Investments

Investments
Teaching notes
João Pedro Pereira

Nova School of Business and Economics
Universidade Nova de Lisboa
joao.pereira@novasbe.pt
https://sites.google.com/view/jpereira
November 14, 2022

Contents
1 Portfolio Choice 5
1.1 Risk and return . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.1.1 Holding period return . . . . . . . . . . . . . . . . . . . . . 5
1.1.2 Expected return and standard deviation . . . . . . . . . . . 6
1.1.3 Risk premium and Sharpe Ratio . . . . . . . . . . . . . . . 7
1.1.4 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.2 Risk Aversion and Capital Allocation . . . . . . . . . . . . . . . . . 8
1.2.1 Risk Aversion . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.2.2 Portfolio rate of return . . . . . . . . . . . . . . . . . . . . . 10
1.2.3 Portfolios of one risky asset and a risk-free asset . . . . . . 12
1.2.4 Risk tolerance and asset allocation . . . . . . . . . . . . . . 13
1.2.5 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
1.3 Optimal Risky Portfolios . . . . . . . . . . . . . . . . . . . . . . . . 15
1.3.1 Diversification and portfolio risk . . . . . . . . . . . . . . . 15
1.3.2 Portfolios of two risky assets . . . . . . . . . . . . . . . . . 16
1.3.3 Portfolios of two risky assets and a risk-free asset . . . . . . 18
1.3.4 The Markowitz portfolio optimization model . . . . . . . . 20
1.3.5 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2 The Capital Asset Pricing Model 22

2.1 Assumptions and derivation . . . . . . . . . . . . . . . . . . . . . . 22
2.2 Efficient frontier . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.2.1 Capital Market Line . . . . . . . . . . . . . . . . . . . . . . 24
2.2.2 Risk aversion and the market risk premium . . . . . . . . . 25
2.3 Expected returns on individual securities . . . . . . . . . . . . . . . 26
2.3.1 Security Market Line . . . . . . . . . . . . . . . . . . . . . . 26
2.3.2 Interpretation of beta . . . . . . . . . . . . . . . . . . . . . 27
2.3.3 Beta of a portfolio . . . . . . . . . . . . . . . . . . . . . . . 28
2.4 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
3 Arbitrage Pricing Theory and Factor Models 30

3.1 Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
3.2 Example with 1 factor: the Market model . . . . . . . . . . . . . . 33
2
Contents 3
3.3 Example with 3 factors: the Fama-French model . . . . . . . . . . 34

3.3.1 Pricing equation . . . . . . . . . . . . . . . . . . . . . . . . 34
3.3.2 Details on the factors . . . . . . . . . . . . . . . . . . . . . 34
3.4 Portfolio performance evaluation . . . . . . . . . . . . . . . . . . . 35
3.4.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
3.4.2 Adjusting returns for risk . . . . . . . . . . . . . . . . . . . 35
4 Bond Markets 38
4.1 Price quotes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
4.1.1 Face value and coupons . . . . . . . . . . . . . . . . . . . . 39
4.1.2 Bond price units . . . . . . . . . . . . . . . . . . . . . . . . 39
4.1.3 Quoting conventions . . . . . . . . . . . . . . . . . . . . . . 40
4.2 Rates of return . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
4.2.1 Annual Percentage Rate . . . . . . . . . . . . . . . . . . . . 42
4.2.2 Effective Annual Rate . . . . . . . . . . . . . . . . . . . . . 42
4.3 Bond Prices and Yields . . . . . . . . . . . . . . . . . . . . . . . . 43
4.3.1 Bond Pricing . . . . . . . . . . . . . . . . . . . . . . . . . . 43
4.3.2 Yield to Maturity . . . . . . . . . . . . . . . . . . . . . . . 44
4.3.3 Realized Yield . . . . . . . . . . . . . . . . . . . . . . . . . 46
4.3.4 Default risk and ratings . . . . . . . . . . . . . . . . . . . . 46
4.3.5 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
4.4 Term structure of interest rates . . . . . . . . . . . . . . . . . . . . 47
4.4.1 Spot rates . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
4.4.2 Forward rates . . . . . . . . . . . . . . . . . . . . . . . . . . 48
4.4.3 Expectations of future interest rates . . . . . . . . . . . . . 49
4.4.4 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
4.5 Yield curve fitting . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
4.5.1 Continuous compounding . . . . . . . . . . . . . . . . . . . 51
4.5.2 Discount factors and interest rates . . . . . . . . . . . . . . 51
4.5.3 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
4.5.4 Bootstrap . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
4.5.5 Nelson-Siegel Model . . . . . . . . . . . . . . . . . . . . . . 56
4.5.6 Svensson Model . . . . . . . . . . . . . . . . . . . . . . . . . 60
4.5.7 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
4.6 Bond management . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
4.6.1 Macaulay Duration . . . . . . . . . . . . . . . . . . . . . . . 62
4.6.2 Active bond management . . . . . . . . . . . . . . . . . . . 64
4.6.3 Duration for general term structures . . . . . . . . . . . . . 65
4.6.4 Passive bond management: Immunization . . . . . . . . . . 66
4.6.5 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
Contents 4
5 The efficient market hypothesis 71

5.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
5.2 Random walks and the EMH . . . . . . . . . . . . . . . . . . . . . 71
5.3 Implications of the EMH . . . . . . . . . . . . . . . . . . . . . . . . 72
5.4 Are markets efficient? . . . . . . . . . . . . . . . . . . . . . . . . . 73
5.5 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
6 Futures markets 75
6.1 Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
6.1.1 Forward contract . . . . . . . . . . . . . . . . . . . . . . . . 75
6.1.2 Futures contract . . . . . . . . . . . . . . . . . . . . . . . . 77
6.1.3 Differences between Futures and Forwards . . . . . . . . . . 78
6.2 Trading strategies . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
6.2.1 Speculation and leverage . . . . . . . . . . . . . . . . . . . . 80
6.2.2 Hedging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
6.2.3 Spread trading . . . . . . . . . . . . . . . . . . . . . . . . . 83
6.3 Single-stock futures . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
6.4 Futures prices of stock indices . . . . . . . . . . . . . . . . . . . . . 86
6.5 Investment strategies with stock-index futures . . . . . . . . . . . . 89
6.5.1 Creating synthetic stock positions . . . . . . . . . . . . . . 89
6.5.2 Hedging an equity portfolio from market risk . . . . . . . . 91
6.6 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
7 Swaps 94
7.1 Interest Rate Swaps . . . . . . . . . . . . . . . . . . . . . . . . . . 94
7.2 Hedging with an IRS . . . . . . . . . . . . . . . . . . . . . . . . . . 95
8 Solutions to Problems 97
Bibliography 98
Chapter 1
Portfolio Choice
1.1 Risk and return
Based on chapter 5 of Bodie, Kane, and Marcus (2014)
1.1.1 Holding period return
Definition 1.1.1: Holding Period Return
The Holding Period Return (HPR) on an investment from today (time 0)

until T years from now is
P (T ) − P (0) + Payouts
HPR = (1.1)
P (0)
where
• P (t) is the price of the asset at time t
• “Payouts” represent cash income from the asset between 0 and T

(dividends for stocks, coupons for bonds), assumed to be paid at the
end of the holding period.
Instead of HPR, will usually just say “rate of return” or simply “return”, and
denote it by r.
5
1.1. Risk and return 6
1.1.2 Expected return and standard deviation
True population moments
We want to characterize the probability distribution of returns across future states

of nature (“scenarios”).
Definition 1.1.2: Mean and Variance

For the random rate of return r,
S
X
µ := E[r] = p(s)r(s)
s=1
XS
σ 2 := Var[r] = p(s)[r(s) − µ]2
s=1
where
• s = 1, . . . , S are the possible states of nature (scenarios)
• p(s) is the probability of state s occuring.

p
The standard-deviation is σ = Var[r].
Example 1.1.1. Stock X is trading at $10. You estimate the follow-

ing scenarios for next year:
State of Market Prob Year-end price Dividends

Expansion 0.6 $ 13 $1
Recession 0.4 $8 $0
Compute the mean and standard-deviation of returns.
Time-series estimation
The “true” moments are not observable, so we have to estimate the inputs to our
models
1.1. Risk and return 7
Estimation of mean and variance

Using a sample of T past observations or realized returns,
T
1X
µ̂ = rt
T
t=1
T
1 X
σ̂ 2 = [rt − µ̂]2
T −1
t=1
1.1.3 Risk premium and Sharpe Ratio
Terminology for returns in excess of the risk-free rate:
1. Excess return is the difference in any particular period between the actual
rate of return on a risky asset and the actual risk-free rate.
2. Risk premium is the difference between the expected HPR on a risky asset
and the risk-free rate
The Sharpe Ratio, or Reward-to-Volatility Ratio, is an important measure of

the trade-off between reward and risk.
Definition 1.1.3: Sharpe Ratio
Given the forward-looking true population moments and a constant risk-free

rate, the Sharpe Ratio (SR) is
Risk premium
SR =
Std-dev of return
To compare historical returns on different portfolios, given that the risk-free

rate also varies through time, we use the following alternative:
f = Average excess return

SR
Std-dev of excess return
Example 1.1.2. Continuing the previous example for stock X, sup-

pose that the 1-year risk free rate is 3%. Compute the Risk Premium
and Sharpe Ratio.
Answer: SR = 0.4423
1.2. Risk Aversion and Capital Allocation 8
1.1.4 Exercises
Ex. 1.1 — Bodie, Kane, and Marcus (2014) problems at the end of chapter 5:
CFA problems 1 through 5 (same CFA problems in 2018 and 2021 editions).
1.2 Risk Aversion and Capital Allocation

1.2.1 Risk Aversion
Definition
A fair game is a gamble with an expected payoff of zero. Example: toss a coin
and double or loose your monthly salary.
Definition 1.2.1: Risk aversion

An investor is risk averse if he wishes to avoid a fair game.
Depending on whether you like a fair game or not, you are either:
- Risk Averse
- Risk Neutral
- Risk Lover
We assume that most investors are risk averse:
• They only take risk if expecting a positive risk premium.

• Market prices show that most investors are risk averse
Modelling risk aversion
Different risky portfolios can be ranked with an expected utility function.
Expected utility function
If the investor has mean-variance preferences, the expected utility function

is:
1
U (r) = E[r] − × A × Var[r] (1.2)
2
Remarks:
1. The parameter A characterizes the investor

• A > 0 means risk averse
• A = 0 means risk neutral
• A < 0 means risk lover (these investors are usually at the casino. . . )
2. This utility score U can be interpreted as a certainty equivalent rate of

return.
• The investor is indifferent between a given risky investment that gen-
erates a utility U and a risk-free investment at rate U
Example 1.2.1. (This is concept check 6.2 in BKM). A portfolio has

an expected rate of return of 20% and standard deviation of 30%.
T-bills offer a safe rate of return of 7%.
1. Would an investor with risk-aversion parameter A = 4 prefer to
invest in T-bills or the risky portfolio?
2. What if A = 2?
Mean-Variance Dominance
We say that portfolio x mean-variance dominates portfolio y if all risk-averse

investors with mean-variance preferences prefer x to y, regardless of their particular
value of A (their degree of risk aversion).
Proposition 1.2.1: Mean-Variance Dominance
Asset x mean-variance dominates asset y iff:
µx ≥ µy and σx < σy
or µx > µy and σx ≤ σy
p
where µi = E[ri ] and σi = Var[ri ]
E[r]
✻
✲
σ
1.2.2 Portfolio rate of return
Definition
Definition 1.2.2: Rate of return on a portfolio
The return on a portfolio is:

I
X
rp = wi ri (1.3)
i=1
where
• I = number of securities in the portfolio
• wi = proportion of funds invested in security i
• ri = return on security i
Example 1.2.2. Consider the following portfolio:
Stock A Stock B
Initial Investment $40,000 $60,000
P0 $20 $10
Initial Number of shares ... ...
P1 $24 $11
Compute the portfolio’s terminal total value:

V1 = . . .
This implies a portfolio return of 114 000/100, 000 − 1 = 0.14.

Now check that (1.3) gives the same number:
rp = wa ra + wb rb = . . . = 0.14
Mean and variance
Proposition 1.2.2: Portfolio mean
For a portfolio with 2 assets (I=2):

" 2 #
X
µp := E[rp ] = E wi ri = w1 E[r1 ] + w2 E[r2 ]
i=1
and
Proposition 1.2.3: Portfolio variance
For a portfolio with 2 assets (I=2):

" 2 #
X
σp2 := V ar[rp ] = Var wi ri = w12 σ12 + w22 σ22 + 2w1 w2 Cov(r1 , r2 )
i=1
Remarks about covariance:
1. By definition, Cov(r1 , r2 ) = E[(r1 − E[r1 ])(r2 − E[r2 ])].
2. We sometimes denote σ12 := Cov(r1 , r2 ).
3. The linear correlation coefficient between r1 and r2 is a more intuitive mea-

sure and is defined as
Cov(r1 , r2 )
ρ :=
σ1 σ2
Always have −1 ≤ ρ ≤ +1.
Example 1.2.3. Consider two stocks with the following parameters:
Stock µ σ
X 10% 20%
Y 20% 40%
The correlation is 0.1. We form the following portfolio: wx = 0.4, wy =

0.6.
Check that the portfolio mean is 0.16 and the standard-deviation is
0.2605.
1.2.3 Portfolios of one risky asset and a risk-free asset
Mean and variance
Consider a complete or combined portfolio c composed of two assets:
• A risk-free asset, such as a T-Bill (denote it by f ).

• A risky asset, which may itself be a portfolio of risky assets (denote it by p).
The risk-free asset is not random and thus has Var[rf ] = 0 and σf p = 0.
Therefore,
E[rc ] = wf rf + wp E[rp ]
V ar[rc ] = wp2 σp2 ⇒ σc = wp σp
where
rc = return on the complete portfolio.
Capital allocation line
Proposition 1.2.4: Capital allocation line
All possible combinations of the risk-free asset and the risky portfolio p lie
on the following straight line:
E[rp ] − rf
CAL(p): E[rc ] = rf + σc
σp
Proof. Using the previous equations,

E[rc ] = wf rf + wp E[rp ]
⇒E[rc ] = (1 − wp )rf + wp E[rp ]
σc σc
⇒E[rc ] = (1 − )rf + E[rp ]
σp σp
E[rp ] − rf
⇒E[rc ] = rf + σc
σp
Example 1.2.4. Consider rf = 7% and a portfolio with E[rp ] = 15%

and σp = 22%. All possible combinations of these two assets plot
along the Capital Allocation Line:
E[r]
✻
✲
σ
For example, plot the combined portfolio for wp = 0.75.

Plot also a leveraged combined portfolio with wp = 1.5. This requires
wf = −0.5 (what does this mean?).
E[r ]−r
p f
The slope of the CAL is σp . This slope is called “reward-to-variability”
ratio. It is also called the Sharpe ratio. The higher this ratio, the better.
1.2.4 Risk tolerance and asset allocation
Proposition 1.2.5: Best complete portfolio
For an investor with mean-variance preferences, the best combination of the

risk-free asset and the risky portfolio p is:
E[rp ] − rf
wp∗ = and wf∗ = 1 − wp∗ (1.4)
Aσp2
Proof. The investor gets the following utility from a portfolio on CAL(p):
U (rc ) = E[rc ] − 0.5 × A × Var[rc ]

= wf rf + wp E[rp ] − 0.5 × A × wp2 σp2
The investor picks the particular portfolio on the CAL that maximizes his utility
function.
maximize (1 − wp )rf + wp E[rp ] − 0.5 × A × wp2 σp2
wp
The first-order condition is:

− rf + E[rp ] − A × σp2 × wp = 0
E[rp ] − rf
⇒wp∗ =
Aσp2
Example 1.2.5. (Example 6.4 in BKM) Continuing the previous ex-

ample, check that the optimal portfolio for an investor with A = 4 is
wp∗ = 0.41. Plot the optimal portfolio on the CAL.
E[r]
✻
✲
σ
1.2.5 Exercises
2, 13–19. Same problems in 2018 and 2021 editions.
Ex. 1.3 — The return on stock x over the next year will depend on the state of
the market as follows:
State of Market Prob rx
Expansion 0.6 0.3
Recession 0.4 -0.1
The risk-free rate is 2%.
What is the Sharpe Ratio of a portfolio that invests 60% in stock x and 40% in
the risk-free asset?
1.3. Optimal Risky Portfolios 15
1.3 Optimal Risky Portfolios
1.3.1 Diversification and portfolio risk
The main idea is “Don’t put all your eggs in one basket”. Different stocks respond
differently to economic shocks. Diversifying your funds into several assets reduces
the risk of the overall portfolio.
The total risk can be decomposed into two parts:
1. The part of the risk that can be easily eliminated through diversification.
This is called:
(a) unique risk or firm-specific risk;
(b) nonsystematic risk; or
(c) diversifiable risk
2. The part of the risk that cannot be eliminated and remains even after di-
versifying. This is called:
(a) market risk;
(b) systematic risk; or
(c) nondiversifiable risk
Figure 1.1: Portfolio diversification (fig 7.2 in BKM)

1.3.2 Portfolios of two risky assets
Mean and variance
Suppose there are just two risky assets (stocks). Recall that the mean and variance
of the return on the portfolio formed by these two assets is:
E[rp ] = w1 E[r1 ] + w2 E[r2 ]

V ar[rp ] = w12 σ12 + w22 σ22 + 2w1 w2 σ1 σ2 ρ
where ρ is the correlation coefficient between r1 and r2 (−1 ≤ ρ ≤ +1).
Correlation effects
The investment opportunity set depends critically on the correlation between

stocks. The smaller the correlation coefficient, the greater the benefits from diver-
sification.
Perfect positive correlation (ρ = 1). There is no gain from diversification

since the assets are essentially identical (the return on one asset is a linear function
of the other). The portfolio standard-deviation is equal to the weighted average
of the two standard-deviations
σp = w1 σ1 + (1 − w1 )σ2
which means that all possible portfolios lie on the straight line between the two
assets (in σ, µ - space).
Imperfect correlation (−1 < ρ < +1). Now we have the diversification
benefit. At each level of µp , the corresponding σp is less than in the ρ = 1 case.
This is because σp2 increases in ρ (∂σp2 /∂ρ = 2w1 w2 σ1 σ2 > 0).
Whereas the expected return on the portfolio is always the weighted average of
expected returns on the individual assets, the standard-deviation of the portfolio
is now less than the weighted average of the individual standard-deviations.
Note that only the portfolios on the upper part of the curve are efficient, that
is, they (mean-variance) dominate the ones on the lower part of the curve.
Perfect negative correlation (ρ = −1). For this (theoretical) case we

would be able to construct a risk-free asset.
Plot all these cases:
E[r]
✻
✲
σ
Minimum variance portfolio
The Minimum-Variance Portfolio (MVP) has the smallest possible risk.
Proposition 1.3.1: Minimum variance portfolio
Given two risky assets, the MVP is given by
σ22 − σ12
w1 = and w2 = 1 − w1
σ12 + σ22 − 2σ12
Proof. The problem is :
minimize σp2
w1 ,w2
s.t. w1 + w2 = 1
or
minimize w12 σ12 + (1 − w1 )2 σ22 + 2w1 (1 − w1 )σ12
w1
The foc for w1 is:
2w1 σ12 − 2(1 − w1 )σ22 + 2(1 − w1 )σ12 − 2w1 σ12 = 0

σ22 − σ12
⇒w1 =
σ12 + σ22 − 2σ12
Example 1.3.1. Consider two stocks:
Security µ σ
1 0.10 0.20
2 0.20 0.30
with ρ12 = 0.1.

Check that the return on MVP has a mean of 12.88% and a standard
deviation of 17.38%. Note that MVP standard deviation is smaller
than both individual standard deviations.
Example 1.3.2. Continuing the previous example, assume instead a

negative correlation of ρ12 = −0.3. Check that the return on MVP
has a mean close to the previous case (13.49%), but a much smaller
standard deviation of σM V P = 14.05%.
1.3.3 Portfolios of two risky assets and a risk-free asset
Tangency portfolio
The optimal risky portfolio to combine with a risk-free asset is the one that pro-
duces the steepest Capital Allocation Line. It is called the Tangency portfolio.
E[r]
✻
✲
σ
Proposition 1.3.2: Tangency portfolio
Given two risky assets and a risk-free rate rf , the tangency portfolio is
r1e σ22 − r2e σ12

w1 = and w2 = 1 − w1
r1e σ22 + r2e σ12 − (r1e + r2e )σ12
with rje = E[rj ] − rf , for j = 1, 2.
Proof. We want to find the portfolio that maximizes the slope of the CAL going
through it:
E[rp ] − rf
maximize
w1 ,w2 σp
s.t. E[rp ] = w1 E[r1 ] + w2 E[r2 ]
σp = [w12 σ12 + w22 σ22 + 2w1 w2 σ12 ]1/2
w1 + w2 = 1
Substitute w1 = 1 − w2 and solve the foc.
Example 1.3.3. Continuing the previous example with ρ = 0.1, fur-

ther assume rf = 5%. Check that the Tangency portfolio is (w1 =
0.3871, w2 = 0.6129). This portfolio has E[rp ] = 0.1613 and σp =
0.2065.
Optimal complete portfolio
Once we have the Tangency portfolio, we can find the optimal complete portfo-
lio that combines T and the risk-free asset. For an investor with mean-variance
preferences (equation 1.2), the optimal solution is given by (1.4):
E[rT ] − rf
wT∗ =
AσT2
Example 1.3.4. Continuing the previous example, assume A = 5.

Check that the optimal combined portfolio is
(wT = 0.5219, wf = 0.4781)
or equivalently,
(w1 = 0.2020, w2 = 0.3199, wf = 0.4781)
The combined portfolio has E[rc ] = 0.1081 and σc = 0.1078

Plot all these portfolios:
E[r]
✻
✲
σ
1.3.4 The Markowitz portfolio optimization model
The previous concepts can be extended to the case of 1 risk-free asset and N risky
stocks. This is called the Markowitz portfolio model.
Extension to N risky assets. Intuitively, the analysis can be generalized

to 3 risky assets by taking one of the possible two-asset portfolios and a new 3rd
asset. Proceeding with these iterations, we could get to N risky assets.
Extension to N risky assets plus 1 risk-free asset. The investor will

pick one particular portfolio on the mean-variance frontier — the tangency portfo-
lio — to combine with the risk-free asset. The straight line going through rf and
µT is the efficient frontier.
E[r]
✻
✲
σ
1.3.5 Exercises
4–10. Same problems in 2018 and 2021 editions.
Ex. 1.5 — There are two funds available for investment:

Fund E[r] σ2
Bond fund 0.04 (0.1)2
Equity fund 0.08 (0.2)2
The correlation between the two funds is 0.3. Additionally, it is possible to invest
in a risk-free asset with rf = 0.02.
An investor is willing to tolerate a maximum standard-deviation of 0.15. What is
the expected return on the best portfolio for this investor?
Ex. 1.6 — Two stocks have the following means and variances:
Stock E[r] σ2
a 0.08 (0.2)2
b 0.12 (0.4)2
The correlation between the two stocks is zero. There is no risk-free asset.
Investors have mean-variance preferences given by U (r) = E[r] − A2 Var[r].
Find the optimal portfolio of the two stocks for an investor with A = 3, that is,
indicate the optimal weights wa and wb .
Chapter 2
The Capital Asset Pricing

Model
(Based on chapter 9 of Bodie, Kane, and Marcus (2014))
The value of any asset is the present value, or discounted value, of its future
cash flows. The CAPM gives us a formula for the discount rate. Hence, it is used
everyday by corporations and investors to price investment projects, stocks, mutual
funds, etc. The CAPM was developed almost simultaneously in three papers by
Sharpe in 1964, Lintner in 1965, and Mossin in 1966.
2.1 Assumptions and derivation
Assumptions:
1. All investors are mean-variance optimizers, i.e., they all use the Markowitz
model.
2. Investors have homogeneous expectations.
3. Investors can borrow or lend at a common (exogenous) risk-free rate.
4. All assets are publicly traded and short positions are allowed.
5. The investors’ planning horizon is a single period.
6. There are no taxes.
7. There are no transaction costs.
22
2.1. Assumptions and derivation 23
Derivation:
1. Since all investors estimate the same inputs (means and covariances), the
tangency portfolio is the same for every investor.
2. The efficient frontier (namely, the straight line through rf and T ) is the
same for every investor.
3. Two fund separation:

(a) Every investor allocates his wealth between two portfolios: the risk-free
asset and the Tangency portfolio.
(b) Note that the weights in the risk-free asset and in the tangency portfolio
may differ across investors due to different degrees of risk aversion, but
still everybody invests in just those two assets and in no other portfolio.
4. In equilibrium, all risky assets must belong to T .

T
To see this, suppose that IBM is not in T (wIBM = 0). Then, there would
i T
be no demand for this stock, (wIBM = wIBM = 0, for every investor i). We
would thus have Demand 6= Supply, which is not equilibrium. Therefore, in
equilibrium, wjT > 0, ∀ asset j.
5. Furthermore, for every asset, the weight in T must be the same as in the
whole market:
Market Capj
wjT = P =: wjM , ∀ asset j
n Market Capn
If we all put 2% of our risky money into IBM stock, then IBM will have
2% of all money invested in the stock market, meaning that the market
capitalization of IBM will be worth 2% of the whole market capitalization.
Note that different investors may put different amounts of money at risk, ie,
in the tangency portfolio. But from these amounts, each investor allocates
the same 2% to IBM.
6. Hence, the Market portfolio is the Tangency portfolio, M = T . This is the

economic content of the CAPM.
CAPM
The CAPM states that the Market portfolio is mean-variance efficient, that
is,
Market portfolio = Tangency portfolio
2.2. Efficient frontier 24
2.2 Efficient frontier
2.2.1 Capital Market Line
When we use M instead of T, the efficient frontier is called Capital Market Line:
E[r]
✻
✲
σ
Proposition 2.2.1: Capital Market Line
Under the CAPM, the efficient frontier is
E[rM ] − rf
CM L : E[rp ] = rf + σp
σM
where p is an efficient portfolio.
Recall that any p ∈ CML is a combination of the risk-free and the market
portfolio, thus σp = wM σM .
Example 2.2.1. You expect the stock market to go up by 10% over

the next year. The standard deviation of the market return is 20%.
You can buy 1 year government bonds yielding 4%. You have $100,000
to invest and you are willing to tolerate a risk (standard deviation) of
15%.
1. What is the best allocation of your money?
2.2. Efficient frontier 25
2. How much money do you expect to have one year from now?
(Answer: $108,500)
2.2.2 Risk aversion and the market risk premium
Proposition 2.2.2: Risk premium of the market portfolio
2
E[rM ] − rf = ĀσM
where Ā is the risk-aversion of the “representative” investor.
Proof. Under the CAPM, the best complete portfolio for an investor with mean-
variance preferences must be a CML portfolio. More precisely, for an investor with
risk-aversion parameter A, the optimal portfolio is determined as in equation (1.4):
E[rM ] − rf
wM = 2 and wf = 1 − wM
AσM
In the aggregate, the amount of borrowing and lending among investors has to net
out to zero. Hence, the “representative” investor will have wf = 0 ⇒ wM = 1.
E[rM ]−rf
Denoting by Ā the risk aversion of the representative investor, 1 = Āσ 2 .
M
Example 2.2.2. (Concept check 9.2 in BKM) Data from the last 8
decades for the S&P500 index yield the following stats: average excess
return, 7.9%; standard-deviation, 23.2%.
To the extent that these averages approximated investor expectations
for the period, what must have been the average coefficient of risk
aversion?
2.3. Expected returns on individual securities 26
2.3 Expected returns on individual securities
2.3.1 Security Market Line
Proposition 2.3.1: Security Market Line
Under the CAPM, the equilibrium expected return for any asset j is given
by
SM L : E[rj ] = rf + βj ( E[rM ] − rf )
where
Cov(rj , rM )
βj :=
Var[rM ]
Different stocks have different betas and thus different expected returns:
E[r]
✻
✲
β
Important remarks about the SML:
• The SML applies to every single asset or portfolio (not necessarily on the
CML).
• For any asset or portfolio j, the relevant measure of risk is βj , not its vari-
ance.1
Example 2.3.1. Industry-type application of the CAPM. Suppose E[rM ] =

10% and rf = 4%. You estimate stock a will pay a dividend of $2 one
year from now. After that, you expect dividends to grow at 5% per
1
But for an efficient portfolio on the CML, including the market portfolio, we can use
either the SML or CML to get its expected return.
year. You also estimate the beta of the stock to be βa = 0.9. What is
the equilibrium price of the stock?
Note: recall that the present value of a stream of dividends growing
D1
at rate g is P0 = r−g , where r is the discount rate. Thus, you just
need to use the CAPM to estimate the required discount for stock a.
Answer: Pa = $45
2.3.2 Interpretation of beta
The SML implies, mechanically,
βa > βb ⇒ E[ra ] > E[rb ]
But what is the economic intuition?
Why do betas explain differences in expected returns?
1. What matters in βj is Cov(rj , rM ) (the market variance is the same for all
stocks)
2. Hence, high beta means high covariance with the market.

2 is proportional to Cov(r , r ).
3. The contribution of an individual security j to σM j M
Intuitively, high Cov(rj , rM ) means that j wins/looses in the same states
as the market wins/looses, which increases the variance of the market. See
BKM for a formal proof.
4. Since all investors hold the market portfolio (in some combination with the
risk-free rate), investors do not like stocks with high betas (why?)
5. Therefore, investors require high expected returns to hold high-beta stocks
Example 2.3.2. Consider the following two companies:

• Duke Energy Corporation (DUK): an energy company based in
North Carolina, with a significant part of its operations in regu-
lated markets.
• Ralph Lauren Corporation (RL): sells lifestyle products (cloth-
ing, accessories, fragrances).
Which one is likely to have a higher beta? Why?
On 2015-06-24, βDU K = 0.28, βRL = 0.93
2.3.3 Beta of a portfolio
Proposition 2.3.2: Portfolio β
For a portfolio p with N securities,

N
X
βp = wi βi
i=1
where wi are the portfolio weights.
Proof. From the definition of beta,
XN
βp := Cov(rp , rM )/ Var(rM ) = Cov( wi ri , rM )/ Var(rM )
i=1
N
X XN
= wi Cov(ri , rM )/ Var(rM ) = wi βi
i=1 i=1
Example 2.3.3. (This is concept check 9.3 in BKM). Suppose that

the risk premium on the market portfolio is estimated at 8% with
a std of 22%. What is the risk premium on a portfolio p invested
25% in Toyota and 75% in Ford, if they have betas of 1.10 and 1.25,
respectively?
Note that p in the previous example is not efficient: it will surely have too
much σp2 . However, the CAPM does not reward all of σp2 ; it only rewards the part
that is nondiversifiable, which depends on Cov(rp , rM ) or βp .
2.4. Exercises 29
Example 2.3.4. Continuing the previous example:

1. Construct an efficient portfolio q on the CML with βq = 1.2125.
2. Since the SML applies to any portfolio and the betas are the same,
the risk premium on q is the same as p:2
E[rq ] − rf = . . .
This means that E[rq ] = E[rp ].

3. Plot the CML and the approximate location of p and q
2.4 Exercises
3–7, 17, 21. Same problem numbers in the 2018 and 2021 editions.
2
Alternatively, to compute the risk premium on q without using the SML, Note that
rq − rf = wM rM + wf rf − rf = wM rM + wf rf − (wM + wf )rf = wM (rM − rf ).
Chapter 3
Arbitrage Pricing Theory and

Factor Models
Based (somewhat) on chapters 10 and 24 of Bodie, Kane, and Marcus (2014), with
additional materials from other sources.
3.1 Theory
The APT was developed by Ross (1976).
1. Factor model. The APT starts by assuming that stock returns are gener-
ated by K factors:
K
X
rj = aj + βjk Fk + εj , j = 1, 2, . . . , N (3.1)
k=1
with the following assumptions:

A1: E[εj ] = 0, ∀j
A2: Cov(Fk , εj ) = 0, ∀k, j
A3: Cov(εj , εi ) = 0, ∀j 6= i
The point is to find a small number of factors (K << N ) that satisfy this
model.1
1
Equation (3.1) is sometimes stated as deviations from means. Take expectations on
PK
both sides to get E[rj ] = aj + k=1 βjk E[Fk ]. Plug the resulting value for aj into (3.1)
30
3.1. Theory 31
2. Diversification. Under (3.1), the return on a portfolio p with N stocks is

K
X
rp = ap + βpk Fk + εp
k=1
and its risk is

" K
#
X
Var[rp ] = Var βpk Fk + Var[εp ]
| {z } | {z }
k=1
total risk | {z } nonsystematic risk
systematic risk
It can be shown that the second term goes to zero in a large, well-diversified
portfolio (wj = 1/N ):
N →∞
Var[εp ] −−−−→ 0
Hence, εp ≡ 0 in a very large portfolio (in the limit), and
K
X
rp = ap + βpk Fk
k=1
3. No arbitrage. If we can trade the factors to hedge the random part of rp ,

then it can be shown that there are no arbitrage opportunities only if
K
!
X
ap = 1 − βpk rf
k=1
Then, the expected return on a well-diversified portfolio must be:

K
X
E[rp ] = rf + βpk ( E[Fk ] − rf ) (3.2)
k=1
when Fk are traded portfolios.

For an individual security (which is not a well-diversified portfolio), we only
get an approximation.
to get
K
X
rj = E[rj ] + βjk F̂k + εj
k=1
with F̂k := Fk − E[Fk ] and thus E[F̂k ] = 0. Stock returns deviate from their means as
a result of unexpected realizations of risk factors. We can further subtract rf from both
sides to get
XK
rj − rf = E[rj − rf ] + βjk F̂k + εj
k=1
which is the version presented in BKM. Note that this is just a mathematical manipulation
of (3.1); it is still not saying anything about what variables explain E[rj ].
3.1. Theory 32
In summary,
Definition 3.1.1: Arbitrage Pricing Theory (APT)
If there are no arbitrage opportunities, the expected return on an individual

security j is
K
X
E[rj ] ≈ rf + βjk FRPk (3.3)
k=1
where:
• FRPk , the k-th Factor Risk Premium, is:
– If factor k is a traded positive-investment portfolio,

FRPk = E[Fk ] − rf = E[Fk − rf ]
– If factor k is a traded zero-investment long-short portfolio,
FRPk = E[rlong − rshort ]
– If factor k is not a traded portfolio, FRPk is a free parameter to
be estimated.
• βjk is the loading of asset j on factor k. In practice, the loadings are

estimated through a time-series regression of excess returns for asset
j on the factors.
Remarks:
• In practical applications we typically ignore the nuances and replace the ≈

with =
• The different alternatives for the FRP is advanced material not covered in
BKM, you will not be tested on this, and we will only apply the traded-
factors version (Fama-French model below).2
For empirical applications with traded factors, the following modification of

(3.2) is more useful:
2
In any case, the intuition is P
the following. The general statement of APT, for any type
K
of factor, is just that E[rpe ] = k=1 βpk F RPk , with F RPk 6= 0, where rpe is the excess
return on a well-diversified portfolio. But for traded factors, we can be more specific. If,
for example, factor 1 is traded, we can put F1 on the LHS of the equations above. The
factor model would obviously have β11 = 1 and β1k = 0, k = 2, . . . , K. Then, the pricing
equation with F1 on the LHS would give E[F1e ] = F RP1 , thus pinning down the FRP for
factor 1.
3.2. Example with 1 factor: the Market model 33
Proposition 3.1.1: Alternative “alpha statement” of the APT
The risk premium on a well-diversified portfolio p is

K
X
E[rp − rf ] = αp + βpk FRPk (3.4)
k=1
If all factors are traded and there are no arbitrage opportunities, we must
have
αp = 0
Advantages and disadvantages of the APT:
• The main advantage of the APT is that it can accommodate several sources
of systematic risk.
Natural macroeconomic factor candidates are: interest rates, GDP growth,
inflation, energy prices, etc.
• The main disadvantage is that the theory does not specify which are the
“true” factors.
It has been and empirical quest, with the current number of factors in the
literature at over 300 and counting... A horse that has won many races is
the FF3.
3.2 Example with 1 factor: the Market model
The Market Model states that there is just one factor: the market. The return
generating process is:
rj = aj + βj rM + εj , ∀j (3.5)
If there are no arbitrage opportunities, expected returns are given by:
E[rj ] = rf + βj ( E[rM ] − rf ) , ∀j
which is the twin brother of the CAPM’s SML.

3.3. Example with 3 factors: the Fama-French model 34
3.3 Example with 3 factors: the Fama-French

model
3.3.1 Pricing equation
Fama and French (1993) propose the following 3-factor asset pricing model:
Definition 3.3.1: Fama and French 3-factor model

The expected return on stock j is
E[rj ] − rf = βjM ( E[rM ] − rf ) + βjs E[SM B] + βjh E[HM L] (3.6)
where the loadings (βjM , βjs , βjh ) are the slopes in the time-series regression
(rj − rf )t = aj + βjM (rM − rf )t + βjs SM Bt + βjh HM Lt + εjt (3.7)
3.3.2 Details on the factors
To form the two new factors, FF divide all firms into six buckets depending on their
size (market equity, ME) and the ratio of book equity to market equity (BE/ME):3
50th ME prct
Small Value Big Value > 70th BE/ME prct
Small Neutral Big Neutral
Small Growth Big Growth < 30th BE/ME prct
“Small” stocks have ME smaller than the median ME. Typically, small stocks
perform better than what the CAPM predicts (this is a so called anomaly).
“Value” stocks have BE/ME higher than the 70th BE/ME percentile; their
book-to-market ratio is High. “Growth” stocks have BE/ME lower than the 30th
BE/ME percentile; their book-to-market ratio is Low. Typically, BE/ME is high
when the ME (denominator) is low. This happens when the firm has had low
returns and is now near financial distress. Nonetheless, most of these firms usually
rebound and thus, if you hold a large portfolio of these firms, you end up making
more money than their CAPM beta would suggest (another CAPM anomaly).
Each month, the factors are computed in the following way:

3
See the details at http://mba.tuck.dartmouth.edu/pages/faculty/ken.french
3.4. Portfolio performance evaluation 35
• SMB (Small Minus Big) is the average return on the three small portfolios
minus the average return on the three big portfolios,
SMB = 1/3 (Small Value + Small Neutral + Small Growth)
- 1/3 (Big Value + Big Neutral + Big Growth)
Historically, the SMB portfolio generated an annual return somewhere be-
tween 1.5% and 3%. This is the size premium.
• HML (High Minus Low) is the average return on the two value portfolios
minus the average return on the two growth portfolios,
HML = 1/2 (Small Value + Big Value) - 1/2 (Small Growth + Big Growth)
Historically, the HML portfolio generated an annual return somewhere be-
tween 3.5% and 5%. This is the value premium.
This model has had considerable empirical success in explaining CAPM anoma-
lies (portfolios that don’t plot on the SML) and in capturing the variation in the
cross-section of expected returns. Thus, Fama and French (1996) argue that SMB
and HML mimic combinations of two underlying risk factors of special concern to
investors.
3.4 Portfolio performance evaluation
3.4.1 Motivation
1. One important question in finance is: How to assess the performance of a
fund manager?
2. We cannot just look at raw realized returns because we want to distinguish

stock-picking skills from simple risk taking. If we see a big return, was it
because the manager was able to identify mispriced stocks or was it because
he took large risks and got lucky?
3. Therefore, we need to compute risk adjusted returns, that is, we need to

measure the difference between the empirical realized returns and the returns
“appropriate” for the risk of the fund.
3.4.2 Adjusting returns for risk
The most widely used performance measure is Jensen’s alpha. It is the empirical
counterpart to the APT equation (3.4).
Definition 3.4.1: Jensen’s alpha
Jensen’s α for fund p is the intercept in the time-series regression

K
X
e
(rp − rf )t = αp + βpk Fkt + εpt , t = 1, . . . , T
k=1
e is the excess return on the traded factor k at time t:

where Fkt
• If factor k is a traded positive-investment portfolio, Fkt
e = (F − r )
k f t
• If factor k is a traded zero-investment long-short portfolio, Fkt

e =
(rlong − rshort )t
We now have two models to adjust returns for risk.
CAPM
• To evaluate the performance of fund p, estimate the following time-series

regression:
(rp − rf )t = αp + βp (rM − rf )t + εpt (3.8)
• According to the CAPM, or the 1-factor Market Model APT, we should find
αp = 0.
• If instead we find that αp is (statistically significantly) positive, we can

conclude that the fund returns are higher than what its level of risk would
require (according to the CAPM). In other words, the manager has skill.
• Graphically, a positive Jensen’s alpha implies that the portfolio lies above
the SML:
E[r]
✻
✲
β
Remark. Model (3.8) is the standard regression to estimate the CAPM beta.
BKM call this the “single-index model”. The market model equation (3.5) is
sometimes also used in the industry to estimate the CAPM beta, but this is only
equivalent to (3.8) when the interest rate is constant (which is not the case in
reality).
FF3
• If we don’t believe that CAPM is a good model to adjust returns for risk,
we can use the Fama-French model.
• Run the regression
(rp − rf )t = αp + βpM (rM − rf )t + βps SM Bt + βph HM Lt + εpt
• Again, if αp > 0 (statistically), the manager has skill.
Note that the βpM estimator that comes out of this regression is not the CAPM
beta (due to the presence of other regressors).
Chapter 4
Bond Markets
Figure 4.1: U.S. Treasury Note
38
4.1. Price quotes 39
4.1 Price quotes
4.1.1 Face value and coupons
The cash flows of a bond are defined by its face value (or par value or principal
amount), coupon rate, and maturity date.
Example 4.1.1. If you hold 10 bonds like the one if Figure 4.1, that
is, a total face value of $10,000, what cash flows will you receive?
4.1.2 Bond price units
Bonds used to be priced in dollars.
Example 4.1.2. The bond in Figure 4.1 is priced at $1,100. There-

fore, to buy 10 bonds I need to spend
Investment = price x quantity = ...
In most markets today, however, bond prices are represented as percentages of

face values.
Example 4.1.3. The bond in Figure 4.1 is priced at 110% of face

value. Therefore, to buy 10 bonds I still need to spend
Investment = price x quantity = (110% × 1000 $) × 10 =...
Furthermore, bonds are now dematerialized: there is no paper certificate like

in Figure 4.1, there is just an electronic record. How do we measure “quantity”?
With face value.
Example 4.1.4. To buy a face value of 10 000 $ (“quantity”) of a

bond priced at 110%, I need to spend
Investment = price x quantity = 110% × (10 000 $) =...
4.1. Price quotes 40
Meaning of Price in percentage of Face value
Total Investment ($) = Price (%) × Face Value ($)
The price is unitless — interpret the symbol “%” as meaning “×0.01” to avoid
messing up the units in your calculations. Think of Face Value as a “quantity” of
dollars.
Example 4.1.5. A Bond with 5% annual coupons is priced at 102.1234%.

This means that:
• If we invest a total of 1 M$, we are able to buy the following face
value (“quantity”)1
F V = . . . = $979 207.51
• Our next coupon will be C = . . . = $48 960.38
• At maturity, we will receive a total of F V ∗ 1.05 = $1 028 167.88
Alternative units:
1. Sometimes prices still appear in $, typically for a $100 or $1000 face value.
Eg, B = $102.1234
2. In US markets, some prices are represented as (weird) “dollars and thirty-

seconds of a dollar”.
Eg, B = 96-18 means B = 96 + 18/32 = $96.5625
4.1.3 Quoting conventions
Prices of coupon-paying bonds
Invoice price = Quoted price + Accrued interest
1. The Quoted (or clean, or flat) price quoted by a dealer does not include
interest that accrued since the last coupon date.
2. The Invoice (or dirty, or full, or cash) price is the price that the buyer
actually has to pay the seller.2
1
In practice, trading is usually done in round lots of face value, with a minimum of
1 000 000 $ for Treasuries and 100 000 $ for corporate bonds. We ignore these constraints.
2
In the U.S., the quoted prices are typically clean prices; in Europe, traders may show
the dirty price.
4.2. Rates of return 41
3. The Accrued Interest (AI) is:
N. days since last coupon

AI = × Interest due in full period
N. days between coupons
• The numerator is the number of days from (and including) the last
coupon payment date, up until (but not including) the trade’s settle-
ment date
• The settlement date typically ranges from t + 3 for corporate bonds to
t + 1 for treasuries
• Day-count conventions (may be different outside the U.S.):
– actual / actual(in period) — for T-Notes and T-Bonds.
– 30/360 — for corporate bonds.
– actual/360 — for T-Bills and money markets.
Example 4.1.6. Consider a 15/Jul/2020 4% Bond, paying semi-annual

coupons. The settlement date is 10/April/2012 (note that 2012 is a
leap year).
• With actual/actual day count,
86
AI = × 0.02 = 0.9451% of FV
182
• With 30/360 day count,
85
AI = × 0.02 = 0.9444% of FV
180
In Excel, use the functions COUPDAYBS() and COUPDAYS() to count days.
4.2 Rates of return
Returns are usually expressed in annual terms. There are two alternative rep-
resentations of the same underlying Holding Period Return (HPR, as defined in
equation 1.1).
4.2. Rates of return 42
4.2.1 Annual Percentage Rate
Example 4.2.1. A bank offers a “6-month deposit at 6%”. This

typically means that if you deposit 1000$, at the of 6 months you will
get
CF at 6 months = ...
The Holding Period Return is 3%.
The rate of 6% is an “APR with semiannual compounding (r2 )”.
Definition 4.2.1: Annual Percentage Rate
The Annual Percentage Rate (APR) with compounding frequency n, de-

noted rn , corresponding to an investment with a given HPR over T < 1
years is
rn = HPR × n
where n = 1/T is the number of compounding periods per year
Example 4.2.2. A 1-month T-bill (zero coupon bond) is trading at

99.75% of par value. Check that the APR with monthly compounding
is
r12 = . . . = 3.0075%
Note that complete specification of an APR always needs to state n, unless it

is clear from the security specifications.
4.2.2 Effective Annual Rate
Example 4.2.3. Today, Euribor 6 months is quoted at 2.40%.

A bank borrows 1 M$ for 6 months, knowing that at the end it will
borrow again for another 6 months the whole principal and interest. If
we assume that Euribor remains constant at 2.40%, how much money
will the bank have to pay 1 year from now?
V (1) = V (0) × . . .
The effective rate that the bank will pay in one year is thus 2.41%. In
other words, 2.41% is equivalent to an APR of 2.40% (with semiannual
compounding).
4.3. Bond Prices and Yields 43
Definition 4.2.2: Effective Annual Rate

The Effective or Equivalent Annual Rate (EAR) corresponding to an in-
vestment with a given HPR over T years is
EAR : 1 + EAR = (1 + HPR)n
where n = 1/T is the number of compounding periods per year

99.75% of par value. Check that:
HP R(T = 1/12) = . . . = 0.2506%
and
EAR = . . . = 3.0493%
Note that when n = 1, the APR is also the EAR: r1 = EAR
Example 4.2.5. Today, Euribor 12 months is quoted at 3%.

The market convention is to quote this rate as an APR with annual
compounding, r1 = 3%. Hence, we also have EAR = 3%.
4.3 Bond Prices and Yields
4.3.1 Bond Pricing
Zero-Coupon Bonds (aka zeroes, or pure discount bonds) pay no coupons.
Definition 4.3.1: Price of a Zero-Coupon Bond (ZCB)
The price of a ZCB maturing in T years is
100%
P =
[1 + r(0, T )]T
where r(0, T ) is the discount rate or required return from today (t = 0)

until T years, expressed as an Equivalent Annual Rate (EAR)
Example 4.3.1. The price of a 1-month T-Bill, given that the re-
quired EAR is 3.0493%, is
P = . . . = 99.75%
Proposition 4.3.1: Price of fixed-coupon bond

m
X c/n 100%
P = t
+ (4.1)
[1 + r(0, ti )] i [1 + r(0, tm )]tm
i=1
where
• P is price of the bond today
• n is the number of coupon payments per year
• c is the annual coupon rate with compounding frequency n
• t1 , t2 , . . . , tm = T are the years until the coupon payment dates (with

T denoting the maturity)
• r(0, ti ) is the discount rate or required return from today (t = 0) until

ti years, expressed as an Equivalent Annual Rate (EAR)
Proof. A coupon-bearing bond is a portfolio of zero-coupon bonds. The relation

follows by no-arbitrage.
Example 4.3.2. Compute the price of a 1.5-year 5% T-Bond (recall:

semi-annual coupons) if the required returns are r(0, 0.5) = 3.5%,
r(0, 1) = 4.0%, r(0, 1.5) = 4.4% (all with annual compounding).
P = . . . = 100.95%
4.3.2 Yield to Maturity
Definition 4.3.2: Yield to Maturity
Given the price of a fixed-coupon bond, the YTM (y, as an EAR) is the
constant discount rate that solves
m
X c/n 100%
y: P = t
+ (4.2)
[1 + y] i [1 + y]tm
i=1
Example 4.3.3. A one-month T-Bill sells for 99.67367% (of par value).
Check that the one-month yield is y = 4% (EAR).
Plot Price versus Yield and note the inverse relation.
Remarks:
• In general, we need Excel (YIELD or IRR function, or Solver) or a financial

calculator to compute the YTM of a bond with multiple coupons.
• YTM are sometimes quoted as APR, aka “bond equivalent yields”.
Example 4.3.4. Price = 95%, 10-yr Maturity, Coupon rate = 7%

with semiannual coupons.
The IRR() fn in Excel outputs 3.8635%, which is a 6-month rate.
Then, we would express the YTM in one of the following ways:
• Bond Equivalent Yield (APR).
y2AP R = 3.86% × 2 = 7.72%
• Effective Annual Yield (EAR).
y1EAR = (1.0386)2 − 1 = 7.88%
• Traders also talk about a Current Yield = Annual interest / market price.
For the previous example, y CY = 7/95 = 7.37%.
Interpretation. The YTM will be the realized yield if:
• you hold the bond until maturity; and
• all coupons are reinvested at the ytm.
Example 4.3.5. 2-year 8% bond with semiannual coupons, trading

at a ytm of y2 = 8% and a price of 100%. Check that under the two
conditions specified above, the realized yield at the end of 2 years will
be r2 = 8%.
Remark. From the previous example, note that at the ex-coupon date, Y T M =
Coupon rate ⇒ P = 100%. Check that Y T M > Coupon rate ⇒ P < 100% and
Y T M < Coupon rate ⇒ P > 100%
4.3.3 Realized Yield
The Realized Yield, or Holding-Period Return, or Realized Compound Return de-

pends on:
• Selling price of the bond;
• Reinvestment rate for the coupons.
Example 4.3.6. 2-year 8% bond with semiannual coupons, trading

at a ytm of y2 = 8% and a price of 100%. In six months the ytm falls
to 7%.
1. Check that P6m = 101.4%
2. Check that the realized yield for 6 months is r2 = 10.8016% (APR).
Example 4.3.7. Consider a 2-year, 10% bond with annual coupons,

trading at a ytm of 10%. You hold the bond until maturity, but are
only able to reinvest the coupons at 8%. Check that the realized yield
is 9.9% (EAR).
4.3.4 Default risk and ratings
Corporate bonds trade at an higher yield than risk-free government bonds because
corporations may default. Default risk is measured by a credit rating.
Long Term Obligation Ratings

S&P Moody’s Meaning
AAA Aaa Highest quality, minimal credit risk
AA Aa
A A
BBB Baa Adequate protection, moderate credit risk
BB Ba Speculative, significant risk
B B
CCC Caa
CC Ca
C C Nonpayment highly likely; in default (Moody’s)
D – in Default
4.4. Term structure of interest rates 47
Obligors rated BBB or better are called “investment-grade”; investors rated

BB or worse are called “speculative-grade”. Each category can also be appended
with a + or - sign (S&P) or with 1,2,3 numbers (Moody’s) to show relative standing
within the category.
4.3.5 Exercises
Ex. 4.1 — Bodie, Kane, and Marcus (2014) problems at the end of chapter 14
(p. 480): 6, 9. Same problem numbers in the 2018 and 2021 editions.
4.4 Term structure of interest rates
4.4.1 Spot rates
Definition 4.4.1: Spot rate
The spot rate r(0, T ) is the yield to maturity on a zero-coupon bond with
maturity T .
Definition 4.4.2: Term structure of interest rates

The term structure of interest rates, or zero-coupon yield curve, or simply
yield curve, is the set of spot rates r(0, T ) for different maturities T .
Example 4.4.1. The following Zero-Coupon bonds are trading in the

market:
Time to Maturity Price (%) YTM
1 95.238 5%
2 88.997 6%
3 81.630 7%
Hence, the term structure of interest rates is:
r(0,1) = 5%
r(0,2) = 6%
r(0,3) = 7%.
The spot rates are related to the price of coupon-bearing bonds through (4.1).
Example 4.4.2. Consider a bond with 5% annual coupons and 2

years to maturity. The total value must be the sum of the present val-
ues of all cash flows. Using the spot rates from the previous example:
P = . . . = 98.212%
Remarks:
• The term “yield curve” is sometimes also used to denote the set of YTM on
coupon bonds (which are not the same as zero-coupon yields).
• In addition to the spot rate curve defined above, we can also have forward
rate curves (using the fwd rates defined below)
4.4.2 Forward rates
A forward rate applies to a time period in the future.
Definition 4.4.3: Forward rate

Given a set of spot rates, the forward rate between t1 and t2 (0 < t1 < t2 ),
expressed as an EAR, is given by
f (t1 , t2 ) : [1 + r(0, t1 )]t1 [1 + f (t1 , t2 )](t2 −t1 ) = [1 + r(0, t2 )]t2
Interpretation. The fwd(1,2) is the rate of return on a future investment that

will make me indifferent between the following alternatives:
1. Invest in a 2-year zero coupon bond.
2. Invest in a 1-year zero coupon bond. After 1 year reinvest the proceeds in
another 1-year bond.
Example 4.4.3. Continuing the previous example, check that f (1, 2) =

7.0095% and f (2, 3) = 9.0284%.
4.4.3 Expectations of future interest rates
Term structure under certainty
If all investors know for sure the path of future interest rates (i.e, there is no risk),
then:
1. The forward rate computed today is the value that the spot rate will take
on in the future.
2. All bonds provide the same return over any given holding period.
Example 4.4.4. Consider the following bonds (all have annual coupons):
Bond A B C
Maturity (yrs) 1 2 3
Coupon rate – 4% 6%
Price (%) 90.909 88.045 85.773
1. Compute the spot rates (this procedure is called Bootstrap)

2. Compute the sequence of 1-yr forward rates.
3. Check that the YTM for bond A, B and C are respectively 10%,
10.98%, and 11.91%.
4. Plot the YTM curve, the spot-rate curve, and the forward-rate
curve.
5. Check that if we invest for 1 year, all bonds produce the same
return (recall that under certainty we know the future value of
interest rates).
Uncertainty and expectations
In reality we don’t know the future value of interest rates. What do forward rates
tell us about future spot rates?
Expectations Hypothesis
The forward rate equals the market consensus expectation of the future spot
interest rate. That is,
f orward0 (t1 , t2 ) = E0 [spott1 (t1 , t2 )]

Under EH, an upward sloping spot curve, which is equivalent to increasing

forward rates, means that the market is expecting spot rates to go up in the
future.
However, this simple theory does not fit the data well. Most investors are
risk averse and will only hold long-term bonds if they receive a risk premium (as
discussed in the Duration section below, long-term bonds have higher duration
and are thus riskier than short-term bonds).
Liquidity Preference Hypothesis
To hold long-term zero-coupon bonds, investors require a premium over

short-term (“liquid”) bond yields. This risk premium creates an upward
sloping spot curve. Hence, forward rates exceed expected spot rates by a
(“liquidity”) risk premium:
f orward0 (t1 , t2 ) = E0 [spott1 (t1 , t2 )] + risk premium
Implications:
1. Expectations of increases in future interest rates must always result in a

rising yield curve.
2. However, the converse is not necessarily true. Even if the yield, spot, and
forward curves are all rising, the market may be expecting constant or even
decreasing future interest rates (why? risk premium).
3. In practice, interest rates are hard to predict...
Relation to the business cycle
The slope of the yield curve is typically a good leading indicator of the business
cycle:
• The normal state is an upward sloping curve, which usually means “good
times” ahead
• A flat or downward sloping curve usually forecasts “bad times” ahead, that
is, an economic downturn or recession.
When short-term yields are higher than long-term yields, it suggests that
investors expect interest rates to decline in the future, usually in conjunction
with a slowing economy and lower inflation.
4.5. Yield curve fitting 51
4.4.4 Exercises
(p. 507): 11, 12. Same problem numbers in 2018 and 2021 editions.
4.5 Yield curve fitting
4.5.1 Continuous compounding
Definition 4.5.1: Annual rate with continuous compounding
The annual rate with continuous compounding, denoted r∞ , corresponding

to an investment with a given HPR over T years isa
1
eT ×r∞ = 1 + HP R ⇒ r∞ = ln(1 + HP R) = ln(1 + EAR)
T
a
Recall that limn→∞ (1 + ny )n = ey

99.75% of par value. Check that
r∞ = . . . = 3.0038%
4.5.2 Discount factors and interest rates
Discount factor
Definition 4.5.2: Discount factor

The discount factor Z(t, T ) is the price at time t of $1 to be received at a
later date T .
Example 4.5.2. A 90-day T.Bill is trading at 99.1234%. Hence,

Z(0, 0.25) = 0.991234
Notation:
• Z(.) stands for Zero because the discount factor equals the price of the cor-
responding ZCB. Hence, we will also denote the discount factor by P (t, T ).
• When it is clear that we are computing values as of today, we will simply

write Z(T ) or P (T ).
Equivalence between discount factors and interest rates
The bond pricing formulas in section 4.3 assume nicely spaced coupons and that we
have r1 (.) rates. If that is not the case, we can use the following general formulas.
Equivalent representations for the price of a ZCB
Let P (t, T ) denote the market price of a ZCB. This same price can be
represented in three alternative ways:
• Discount factor:
P (t, T ) = 1 × Z(t, T )
• Interest rate with discrete compounding of n periods per year:

−n×(T −t)
1 rn (t, T )
P (t, T ) = h in×(T −t) = 1+
rn (t,T ) n
1+ n
• Interest rate with continuous compounding:

1
P (t, T ) = = e−r∞ (t,T )×(T −t)
er∞ (t,T )×(T −t)
Example 4.5.3. A 1-month T-bill yields r12 (0, 1/12) = 3.0075%. As-
sume that in the future you will be able to reinvest at the same T-bill
rate that is available now. How much money do you need to invest
today to have $ 100 000 in 3 months?
V (0) = . . . = 99 251.87
Example 4.5.4. The term structure today (t=0) is

Ti (yr) r2 (0, Ti )
0.25 3.00%
0.50 3.50%
0.75 4.00%
1.00 4.20%
Compute the price of 5% Treasury Bond that matures in 9 months.

Recall that Treasury Notes and Bonds pay coupons semi-annually.
The stated coupon rate is always annual with semi-annual compound-
ing.
P (0) = . . . = 101.98%
4.5.3 Motivation
• Spot interest rates are not directly observable in the market (except for
money market rates up to 1 year).
• Instead, what we can observe in the market are bond prices.
• Hence, we need to back out the interest rates that are implied by the observed
market prices.
• We will focus on Treasury Bonds since these are the most liquid. Once we
have the risk-free term structure, we can add a credit spread to get the term
structure for a different credit risk level.
Notation:
• To simplify the notation, we will denote the spot rate r(0, T ) by r(T ) and
the spot discount factor Z(0, T ) by Z(T ).
• In this section, r(.), without subscript, denotes a continuously compounded

rate.
• Recall that the price of a bond that pays a fixed coupon at rate c with
frequency of n periods per year, at dates T1 , . . . , Tm , is
m m
cX c X −r(Ti )Ti
B= Z(Ti ) + Z(Tm ) = e + e−r(Tm )Tm
n n
i=1 i=1
4.5.4 Bootstrap
Method
Start from short-maturity ZCB and work your way up to long-term coupon bonds.
Definition 4.5.3: Bootstrap iterative procedure
1. From the price of a ZCB maturing in T1 , we get Z(T1 ):
P = 1 × Z(T1 )
2. Then, from a second bond with cash flows in T1 and T2 , we get Z(T2 ):
B = c/n × Z(T1 ) + (1 + c/n) × Z(T2 )
3. And so on...
Just like passing the strap through your boots.
What if there are any holes? Do linear interpolation.
Linear interpolation
Given Z(T1 ) and Z(T2 ), the discount factor Z(s), for T1 < s < T2 , can be
obtained through linear interpolation:
Z(T1 ) − Z(s) Z(T1 ) − Z(T2 )

=
s − T1 T2 − T1
Z(T2 ) − Z(T1 )
⇒Z(s) = Z(T1 ) + (s − T1 )
T2 − T1
Remark: it is better to do linear interpolation on discount factors rather than

interest rates because Z(.) is more linear than r(.) — see figure 4.2.
Figure 4.2: Interest rates vs discount factors
Application
Example 4.5.5. Consider the following bond market data (all coupons
are paid annually):
Security Market Price (P) Maturity (years) Coupon rate

ZCB-3m 99.26% 0.25 —
ZCB-6m 98.44% 0.50 —
ZCB-12m 96.39% 1.00 —
CBB-A 107.78% 1.50 7.00%
CBB-B 103.32% 2.00 6.00%
CBB-C 100.19% 4.00 5.00%
1. Using the first ZCB-3m,
0.9926 = 1 × Z(0.25)
2. Similarly for the other two ZCB:

Z(0.5) = 0.9844, Z(1) = 0.9639
3. Using CBB-A and d(0.5),

1.0778 = 0.07 × Z(0.5) + 1.07 × Z(1.5) ⇒ Z(1.5) = 0.9429
4. Using CBB-B and d(1),
. . . Z(2) = 0.9202
5. As an aside, suppose we wanted to know Z(1.25). We can inter-

polate between Z(1) and Z(1.5):
Z(1.25) = . . . = 0.9534
6. Continue to CBB-C:
1.0019 = 0.05 ∗ 0.9639 + 0.05 ∗ 0.9202 + 0.05 ∗ Z(3) + 1.05 ∗ Z(4)
One equation for two unknowns!

(a) Use interpolation to write Z(3) as function of Z(2) = 0.9202
and Z(4):
3−2
Z(3) = 0.9202 + [Z(4) − 0.9202]
4−2
(b) Plug this expression for Z(3) into the previous equation and
solve for Z(4):
The result is Z(4) = 0.8230
Which then implies Z(3) = 0.8716
Summary of bootstrap Z(T ) and corresponding r(T )
T Z(T) r(T)
0.25 0.9926 2.9710%
0.50 0.9844 3.1446%
1.00 0.9639 3.6768%
1.25 0.9534 3.8181%
1.50 0.9429 3.9204%
2.00 0.9202 4.1606%
3.00 0.8716 4.5822%
4.00 0.8230 4.8709%
4.5.5 Nelson-Siegel Model
Method
• Since real bonds have non-standard maturities, it is convenient to model

interest rates as continuous functions of time.
• The model of Nelson and Siegel (1987, Journal of Business) proposes the
following convenient parametric function for the continuously compounded
spot interest rate:
Definition 4.5.4: Nelson-Siegel model
The continuously compounded spot rate for maturity T is

" #
1 − e−T /λ 1 − e−T /λ −T /λ
r(T ) = β1 + β2 + β3 −e
T /λ T /λ
where β1 , β2 , β3 , λ are parameters to be estimated.
To estimate the parameters β1 , β2 , β3 , λ we use a numerical optimizer (e.g.,

Solver in Excel) to solve the following problem.
Nelson-Siegel optimization procedure
I
X
minimize (Bimkt − Bins )2
β1 ,β2 ,β3 ,λ
i=1
where:
• Bimkt is the observed market price of bond i;
c Pm
• Bins is the model value of bond i, ie, Bins = n j=1 e
−r(Tj )Tj +
e−r(Tm )Tm , using the NS rate defined above;
• I is the number of available bonds.
Note that we minimize differences in prices, not yields.3
Meaning of the parameters
The optimization procedure is sensitive to the initial guesses. It is thus important

to understand the meaning of the parameters.
• The instantaneous short rate is:
lim r(T ) = β1 + β2
T →0
3
It seems that there is some online video recommending that you minimize the difference
between NS rates and YTM on coupon bonds. That is wrong because the NS r(T ) rate
is a zero-coupon rate, and thus should be different from the YTM on a coupon bond with
maturity at T .
Hence, β1 + β2 should be relatively close to the yield on the shortest ZCB

(or the overnight rate).
• The long term rate is

lim r(T ) = β1
T →∞
Thus β1 can be interpreted as the level of the curve. It should be relatively

close to the yield on a very long bond.
• Hence, the slope is −β2 .
The following figures illustrate the effect of all parameters.
Figure 4.3: β1 determines the level

β1 = see legend, β2 = 0, β3 = 0, λ = 0
Figure 4.4: β2 determines the slope

β1 = 0.05, β2 = see legend, β3 = 0, λ = 2
Figure 4.5: β3 determines the curvature

β1 = 0.05, β2 = 0, β3 = see legend, λ = 2
Figure 4.6: λ determines the position of the hump

β1 = 0.05, β2 = −0.02, β3 = 0.1, λ = see legend
Application
Example 4.5.6. For the same bonds in the previous example, we get
β1 = 0.0396, β2 = −0.0112, β3 = 0.0624, λ = 4.1219.
Interest rates (cts. comp.)

T Nelson-Siegel Bootstrap
0.25 3.0551% 2.9710%
0.50 3.2544% 3.1446%
1.00 3.6101% 3.6768%
1.25 3.7683% 3.8181%
1.50 3.9147% 3.9204%
2.00 4.1745% 4.1606%
3.00 4.5822% 4.5822%
4.00 4.8707% 4.8709%
4.5.6 Svensson Model
Method
• The model of Svensson (1994, NBER Working Paper) is an extension of

Nelson-Siegel that allows more flexible curves.
• It is used by the European Central Bank. Check out the euro area yield
curve at
http://www.ecb.europa.eu/stats/money/yc/html/index.en.html
Definition 4.5.5: Svensson model

The continuously compounded spot rate for maturity T is
" #
1 − e−T /λ1 1 − e−T /λ1 −T /λ1
r(T ) = β1 + β2 + β3 −e
T /λ1 T /λ1
" #
1 − e−T /λ2
+ β4 − e−T /λ2
T /λ2
where β1 , β2 , β3 , β4 , λ1 , λ2 are parameters to be estimated.
Aplication
Example 4.5.7. For the same bonds in the previous example, we get:
Svensson Nelson-Siegel
β1 0.0612 0.0396
β2 -0.0366 -0.0112
β3 0.1002 0.0624
λ1 10.8919 4.1219
β4 0.0200
λ2 0.7457
The resulting rates (cts. comp.) are in the following table and figure.
T Svensson Nelson-Siegel Bootstrap YTM

0.25 2.8929% 3.0551% 2.9710% 2.9710%
0.50 3.2091% 3.2544% 3.1446% 3.1446%
1.00 3.6427% 3.6101% 3.6768% 3.6768%
1.25 3.7996% 3.7683% 3.8181%
1.50 3.9330% 3.9147% 3.9204% 3.9032%
2.00 4.1576% 4.1745% 4.1606% 4.1467%
3.00 4.5327% 4.5822% 4.5822%
4.00 4.8729% 4.8707% 4.8709% 4.8280%
4.6. Bond management 62
Figure 4.7: Svensson vs Nelson-Siegel vs Bootstrap vs YTM
4.5.7 Conclusion
How to estimate interest rates implied by observed bond market prices?
1. Bootstrap method
• Robust and easy to understand.
• Cumbersome to implement in real-life cases.
2. Nelson-Siegel model
• Easy to implement in real-life cases.
• Numerical optimization is sensitive to initial guesses for the parame-
ters.
3. Svensson model
• More flexible function than NS.
• Even more parameters to fine-tune than NS.
4.6 Bond management
4.6.1 Macaulay Duration
How to measure the sensitivity of the bond price to changes in interest rates, i.e.,
the risk of a bond?
Definition 4.6.1: Macaulay Duration
The Macaulay Duration is

m
1 X Cti
D := ti (4.3)
P (1 + y)ti
i=1
where
• y is the YTM (EAR)
• ti are the coupon payment dates (in years)
• Cti is the total cash flow at date ti : Cti = c/n for i < m, and
Cti = c/n + 1 for i = m, where n is the coupon payment frequency.
Example 4.6.1. Consider a 3-year 5% bond paying annual coupons.

The bond is trading at y = 4%. Check that P = 1.027751 and
D = . . . = 2.8615
Proposition 4.6.1: Price sensitivity to interest rate changes
For a small change in YTM (∆y), the bond price change is well approxi-
mated by
∆P ∼ 1
= −D ∆y (4.4)
P 1+y
Proof. The change in price due to a change in interest rates is ∆P ≡ P (y +

∆y) − P (y). If the change in interest rates is small and instantaneous (i.e., time is
standing still), the change in price can be approximated by taking a Taylor series
expansion of P around y:
dP d2 P
P (y + ∆y) = P (y) + ∆y + 1/2 ∗ (∆y)2 + . . .
dy dy 2
Taking derivatives and dividing through by P, we get
∆P 1 1
= −D ∆y + C(∆y)2 + . . .
P 1+y 2
where D is as defined in (4.3) and C, denoted Convexity, is
Pm Cti
i=1 ti (ti + 1) (1+y)ti 1
C :=
P (1 + y)2
Ignoring term of order higher than ∆y produces the approximation in (4.4).
Meaning of Duration. Bonds with higher Duration are more sensitive to

interest rate changes. Hence, Duration is a measure of the riskiness of a given
bond.
Example 4.6.2. Continuing the previous example, assume that YTM

increases by 20 basis points.
1. The proportional price change for the previous bond is
∆P ∼
P = . . . = −0.5503%
2. Now consider a 3-year ZCB also trading at an initial YTM of
4%. Since
D = ...
the proportional price change in this bond will be
∆P ∼
P = . . . = −0.5769%
Remarks:
• The Modified Duration is D ∗ := D/(1 + y). Using D ∗ instead of D, (4.4)

becomes ∆P ∼
P = −D ∆y.
∗
• From (4.3), we can also think of duration as a weighted average of time, i.e.,
m
X
D= ti × wti
t=i
C ti
where the weights are wti = (1+y) ti /P . Thus, the units of duration are
“years”. In this sense duration measures how fast a bond generates cash
flows. For the same maturity, a bond with higher coupons will have lower
duration. This second interpretation of duration is easier to understand, but
it is nearly worthless for practical applications.
4.6.2 Active bond management
If a manager is able to forecast unanticipated interest rate movements better than

the rest of the market, then equation (4.4) suggests an investment rule:
Active bond management with Duration
• Expect interest rates to decrease ⇒ Increase the portfolio duration
• Expect interest rates to increase ⇒ Decrease the portfolio duration

This rule implies that we readjust the bond portfolio, before interest rates
change, according to the following formula:
Proposition 4.6.2: Duration of a portfolio
The duration of a bond portfolio is:

N
X
Dp = Di wi (4.5)
i=1
where
• N is the number of bonds in the portfolio
• wi = Vi /Vp is the weight in bond i, where Vi is the amount ($) invested

P
in bond i and Vp is the total value of the portfolio, Vp = N i=1 Vi .
Example 4.6.3. A bond portfolio manager has 40 M$ invested in

bond A and 60 M$ invested in bond B. The durations are DA = 7
and DB = 2.
The duration of the portfolio is:
Dp = . . . = 4
The manager expects interest rates to decrease in the coming weeks.
If the fund has no concentration limits, how should the manager re-
balance the portfolio?
Caveat:
1. Strictly speaking, this procedure is only correct under the following assump-
tions:
(a) the term structure is flat (all bonds have the same YTM);
(b) the change in yield is small and instantaneous.
2. Nevertheless, (4.4) is still a good guide for bond management in most real-life
cases.
4.6.3 Duration for general term structures
When the term structure of interest rates is not flat, a more precise measure of
duration uses the appropriate spot rates.
Definition 4.6.2: Duration (general term structure)
The Duration of a coupon bond is is

m
1 X
D := ti Cti Zti (4.6)
P
i=1
where
• ti are the coupon payment dates (in years)
• Cti is the total cash flow at date ti : Cti = c/n for i < m, and
Cti = c/n + 1 for i = m, where n is the coupon payment frequency.
• Zti is the spot discount factor for date ti .
Note:
1
1. With annual compounding (r1 or EAR), Zti = [1+r1 (0,t1 )]ti
2. With continuous compounding, Zti = e−r∞ (0,ti )×ti
3. This duration is also known as “Fisher-Weil Duration”.
Example 4.6.4. Consider a 1-year Treasury Bond paying 4% semi-

annually. The spot rates are r∞ (0, 0.5) = 1% and r∞ (0, 1) = 1.2%
The bond price and duration are
B = . . . = 1.027733
DCBB = . . . = 0.990318
4.6.4 Passive bond management: Immunization
Some financial institutions sell contracts that commit them to a stream of fixed
payments in the future. For example, many insurance companies sell “Guaranteed
Investment Contracts” (GICs). GICs are essentially zero-coupon bonds. They
guarantee a given return over some period and therefore are popular products for
individual’s retirement-savings accounts. The insurance company can fund those
future commitments by investing in bonds today.
There are several investment alternatives:

1. If there are ZCB for all maturities, the company can do a simple cash flow
matching strategy. This would be the perfect hedging strategy. However, it
is unlikely that there are ZCB available for all the required maturities.
2. Following naive strategies will leave the firm exposed to interest rate risk.
For example:
• Investing in long-term bonds leaves the firm exposed to price risk: loose
money if interest rates go up.
• Investing in short-term revolving deposits or bills leaves the firm ex-
posed to reinvestment risk: loose money if interest rates go down.
3. An immunization strategy is a particular way of investing in bonds with the

following advantages:
• Hedges the net value of the financial institution, that is, it eliminates
interest rate risk.
• Allows the institution to choose bonds with good liquidity and lower
transaction costs.
Definition 4.6.3: Immunization strategy
Consider a stream of liabilities due at future dates. The goal of an immu-

nization strategy is to build an asset portfolio that guarantees the payment
of those liabilities with certainty, that is, regardless of what happens to
interest rates until the discharge dates.
Under the assumptions detailed below, immunization can be achieved as fol-

lows:
Proposition 4.6.3: Immunization (infinitesimal parallel shift)
To immunize a stream of liabilities against an instantaneous infinitesimal

parallel change in the term structure of interest rates, select a portfolio of
bonds with:
1. PV assets = PV liabilities
2. Duration assets = Duration liabilities

where Duration is defined in (4.6).
Example 4.6.5. We want to immunize a liability of 1 M$ due in 2

years. The following assets are available:
ZCB: 1 year to maturity
CBB: 5% annual coupons, 3 years to maturity.
The term structure is flat at r∞ (0, T ) = 5%, ∀T .

1) Compute the necessary investment today:
P VA = P VL ⇒ P VA = . . .
2) Compute the price and duration of each asset:
PZCB = . . .
DZCB = . . .
PCBB = . . .
DCBB = . . . = 2.8591
3) Determine the weight in each asset:
DA = DL ⇒ . . .
4) Determine the face value of each asset to purchase:
F VZCB = . . .
F VCBB = . . .
The following table summarizes all the results:
Asset w Investment Price Face Value
ZCB 0.4621 418 143 0.951229 439 582
CBB 0.5379 486 694 0.996547 488 381
Total 904 837
Note that the immunization strategy will work under the following assump-
tions:
1. Interest rates shift in an infinitesimal parallel way
2. This term structure shift happens immediately after setting up the portfolio.
3. Afterwards, the forward rates will become the future spot rates.
Example 4.6.6. Continuing the previous example, suppose that in-

terest rates increase by 40 bp immediately after purchasing the assets.
Then, spot rates remain at r∞ (0, T ) = 5.4%, ∀T , for the next two
years.
Reinvesting the proceeds of the ZCB at t = 1 for one additional year,
will generate a total value at t = 2 of:
VZCB (2) = . . .
For the CBB, reinvest the initial coupon and liquidate everything at
t = 2 for a total of:
VCBB (2) = . . .
Hence, the terminal value of the assets exceeds the liability:
VA (2) = . . . = 1 000 008 > VL (2) = 1 000 000
This means that the immunization strategy worked well!
Intuitively, why does immunization work? There are two opposite effects for a
given change in rates:
1. If interest rates increase, we have a capital loss today, but the investment
grows at a faster rate in the future.
2. If interest rates decrease, we have a capital gain today, but the investment
grows at a lower rate in the future.
The duration is the date by which these two effects cancel out and the value of
the assets matches the liability for any interest rate scenario.
VA (t)
✻
✲
t
D
Alternatively, we can illustrate the same effect by plotting the terminal value
of the assets for different values of the interest rate:
VA (D)
✻
✲
r⋆
r
4.6.5 Exercises
(p. 547): 3, 4. Same problem numbers in 2018 and 2021 editions.
Chapter 5
The efficient market hypothesis
5.1 Motivation
• Do security prices reflect the available information?
• Important implications for

– Investment decisions
– Efficient resource allocation
5.2 Random walks and the EMH
EMH statement:
1. The Efficient Market Hypothesis (EMH) is that stock prices reflect all the
available information.
2. If EFM is true, then prices change only in response to new information,

which is by definition unpredictable
3. The mathematical model for random price changes is a random walk.1

1
To be precise, investors expect to be rewarded for risk, so the expected price change
is positive over time (this is called a submartingale). Hence, the common model for prices
is a positive trend with random fluctuations about the trend.
71
5.3. Implications of the EMH 72
EMH and Competition:
1. Once information becomes available, market participants quickly analyze it

to trade on it.
2. Hence, competition ensures prices reflect information
3. However, there is a Catch-22:

• Stock analysis ensures markets are efficient
• Stock analysts must be compensated for their work, which can only
happen if markets are not efficient
• Equilibrium? Benefit = Cost
Versions of the EMH
1. Weak-form. Prices already reflect all information in past trading data (stock
prices and volumes)
2. Semistrong-form. Prices already reflect all public information (past prices,

plus accounting, macro, media, etc, information)
3. Strong-from. Prices already reflect ALL information — public and private

— related to the firm.
Example 5.2.1. (This is concept check 11.1 in BKM)

1. Suppose high-level managers make superior returns on invest-
ments in their company’s stock. Would this be a violation of
weak-form market efficiency? Would it be a violation of strong-
form market efficiency?
2. If the weak form of the EMH is valid, must the strong form also
hold? Conversely, does strong-form efficiency imply weak-form
efficiency?
5.3 Implications of the EMH
Some implications of the EMH:
• Weak-form efficiency implies that Technical analysis is worthless

5.4. Are markets efficient? 73
• Semistrong-form efficiency implies that:

– Fundamental analysis is worthless;
– the best investment strategy is Passive Management (buy and hold an
index fund), rather than Active Management (stock picking).
Example 5.3.1. (This is concept check 11.3 in BKM). What would

happen to market efficiency if all investors attempted to follow a pas-
sive strategy?
5.4 Are markets efficient?
Evidence on Mutual Fund Performance is mixed, but consistent

with market efficiency
• Some evidence of persistent positive (superstar or “hot-hand” managers)

and negative performance (due to too much trading). But most managers
are not consistent.
• How to correctly adjust returns for risk? Some apparent “anomalies” may
represent poorly understood risk premia.
Case in point: rise and fall of LTCM in late 1990s.
Most economists agree that markets are efficient
• The introduction by John Cochrane in this AFA conference nicely summa-

rizes the most common academic view.
• In this Chicago Booth Review video, Eugene Fama (the author of the EMH)
debates whether markets are efficient with Richard Thaler (a behavioral
economist that contests market efficiency).2
2
Thanks to Leander Opperman, MBA 2020, for pointing these videos.
5.5. Exercises 74
Conclusion:
• “Don’t try this at home!”

Individual investors (with other day jobs) should not engage in stock picking.
It’s better to buy a diversified portfolio.
• Professional analysts are probably making an extra return at least equal to

the cost of gathering information.
Caveat. Two economists are walking down the street. They spot a $20 bill on
the sidewalk. One starts to pick it up, but the other one says, “Don’t bother; if
that bill was real someone would have picked it up already”.
5.5 Exercises
(p. 381): 1, 2, 5, 9, 12. Same problem numbers in 2018 and 2021 editions.
Ex. 5.2 — Bodie, Kane, and Marcus (2014) CFA problems at the end of chapter
11 (p. 384): 1, 2, 3, 4. Same problem numbers in 2018 edition. If you have the
2021 edition, do problems 1 and 2.
Chapter 6
Futures markets
Based on chapter 22 of Bodie, Kane, and Marcus (2014).
6.1 Definition
6.1.1 Forward contract
Definition 6.1.1: Forward contract

A Forward contract is an OTC agreement between two counterparties
whereby:
1. One counterparty (“long forward position”) agrees to buy N units of
a given asset, at a future date T , for a forward price F (0, T ) defined
today
2. The other counterparty (“short forward position”) agrees to sell the

asset under the same conditions.
Example 6.1.1. It’s Oct/2017. An oil refinery negotiates with a

bank a forward contract to buy 100 000 barrels of crude oil one year
from now at a price of 50 dollars per barrel.
In the notation above, F (0, 1) = 50 $/bbl. If we want to make clear
the dates, we may write F (Oct/2017, Oct/2018) = 50 $/bbl. Note that
the maturity date, Oct/2018, will not change.
Note that the spot oil price today, in Oct/2017, does not enter any-
where in this forward contract.
75
6.1. Definition 76
The maturity date T is also denoted settlement date or delivery date.
Payments:
1. At inception there is no payment.
2. At the settlement date T , there are two alternatives according to what the
contract specifies:
• Physical settlement. The short delivers the physical asset and the long
pays the price initially agreed upon, F (0, T ).
• Cash settlement. The payoff to the long forward counterparty is
CF to long forward at T = N × [S(T ) − F (0, T )] (6.1)
where S(T ) is the spot price of the underlying asset at the settlement
date T . The payoff to the short is the symmetric of (6.1).
In either case, the effective price paid for the asset at time T is F (0, T ),
rather than S(T ).
Example 6.1.2. Continuing the previous example, suppose the con-

tract specifies cash settlement. At maturity, Oct/2018, the spot price
of oil is 53 $/barrel.
The payoff to the refinery (long forward) in Oct/2018 is:
CF to refinery = . . . = +300 000 $
The bank (short forward) pays this amount.
Draw the Forward contract profit profile at maturity for each position:
Profit
✻
p ✲
S(T )
F (0, T )
6.1. Definition 77
6.1.2 Futures contract
Definition 6.1.2: Futures contract

A Futures contract is an Exchange-traded contract between two counter-
parties whereby:
1. One counterparty (“long futures position”) agrees to buy N units of
an underlying asset, at a future date T , for a future price F (0, T )
defined today
2. The other counterparty (“short futures position”) agrees to sell the

asset under the same conditions.
Figure 6.1 lists the most common futures contracts.
Figure 6.1: List of typical futures contracts (from BKM)
One of the most important Futures exchanges is http://www.cmegroup.com/.

6.1. Definition 78
Futures curve. At each moment, we can trade several futures on the same
commodity with different delivery dates.
Figure 6.2: Crude oil futures curve
Other important characteristics of futures contracts:
• Settlement type. Some contracts are only cash settled (eg, stock index fu-
tures), whereas others permit physical delivery (eg, crude oil). However,
in practice most contracts are closed out, or reversed, before maturity, so
that actual physical delivery only happens in a small fraction of contracts
(1%–3%).
• Convergence. The futures price and the spot price must converge at matu-
rity; otherwise, there is an arbitrage opportunity. In other words, the basis
must go to zero:1
t→T
[F (t, T ) − S(t)] −−−→ 0
6.1.3 Differences between Futures and Forwards
Futures and forward contracts are very similar. However, there are some important
differences:
1
The basis can be defined as either F − S or S − F .
6.1. Definition 79
1. Futures are exchange-traded. In a Futures, the counterparty to every trader

is the exchange clearinghouse.
2. Futures are standardized. We can only trade the contracts for the underlying
securities, maturities, and quantities that the exchange specifies. This makes
futures more liquid than forward contracts, but less adaptable to specific
needs of traders.
3. Futures are marked-to-market daily. Profits or losses accrue to traders with

daily frequency. If we buy a futures contract today (i.e., enter into the long
position), our cash flow tomorrow is:
CF to long futures at day 1 = N × [F (1 day, T ) − F (0, T )]
This mark-to-market is repeated every day until we close the position. If we

hold the futures until maturity,
CF to long position over life of the futures = N ×

[F (1 day, T ) − F (0, T )
+ F (2 days, T ) − F (1 day, T )
+ ...
+ F (T, T ) − F (T − 1 day, T )]
= N × [S(T ) − F (0, T )]
The last line uses the fact that the futures price must converge to the un-
derlying security price at the futures maturity date: F (T, T ) = S(T ).
Hence, the total payoff is similar to a forward contract (equation 6.1), despite
the timing of the cash flows being different.
4. Margins. Futures traders need to post an “initial margin” with the exchange
(5%–15% of the total value of the contract) and must replenish the account
whenever the daily losses drive the balance below a given “maintenance
margin”.
Margins and daily mark-to-market virtually eliminate the counterparty credit
risk in futures contracts.
6.2. Trading strategies 80
6.2 Trading strategies
6.2.1 Speculation and leverage
Speculation trading rule
• Expect futures price to increase ⇒ Long futures
• Expect futures price to decrease ⇒ Short futures
Example 6.2.1. (Example 22.3 in BKM) Crude oil futures are trad-
ing at 91.86 $/bbl. Each contract is on 1 000 barrels. You believe that
oil prices are going to increase.
1. Trading strategy (today): (long/short) futures
2. Result (some days latter): suppose that the futures price in-
creases by $2 and you decide to close the position.
Profit = . . . = 2 000 $/contract
Trading in futures allows for much higher leverage than trading in the under-
lying spot asset.
Example 6.2.2. (Example 22.4 in BKM) Continuing the previous

example,
• Suppose the initial margin for the oil contract is 10%.
1. Initial margin = . . . = 9 186 $/contract
2. The $2 increase represents a percentage gain of
Return on posted margin = 9 186...$/ctt = 21.77%
• If instead we had invested in spot oil, the return for the same $2
increase would have been only
Return on spot oil = 91.86...$/bbl = 2.177 %
A 10-to-1 ratio!
6.2.2 Hedging
The most basic hedging strategy consists of trading a futures today to guarantee
the price at which we will trade the underlying asset in the future.
Example 6.2.3. On 2018/09/20, CME Corn Futures for March 2019

are trading at 3.58 $/bushel. If we buy futures contracts today, we
guarantee that we will be able to buy corn in March, at that price we
see today.
(Each futures contract is for 5 000 bushels, approximately 127 metric
tons, which makes a lot of corn flakes.)
In general, we should first define our underlying exposure, or “cash” position,

and trade futures appropriately.
Long/short definition. Recall we are long (short) on X if we profit from an

increase (decrease) in the price of X.
Examples:
• If I hold 100 shares of Microsoft, I am long on Microsoft stock.

• An oil producer owning a large field of crude oil is long on crude oil.
• An oil refinery that will need to buy crude oil one month from now has a
short exposure to crude oil because it will profit if the spot oil price decreases
until then.
• A trader that has purchased a crude oil futures is long on the futures because
it will profit if the futures price increases.
• A firm that will invest a revenue at some future date is long on interest rates
(because it will earn more interest if rates go up) and short on bond prices
(because bond prices move in opposite direction to interest rates)
Hedging trading rule
To hedge an underlying cash exposure, take a position in futures such that

the gains/losses in the futures offset the losses/gains in the cash position.
• Long spot position (ie, will sell later) or borrowing sometime in the
future (ie, loose if price decreases) ⇒ Short futures
• Short spot position (ie, will purchase later) or investment sometime

in the future (ie, loose if price increases) ⇒ Long futures
Example 6.2.4. (Example 22.5 in BKM) Consider an oil producer

planning to sell 100 000 barrels of oil in February that wishes to hedge
against a possible decline in oil prices. February futures are trading
at 91.86 $/barrel. Each contract is on 1 000 barrels.
1. To hedge its production, the company should (buy/sell)
(number) contracts.
2. Whatever the spot oil price in February, the profit/loss in the
futures exactly offsets the change in the cash position, such that
the effective sale price is 91.86 $/barrel. Figure 6.3 illustrates
this.
Figure 6.3: Hedging with futures (Fig 22.4 in BKM)
Example 6.2.5. A firm will receive 10 M$ from a client in 6 months.

The firm wants to guarantee the rate at which it will be able to invest
that sum for three months.
Consider a Treasury Bill that matures in 9 months. A bank quotes
a 6-month forward contract on this ZCB at F (0, 0.5) = 0.9901, with
physical settlement.
1. The firm “buys the forward” or “buys the bond forward”, i.e.,
takes the long position in the forward. No payments exchange
hands today.
2. In 6 months, the firm will obtain the T.Bill for the contracted
price. The forward rate implied in the forward T.Bill price is
thus:
r4 (0, 0.5, 0.75) : . . .
This means that the firm is guaranteed to get 3.9996% on its
investment.
3. The notional amount defined in the contract is such that:
Investment = N F (0, Tf ) ⇒ N = . . .
Example 6.2.6. Continuing the previous example, suppose instead

that the forward contract had cash settlement.
Suppose that 6 months from now, r4 (0, 0.25) = 3%.
1. The market price of the Tbill is now
S(0.5) = . . .
2. The forward payoff to the firm is
CF to long forward at T = . . . = N × 0.002456
3. The effective price to buy the bond in the market is thus
Eff Price = N × [S(0.5) − profit from forward] = N × . . .
4. This guarantees that the firm can invest at an effective rate of
r4 (0, 0.25) = 3.9996% (check this), rather than at the r4 (0, 0.25) =
3% currently available in the market.
6.2.3 Spread trading

• Definition: a spread trade consists of a long position in one futures contract
and a short position in another futures contract.
• Goal: Speculate or hedge a possible price divergence/convergence
• Typical spreads:
– Different but related commodities
– Different delivery months of the same commodity
• Advantages:
– Only relative price changes matter, i.e., do not have to guess direction
of the market
– Margin requirements are lower because spreads are less volatile than
absolute levels. Hence, can get more leverage.
Spread trading rule
Let H be the high-price and L the low-price commodity. The spread is

S = H − L.
• If expect S to widen: buy H, sell L
• If expect S to narrow: buy L, sell H
Example 6.2.7. It is now July. Consider the following NYMEX/CME

futures contracts for New York Harbor delivery, each on 42 000 gallons:
Futures January ($/gallon)

Heating oil 2.96
Gasoline 2.67
You believe that the winter in the U.S. East coast will not be as cold as
the market expects, and thus that January heating oil will get closer
to gasoline.
1. What is your trading strategy?
2. Suppose you close the position in December, when the January
futures are trading at: gasoline, 2.7 $/gal; heating oil, 2.8 $/gal.
Compute your profit if you had traded 10 contracts long and 10
short. (Answer: $ 79 800).
Example 6.2.8. Suppose it’s March/2013. You expect the Brent fu-
tures curve in Figure 6.2 to flatten. What is your trading strategy?
Suppose when you close the position the curve is flat at 95 $/bbl.
Compute your profit.
Remarks:
• Spread trading is heavily used for hedging in the Energy industry. Typical
spreads include:
– Crack spread — to hedge the profit of an oil refinery
– Spark spread — to hedge the profit of a gas-fired power plant
– Dark spread — to hedge the profit of a coal-fired power plant
6.3. Single-stock futures 85
• We can also speculate on basis (F − S) changes with the appropriate trades.
Example 6.2.9. (Example 22.6 from BKM). Suppose that gold to-
day sells for 991 $/ounce, and the futures price for delivery in 6 months
is 996 $/ounce. Therefore, the basis is currently 5 $/ounce.
1. To bet that the basis will narrow, what is your trading strategy?
2. Suppose that a few days from now, the spot price increases to 995
$/ounce, while the futures price increases to 999 $/ounce. Check
that your net gain equals the change in the basis, 1 $/ounce
6.3 Single-stock futures
Single-stock futures (SSF) are not heavily traded, but they are a good starting
point to understand index futures pricing.
Suppose a stock that does not pay dividends is trading at $50. Investors can
borrow or lend at a risk-free rate of r1 = 2%.
The arbitrage-free one-year futures price has to be
F (0, 1) = S(0)(1 + r1 ) ⇒ F (0, 1) = $51
Arbitrage strategies:
1. Suppose F (0, 1) = $52.

(a) Borrow $50 to buy the stock
(b) Sell the futures today at $52.
(c) One year from now, repay the loan and deliver the stock for a profit of
...
2. Suppose F (0, 1) = $49.

(a) Short the stock
(b) Invest the proceeds ($50) at the risk-free rate
(c) Buy the futures today at $49.
(d) One year from now, collect you risk-free investment, take delivery of
the stock in the futures for $49, use it to close the short. The net profit
is ...
6.4. Futures prices of stock indices 86
Alternative strategy for F low: even if it is not possible to short the stock,
investors that are long the stock will sell it now, invest the proceeds, and
buy the futures. One year from now, they will recover the stock from the
futures and collect the risk-free investment. They become long stock again
and also receive an additional arbitrage profit. Many investors will do this
until F (0, 1) = S(0)(1 + r1 ) is restored.
6.4 Futures prices of stock indices
Units:
• Stock indices (eg, S&P 500) are usually quoted as unitless numbers (“index
points”)
• Futures on those stock indices are quoted in the same “index points”
• One Futures contract calls for delivery of some dollar multiple of the index
number: $k × index. For example, the most important futures on the S&P
500 has a multiplier of $250, ie, the contract size is $250.
Proposition 6.4.1: Futures pricing
If there are no arbitrage opportunities, the Futures price of a stock index is
F0 = S0 (1 + r1 )T − DT (6.2)
where
• F0 or F (0, T ) is the current futures price for delivery at time T
• S0 or S(0) is the current spot price of the underlying index
• DT is the future value of the dividends paid on the underlying stocks

over the life of the futures (assumed known and compounded to time
T at the risk-free rate).a
• r1 or r1 (0, T ) is the spot risk-free rate (with annual compounding) for

the maturity of the future.b
a
To avoid having different units on the LHS and RHS of this equation, we need
to interpret F0 and S0 as dollar values, that is, as a dollar multiplier ($k) times
the quoted index points. As this scaling will not affect pricing, it is convenient to
work with $k = $1.
b
Note that subscripts under F , S, and D denote time, whereas under r (and d
below) denote compounding frequency.
Intuition. Recall that:

• Stock indices (Dow Jones, S&P 500, etc) usually do not account for dividends
• If you hold a mutual fund that tracks the index, or if you buy the stocks
directly, you obviously receive dividends
Hence, if we invest in the stock portfolio and sell the index forward to guarantee
the capital appreciation, the gross return of
F0 + DT
S0
should be risk free. Equation (6.2) immediately follows.
Example 6.4.1. (This is from section 22.4 of Bodie, Kane, and Marcus
(2014)). We want to price a 1-year futures contract on the S&P 500.
1. Suppose the S&P 500 is currently at 1 000 (this is a unitless
index).
2. Suppose the futures contract calls for delivery of $1 times the
value of the SP500 index.2 Hence, S0 = $1 000.
3. Suppose that if we invest $1000 in a mutual fund that tracks the
SP500, we will receive $20 of dividends over the course of the
year. Assume that all dividends are paid at the end of the year.
4. Assume the risk-free rate is r1 (0, 1) = 3%.
5. Then,
F0 = . . . = $1 010 = 1 010 × $1
which would be quoted as 1 010 index points.
Proof of proposition 6.4.1. If equality does not hold, there is an opportunity for
index arbitrage.
• If F0 > S0 (1 + r1 )T − DT , the arbitrage strategy is
Cash Flow at
Trade t=0 t=T
Borrow S0 risk-free +S0 −S0 (1 + r1 )T
Buy S0 of stocks in index −S0 ST + DT
Short Futures 0 F0 − ST
Total 0 F0 − S0 (1 + r1 )T + DT
2
The two most common contracts traded on CME are actually for $250 or $50 times
the SP500 index, but this scaling does not matter for pricing.
The total CF at T is positive by assumption.
• If F0 < S0 (1 + r1 )T − DT , do the opposite trades. Recall that when you

short stock you need to pay the dividends.
Example 6.4.2. Continuing the previous example, suppose that the

1-year futures is trading at 990. Complete the arbitrage strategy:
Cash Flow at
Trade t=0 t=T
Invest S0 risk-free
Short S0 of stocks in index
Long Futures
Total
This guarantees a $20 profit for each futures contract traded.
Remarks about the spot-futures parity relation (6.2):
1. Alternative with D0 . With continuous compounding, (6.2) becomes F0 =

S0 er∞ T − DT . Let D0 denote the present value of the dividends paid during
the life of the contract and note that DT = D0 er∞ T . Then, (6.2) is equivalent
to
F0 = (S0 − D0 )er∞ T (6.3)
2. Alternative with dividend yield. Equation (6.3) can be further modified to

S0 − D0 r∞ T S0 − D0
F0 = S0 e = S0 exp ln er∞ T
S0 S0

1 S0
= S0 exp − ln T er∞ T
T S0 − D0
or
F0 = S0 e(r∞ −d∞ )T (6.4)
where d∞ = T1 ln S0S−D
0
0
is the annual dividend yield (with continuous com-
pounding) estimated with the dividends paid during the life of the futures.3
3
Equation (6.4) is the standard representation in Hull (2012).
Bodie, Kane, and Marcus (2014) present a discrete-compounding version of (6.4) as
F0 = S0 (1 + r1 − d? )T . When the futures maturity is exactly T = 1 year, d? can be easily
6.5. Investment strategies with stock-index futures 89
3. Investment assets. Equation (6.2) can be applied directly to other assets

held purely for investment purposes, like gold or silver, by setting DT = 0.
4. Bonds. Equation (6.2) can be applied to bonds by replacing Dividends

with Coupons. However, strictly speaking, the equality would be exact for
Forward bond prices, but would only be an approximation for Futures bond
prices. The reason is the different cash flow timing induced by mark-to-
market in futures, which cannot be ignored in interest-rate derivatives.
5. Consumption commodities. (Advanced optional material) Equation (6.2)

can be extended to consumption commodities (like crude oil, corn, etc) by
including storage costs, but we only get a no-arbitrage upper bound on the
futures price: F0 ≤ (S0 + U0 )erT , where U0 is the present value of storage
costs. For further details on the pricing of these other contracts see Hull
(2012).
6.5 Investment strategies with stock-index fu-

tures
Stock-index futures allow for market-timing strategies with much lower transac-
tions costs than trading in the individual stocks of the index.
Assumptions:
• These strategies are typically short-term, so we ignore dividends.
• The horizon of the strategy is the same as the maturity of the futures.
6.5.1 Creating synthetic stock positions
(Based on section 23.2 of Bodie, Kane, and Marcus (2014))
estimated by d? = D1 /S0 . However, I don’t see how to estimate a d? that works for any
T 6= 1, other than by first getting d∞ from d∞ = T1 ln S0S−D 0
0
, and then forcing d? to
(r∞ −d∞ )
be the value that solves 1 + r1 − d? = e . When T 6= 1, BKM seem to estimate
effective rates for a short horizon of 1/n years, r(n) and d(n) . For example, if n = 12, d(12)
is the monthly (not annual) dividend yield from dividends received over the next month.
The futures price is then F (t, T ) = St (1 + r(n) − d(n) )n(T −t) .
Synthetic stock position
To create a synthetic long stock position equivalent to a total initial equity

investment of $V0 for a period of T years,
1. Invest V0 dollars in Treasury Bills maturing at T .
2. Buy N = V0 /S0 index futures for delivery at T . (Note that we thus

have V0 = S0 × N .)
The resulting cash flows are:
Cash Flow at
Trade t=0 t=T
Invest S0 N in T-bills −S0 N +S0 N (1 + r1 )T = F0 N
Long N Futures 0 (ST − F0 )N
Total −S0 N +ST N
where
• N is the number of futures contracts
• S0 is the dollar spot price of the underlying index ($k× index)
• F0 = S0 (1 + r1 )T by prop. 6.4.1 when DT = 0
• r1 is the spot risk-free rate for period T (with annual compounding)
Remark: Note that the rate of return is same as on a real stock portfolio
(ignoring dividends):
VT /V0 − 1 = (ST N )/(S0 N ) − 1 = ST /S0 − 1
Example 6.5.1. (Example 23.3 in Bodie, Kane, and Marcus (2014)).

• An institutional investor wants to bet 140 M$ that the market
will go up over the next 1 month. To minimize trading costs, he
wants to trade in futures.
• The SP500 index is now at 1400.
• The Futures contract has a multiplier of $250.
Investment strategy:
1. The amount to invest in T-bills is 140 M$
2. Check that the number of futures to buy is
N = . . . = 400
6.5.2 Hedging an equity portfolio from market risk
(Based on section 3.5 of Hull (2012))
Market risk hedging
To hedge an existing stock portfolio from market movements over a period

of T years, we should short index futures. The optimal number of contracts
to short in order to minimize the volatility of the combined position is
V0
N =β
F0
where
• V0 is the current market value of the equity portfolio
• F0 or F (0, T ) is the dollar value of 1 futures contract (ie, contract size

times futures quote)
• β is the CAPM beta of the portfolio.
Remarks:
• This rule is from Hull (2012, equation 3.5)
• Recall that beta can be estimated from the time series regression
(rp − rf )t = αp + βp (rM − rf )t + εpt (6.5)
If the fund is fairly price according to the CAPM, we have αp = 0. Hence,

the hedge will work well if
r̃p,t+1 = rf + βp (r̃M,t+1 − rf ) (6.6)
Example 6.5.2. The manager of an equity portfolio worth 50 M$ is

bullish on the market over the long term, but is afraid that there may
be a big drop over the next month. The manager wants to avoid this
risk, but it would be too expensive to sell the whole portfolio now and
rebuild it 1 month from now due to transaction costs.
Other inputs:
• The portfolio has β = 1.5
• The SP500 is at 2050
• The risk-free interest rate with monthly compounding is 3% (ie,
the interest received at the end of 1 month is 0.25%).
6.6. Exercises 92
• The 1-month Futures is quoted at 2055.125. The contract size,

or multiplier, is $250.
HEDGING STRATEGY:
Check that the number of futures to short is
N = . . . ≈ 146
VERIFICATION:
The following table shows alternative scenarios for the market over
the next month. Note that the combined return is always very close
to the risk-free rate! (Check the numbers in the table)
Scenario
Inputs
rm -5.00% -1.00% 0.25% 1.00% 5.00%
SP500 1947.50 2029.50 2055.13 2070.50 2152.50
Ouputs
Futures payoff
Short position 3 928 313 935 313 0.00 - 561 188 - 3 554 188
Cash Portfolio
rp -7.625% -1.625% 0.250% 1.375% 7.375%
Vp 46 187 500 49 187 500 50 125 000 50 687 500 53 687 500
Combined Portfolio
Total value 50 115 813 50 122 813 50 125 000 50 126 313 50 133 313
rc 0.232% 0.246% 0.250% 0.253% 0.267%
6.6 Exercises
Ex. 6.1 — Crude oil futures are trading at 50 $/bbl. Each contract is on 1 000
barrels. The initial margin is 10%. You believe that oil prices are going to decrease
and want to get exposure to 20,000 barrels.
1. Your trading strategy is to [buy/sell] [number] fu-
tures contracts
2. Some days latter you close the position when the futures is trading at 47
$/bbl. Compute the rate of return on your initial margin.
Ex. 6.2 — An investors holds a diversified portfolio of stocks worth 10 M$ in

October/2017. The manager wants to hedge the portfolio from market movements
6.6. Exercises 93
for the next couple of months. The beta of the portfolio is estimated at 1.1. The
S&P500 Dec/2017 futures is quoted at 2526 (the contract multiplier is $250). How
many futures contracts should the investor trade? Indicate whether the investor
should Buy or Sell.
(p. 794): 8, 10, 14. The problems in the 2021 edition are the same, except some
prices and rates are different.
Chapter 7
Swaps
Reference: section 23.4 of Bodie, Kane, and Marcus (2014).
7.1 Interest Rate Swaps
Definition 7.1.1: Interest Rate Swap (IRS)
A plain vanilla fixed-for-floating Interest Rate Swap (IRS) is a contract be-

tween two counterparties whereby they agree to exchange interest payments
on a notional N , at dates Ti (with i = 1, 2, . . . , m), on the following terms:
1. One counterparty (“fixed rate payer”) pays a fixed rate cn at each
date Ti .
2. The other counterparty (“floating rate payer”) pays a variable rate

In (Ti−1 , Ti ) at each date Ti .
where:
• cn is the fixed rate defined at the beginning of the contract (t = 0)
• In (Ti−1 , Ti ) is a money market index, typically the LIBOR, which will

be observed in the market at future time Ti−1
• Both rates are expressed with a compounding frequency n correspond-

ing to the length of the period in the contract, that is, 1/n = Ti −Ti−1
Example 7.1.1. Consider the following IRS:

• Notional = 10 M$
94
7.2. Hedging with an IRS 95
• Firm pays fixed c2 = 4%

• Bank pays LIBOR 6M
• Payment dates: every 6 months.
• Termination in 5 years.
The firm will have to pay every 6 months. The
bank will pay depending on LIBOR.
Suppose that in 6 months LIBOR 6M fixes at 3.5%. Then, 12 months
from now:
Net CF to fixed payer (firm) = . . . = −$25 000
Note that only the net amount exchanges hands.
7.2 Hedging with an IRS
IRS are useful to transform existing liabilities (or assets) from fixed rate to floating
rate, or vice-versa.
Example 7.2.1. Sometime ago, a firm issued floating-rate bonds to

borrow 10 M$ at LIBOR 6M + 0.4%.
• The manager believes that interest rates are likely to increase in
the coming months.
• In principle, the firm could issue fixed-rate bonds and use the
proceeds to buy back the outstanding floaters.
• But it is faster and cheaper to use a swap.
If the firm enters into the swap of the previous example, its CF every
6 months will be:
1. Pay debt at LIBOR 6M + 0.4%
2. Receive IRS floating at
3. Pay IRS fixed at
Hence, the swap transforms the floating-rate debt into synthetic fixed-
rate debt at a rate of
Every 6 months, the firm will pay a net amount of $220 000.
Example 7.2.2. (This is example 23.6 in BKM).

7.2. Hedging with an IRS 96
• A large portfolio currently includes $100 million par value of

long-term bonds paying an average coupon rate of 7%.
• The manager believes interest rates are about to rise. As a result,
he would like to sell the bonds and replace them with either
short-term or floating-rate issues.
• However, it would be exceedingly expensive in terms of trans-
action costs to replace the portfolio every time the forecast for
interest rates is updated.
• A cheaper and more flexible approach is to swap interest pay-
ments.
A swap dealer quotes the following IRS:
• Notional = 100 M$
• Fund (pays/receives) fixed c2 = 7%
• Dealer (pays/receives) LIBOR 6M
• Payment dates: every 6 months.
• Termination in 2 years.
The manager has, in effect, converted a fixed-rate bond portfolio into
a synthetic floating-rate portfolio (see figure 7.1).
Figure 7.1: Swap example (ex 23.6 in BKM)

Chapter 8
Solutions to Problems
Answer (Ex. 1.3) — E[rx ] = 0.14. V ar[rx ] = 0.6 ∗ (0.3 − 0.14)2 + 0.4 ∗ (−0.1 −
0.14)2 = 0.0384 ⇒ σx = 0.1960. Hence,
SRp = SRx = (0.14 − 0.02)/0.196 = 0.6124
Note that all portfolios along the CAL(x) have the same SR.
Answer (Ex. 1.5) — The Tangency portfolio is given by (1 means bond fund,
2 means stock fund):
r1e σ22 − r2e σ12
w1 = = 0.4783
r1e σ22 + r2e σ12 − (r1e + r2e )σ12
and w2 = 0.5217. T hence has
E[rT ] = w1 ∗ 0.04 + w2 ∗ 0.08 = 0.0609
and
σT = ((w12 ∗ 0.12 + w22 ∗ 0.22 + 2 ∗ w1 ∗ w2 ∗ 0.1 ∗ 0.2 ∗ 0.3))0.5 = 0.1272
The optimal portfolio for the investor is
0.15 = wT ∗ 0.1272 ⇒ wT = 1.1796
and wf = −0.1796
The complete best portfolio has
E[rp ] = wT ∗ E[rT ] + wf ∗ 0.02 = 0.0682
97
98
Answer (Ex. 1.6) — Since ρ = 0, the variance is just σp2 = wa2 σa2 + wb2 σb2 . Let
wb = 1 − wa .
The investor picks the portfolio that maximizes his utility function:
A 2 2
maximize [wa E[ra ] + (1 − wa )E[rb ]] − [w σ + (1 − wa )2 σb2 ]
wa 2 a a
The first-order condition is:
A
E[ra ] − E[rb ] − 2wa σa2 − 2(1 − wa )σb2 = 0
2
E[ra ] − E[rb ]
⇒wa = 2
+ σb /(σa2 + σb2 )
A
Using the numbers provided, wa = 0.7333 and wb = 0.2667
Answer (Ex. 6.1) — .

1) Sell 20 contracts
2) Initial margin = 0.1 ∗ 50 ∗ 20000 = 100, 000. Return = 3 ∗ 20000/100000 = 60%
Answer (Ex. 6.2) — Sell N = βV0 /F0 = 1.1 × (10 M $)/(2526 × 250$) = 17.4 ≈
17 contracts.
Bibliography
Bodie, Z., A. Kane, and A. Marcus, 2014, Investments. McGraw-Hill, 10th

edn.
Fama, E. F., and K. R. French, 1993, “Common Risk Factors in the Returns
on Stocks and Bonds,” Journal of Financial Economics, 33, 3–56.
, 1996, “Multifactor explanations of Asset Pricing Anomalies,” Jour-

nal of Finance, 51(1), 55–84.
Hull, J. C., 2012, Options, Futures, and other Derivatives. Pearson Educa-
tion.
Ross, S. A., 1976, “The Arbitrage Theory of Capital Asset Pricing,” Journal
of Economic Theory, 13, 341–360.
99

Investments - Teaching Notes 20221114

Uploaded by

Copyright:

Available Formats

Investments - Teaching Notes 20221114

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Investments - Teaching Notes 20221114

Uploaded by

Copyright:

Available Formats

João Pedro Pereira

November 14, 2022

2 The Capital Asset Pricing Model 22

3 Arbitrage Pricing Theory and Factor Models 30

3.3 Example with 3 factors: the Fama-French model . . . . . . . . . . 34

5 The efficient market hypothesis 71

1.1 Risk and return

Based on chapter 5 of Bodie, Kane, and Marcus (2014)

1.1.1 Holding period return

Definition 1.1.1: Holding Period Return

The Holding Period Return (HPR) on an investment from today (time 0)

• “Payouts” represent cash income from the asset between 0 and T

1.1.2 Expected return and standard deviation

True population moments

We want to characterize the probability distribution of returns across future states

Definition 1.1.2: Mean and Variance

• p(s) is the probability of state s occuring.

Example 1.1.1. Stock X is trading at $10. You estimate the follow-

State of Market Prob Year-end price Dividends

Compute the mean and standard-deviation of returns.

Estimation of mean and variance

1.1.3 Risk premium and Sharpe Ratio

Terminology for returns in excess of the risk-free rate:

The Sharpe Ratio, or Reward-to-Volatility Ratio, is an important measure of

Definition 1.1.3: Sharpe Ratio

Given the forward-looking true population moments and a constant risk-free

To compare historical returns on different portfolios, given that the risk-free

f = Average excess return

Example 1.1.2. Continuing the previous example for stock X, sup-

1.2 Risk Aversion and Capital Allocation

1.2.1 Risk Aversion

Definition 1.2.1: Risk aversion

We assume that most investors are risk averse:

• They only take risk if expecting a positive risk premium.

Modelling risk aversion

Different risky portfolios can be ranked with an expected utility function.

Expected utility function

If the investor has mean-variance preferences, the expected utility function

1. The parameter A characterizes the investor

2. This utility score U can be interpreted as a certainty equivalent rate of

Example 1.2.1. (This is concept check 6.2 in BKM). A portfolio has

We say that portfolio x mean-variance dominates portfolio y if all risk-averse

Proposition 1.2.1: Mean-Variance Dominance

Asset x mean-variance dominates asset y iff:

1.2.2 Portfolio rate of return

Definition 1.2.2: Rate of return on a portfolio

The return on a portfolio is:

• I = number of securities in the portfolio

• wi = proportion of funds invested in security i

Example 1.2.2. Consider the following portfolio:

Compute the portfolio’s terminal total value:

This implies a portfolio return of 114 000/100, 000 − 1 = 0.14.

Mean and variance

Proposition 1.2.2: Portfolio mean

For a portfolio with 2 assets (I=2):

Proposition 1.2.3: Portfolio variance

For a portfolio with 2 assets (I=2):

Remarks about covariance: