let pi be the Sx1 vector of probabilities of
(uncertain) alternatives? each of S states pi(j)
• Decision theory under risk: expected utility • Expected payoff is w’Xpi (1xS)x(Sx1)=1x1
theory – Sum over all states j of pi(j)*payoff(j)

55 56

St Petersburg Paradox
• A casino offers a game of chance for a single player in
which a fair coin is tossed at each stage. The pot starts
at 2 dollars and is doubled every time a tail appears. The
first time a head appears, the game ends and the player
wins whatever is in the pot. Thus the lottery pays 2n if the
first head appears at the n-th toss (and pays zero before
a head occurs and after that).
• What is the expected value of this gamble X, known as
the “St. Petersburg lottery"?
• EX =0.5*2+0.25*4+0.125*8+…=inf
• How much would you pay for this gamble?

57

10