Fixed Income

Teaching notes

João Pedro Pereira

Nova School of Business and Economics
Universidade Nova de Lisboa
[Link]@[Link]

April 6, 2016
Contents

1 Fundamental concepts 5
1.1 Markets and securities . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.1.1 Relative sizes . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.1.2 Government debt market . . . . . . . . . . . . . . . . . . . 5
1.1.3 Money market . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.1.4 Repurchase Agreements . . . . . . . . . . . . . . . . . . . . 8
1.2 Price quotes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.2.1 Bond price units . . . . . . . . . . . . . . . . . . . . . . . . 9
1.2.2 Quoting conventions . . . . . . . . . . . . . . . . . . . . . . 9
1.3 Discount factors and interest rates . . . . . . . . . . . . . . . . . . 11
1.3.1 Discount factors . . . . . . . . . . . . . . . . . . . . . . . . 11
1.3.2 Interest rates . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.4 Spot and Forward rates . . . . . . . . . . . . . . . . . . . . . . . . 14
1.4.1 Term structure of interest rates . . . . . . . . . . . . . . . . 14
1.4.2 Forward discount factors . . . . . . . . . . . . . . . . . . . . 15
1.4.3 Forward interest rates . . . . . . . . . . . . . . . . . . . . . 15
1.5 Fixed coupon bonds . . . . . . . . . . . . . . . . . . . . . . . . . . 17
1.5.1 Bond price . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
1.5.2 Yield to Maturity . . . . . . . . . . . . . . . . . . . . . . . 19
1.6 Floating rate bonds . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
1.6.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
1.6.2 Pricing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
1.7 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

2 Yield curve fitting 27

2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
2.2 Bootstrap . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
2.2.1 Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
2.2.2 Application . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
2.3 Nelson-Siegel Model . . . . . . . . . . . . . . . . . . . . . . . . . . 31
2.3.1 Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
2.3.2 Meaning of the parameters . . . . . . . . . . . . . . . . . . 31
2.3.3 Application . . . . . . . . . . . . . . . . . . . . . . . . . . . 34

2
Contents 3

2.4 Svensson Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

2.4.1 Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
2.4.2 Aplication . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
2.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36

3 Bond portfolio management 38

3.1 The variation in interest rates . . . . . . . . . . . . . . . . . . . . . 38
3.2 Duration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
3.2.1 Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
3.2.2 Duration of a zero coupon bond . . . . . . . . . . . . . . . 39
3.2.3 Duration of a portfolio . . . . . . . . . . . . . . . . . . . . . 41
3.2.4 Duration of a fixed coupon bond . . . . . . . . . . . . . . . 42
3.2.5 Duration of a floating rate bond . . . . . . . . . . . . . . . 42
3.2.6 Traditional definitions of duration . . . . . . . . . . . . . . 45
3.3 Active bond management . . . . . . . . . . . . . . . . . . . . . . . 46
3.4 Passive bond management: Immunization . . . . . . . . . . . . . . 47
3.5 Duration of zero investment portfolios . . . . . . . . . . . . . . . . 50
3.6 Asset-Liability management . . . . . . . . . . . . . . . . . . . . . . 52
3.7 Convexity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
3.7.1 Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
3.7.2 Convexity of a zero coupon bond . . . . . . . . . . . . . . . 55
3.7.3 Convexity of a portfolio . . . . . . . . . . . . . . . . . . . . 56
3.7.4 Convexity of a fixed coupon bond . . . . . . . . . . . . . . 56
3.7.5 Application to immunization . . . . . . . . . . . . . . . . . 57

4 Interest rate derivatives 59

4.1 Forward rate agreements . . . . . . . . . . . . . . . . . . . . . . . . 59
4.1.1 Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
4.1.2 Hedging with a FRA . . . . . . . . . . . . . . . . . . . . . . 60
4.1.3 The value of an existing FRA . . . . . . . . . . . . . . . . . 61
4.1.4 Setting the fixed rate in a FRA . . . . . . . . . . . . . . . . 63
4.1.5 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
4.2 Interest Rate Swaps . . . . . . . . . . . . . . . . . . . . . . . . . . 67
4.2.1 Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
4.2.2 Hedging with an IRS . . . . . . . . . . . . . . . . . . . . . . 68
4.2.3 The value of an existing IRS . . . . . . . . . . . . . . . . . 68
4.2.4 Setting the fixed rate in an IRS . . . . . . . . . . . . . . . . 70
4.2.5 The swap curve . . . . . . . . . . . . . . . . . . . . . . . . . 71
4.2.6 Asset-Liability management with IRS . . . . . . . . . . . . 73
4.2.7 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
4.3 Forward contracts . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
4.3.1 Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
4.3.2 Hedging with a Forward contract . . . . . . . . . . . . . . . 77
4.3.3 Setting the initial Forward price . . . . . . . . . . . . . . . 78
Contents 4

4.3.4 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
4.4 Interest rate Futures . . . . . . . . . . . . . . . . . . . . . . . . . . 80
4.4.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
4.4.2 Eurodollar futures . . . . . . . . . . . . . . . . . . . . . . . 83
4.4.3 Treasury bond futures . . . . . . . . . . . . . . . . . . . . . 88
4.4.4 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
4.5 Interest rate Options . . . . . . . . . . . . . . . . . . . . . . . . . . 93
4.5.1 Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
4.5.2 Bond options . . . . . . . . . . . . . . . . . . . . . . . . . . 95
4.5.3 Caps and Floors . . . . . . . . . . . . . . . . . . . . . . . . 97
4.5.4 Futures options . . . . . . . . . . . . . . . . . . . . . . . . . 99
4.5.5 Swaptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99

5 Term structure dynamics: Vasicek model 101

5.1 Model for the short rate . . . . . . . . . . . . . . . . . . . . . . . . 101
5.2 Bond prices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
5.3 Estimating the parameters . . . . . . . . . . . . . . . . . . . . . . . 103
5.4 Option prices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105

Solutions to exercises 106

Bibliography 110
Chapter 1

Fundamental concepts

1.1 Markets and securities

1.1.1 Relative sizes

Size of fixed income markets is very large and still growing...

Table 1.1: Size of Fixed Income Markets (Dec/2008, Trillion $)

Source: Tab 1.1 in Veronesi (2010)
Market Market Value Notional
US Treasury Debt 5.9
US Municipal Debt 2.7
US Federal Agency Securities 3.2
US Money Market 3.8

Interest Rate Swaps 328.1

Interest Rate Forwards and Futures 58.5
Interest Rate Options 86.5

US Corporate Debt 6.3

1.1.2 Government debt market

Standard instrument to finance government operations in every country.

5
1.1. Markets and securities 6

US Treasury Debt Maturity Coupon rate

Treasury Bills Short term (4, 13, 26, None
or 52 weeks)
Treasury Notes Medium term (be- Fixed, semi-annual
tween 2 and 10 yrs)
Treasury Bonds Long term (up to 30 Fixed, semi-annual
yrs)
Remarks:

• Considered “safe” investment, ie, with low credit risk

• But its return is still risky due to:

– Potential capital losses if sold before maturity

– Reinvestment rates for future coupons
– Nominal coupons and principal, ie, there is inflation risk. (Excep-
tion: TIPS, where principal is indexed to inflation)

• On-the-run securities (most recently issued) are more liquid than off-
the-run.

Zero-Coupon Bonds

Definition 1.1.1: Zero-Coupon Bonds (ZCB)

ZCB are securities that pay only the principal at maturity

Examples:

• Treasury Bills

• STRIPS (Separate Trading of Registered Interest and Principal Secu-

rities).
Financial institutions strip the principal and coupons from Treasury
Notes and Bonds and sell them separately.
1.1. Markets and securities 7

1.1.3 Money market

Short-term borrowing/lending for financial institutions and large corpora-
tions
Important rates:

• LIBOR (London Interbank Offered Rate).

– Average interest rate that banks charge to each other for short-
term uncollateralized borrowing in the London Market.1
– Maturities (as of Jan/2014): Overnight, 1 week, 1 month, 2 m, 3
m, 6 m, 12 m.
– Currencies (as of Jan/2014): EUR, USD, GBP, JPY, CHF
• Euribor (Euro InterBank Offered Rate).
– Average interest rate that banks charge to each other for short-
term uncollateralized borrowing in the Euro area.
– Published daily by Euribor-EBF.
– Maturities (as of Jan/2014): Overnight (“EONIA” - Euro Overnight
Index Average), 1 week, 2 w, 1 month, 2 m, 3 m, 6 m, 9 m, 12 m.
– Currency: EUR.
• Eurodollar rate.
– A Eurodollar is a U.S. dollar deposited in a bank outside the U.S.
(thus not subject to US banking regulations). The term originally
meant a deposit in a European-based bank, but now it means
anywhere outside the US.
– The Eurodollar Deposit Rate is the rate on such deposits.
– However, in practice when we say Eurodollar rate, we usually
mean the USD LIBOR.
– The most important is the 3-month USD LIBOR (reference for
Eurodollar futures).
• U.S. Federal Funds Rate
– Rate that banks charge to each other to lend or borrow balances
kept at the Federal Reserve for 1 day (overnight).
1
Interesting read on the LIBOR manipulation scandal:
[Link]
1.1. Markets and securities 8

1.1.4 Repurchase Agreements

Definition 1.1.2: Repurchase Agreement

A Repurchase Agreement or Repo is a contract to borrow money with
collateral.

1. Today(t): trader sells a security to a dealer.

2. At maturity (T ): trader buys back security from dealer at same

price plus interest.

Remarks:

• A repo can be used to fund a long position (investment) in the under-

lying security.

• The opposite transaction is called a Reverse Repurchase Agreement or

Reverse Repo

1. Today(t): trader buys a security from a dealer.

2. At maturity (T ): trader sells back security to dealer for same price
plus interest.

• A reverse repo allows the trader to create a short position in the security

1. At t: borrow security through reverse repo

2. At t: sell it in the market
3. At T: buy in the market
4. At T: deliver to dealer to close reverse repo

• The interest rate is denoted repo rate. The repo rate is usually close
to, but lower than LIBOR (why?).

• The day-count convention for interest is actual/360.

• Repos are mainly overnight (in the US market), but there also repos
for maturities up to 1 month or even longer (aka term repos).2
2
For more details see the ICMA page at [Link]
1.2. Price quotes 9

1.2 Price quotes

1.2.1 Bond price units

In most markets, bond prices are represented as percentages of face values.

Meaning of Price in percentage of Face value

Total Investment ($) = Price (%) × Face Value ($)

Think of Face Value as a “quantity”.

Example 1.2.1. A Bond with 5% annual coupons is priced at

102.1234%.3 This means that:

• If we invest 1 M$, we are able to buy a face value (“quan-

tity”) of
F = 1 M$/1.021234 = . . . $
• Our next coupon will be C = . . . $
• At maturity, we will receive a total of $1,028,167.88

Alternative units:

1. Sometimes prices 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

1.2.2 Quoting conventions

Treasury Bills

• The price (P ) of a Treasury Bill is always below face or par value.

3
Interpret the symbol “%” as meaning “×0.01” to avoid messing up the units in your
calculations.
1.2. Price quotes 10

• Instead of P , traders quote an annualized “discount” (k):

n
k: P = 100% − k
360
where n is the number of calendar days until maturity (day count “ac-
tual/360”).

• This loosely says that the price is “k% below par”.

• Note that k is not the yield on the Bill.

Example 1.2.2. Assume a 26 week (182 days) bill is auctioned

for $98.123456 per $100 of face value.

• Traders quote 3.7118%

• If we invest $10,000 now, 182 days from now we get back
$10, 000/0.98123456 = $10, 191.24.

Treasury Notes and Bonds

Prices of coupon-paying bonds

Invoice price = Quoted price + Accrued interest

1. The Quoted (or clean, or flat) price quoted by a trader 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:

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
settlement date
1.3. Discount factors and interest rates 11

• 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 1.2.3. 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.

1.3 Discount factors and interest rates

1.3.1 Discount factors

Definition 1.3.1: Discount factor

The discount factor Z(t, T ) is the price at time t of $1 to be received
for sure at a later date T .

Example 1.3.1. A 90-day [Link] is trading at 99.1234%. Hence,

Z(0, 0.25) = 0.991234

Notation:
1.3. Discount factors and interest rates 12

• Z(.) stands for Zero because the discount factor equals the price of the
corresponding 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 ).

1.3.2 Interest rates

Definition

Suppose you can lend $100 at an annual rate of 10%. The amount of money
you will have at the end of the year, V , depends on the compounding fre-
quency:

• Annual: V = $100 × (1 + 0.1)

2
• Semi-annual: V = $100 × 1 + 0.1
2

0.1 360


• Daily: V = $100 × 1 + 360

• ...
0.1 n
= $100 × e0.1


• Continuous: V = $100 × limn→∞ 1 + n

Hence, a complete definition must always specify:

1. The annual rate of interest. (Eg, 10%)

2. The compounding frequency. (Eg, semi-annual compounding).

Definition 1.3.2: Interest rate

rn (t, T ) := annual interest rate from time t to T , with a compounding
frequency of n periods per year.
1.3. Discount factors and interest rates 13

Equivalence between discount factors and interest rates

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:

 −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 1.3.2. A 90-day [Link] is trading at 99.1234%. This

price implies the following equivalent quantities:

• Discount factor: Z(0, 0.25) = 0.991234

• r1 (0, 0.25) = · · · = 3.5846%
A.k.a “Equivalent Annual Rate” (EAR) or “Effective An-
nual Yield”.
• r4 (0, 0.25) = · · · = 3.5374%
A.k.a “Annual Percentage Rate” (APR) with quarterly com-
pounding or “Bond Equivalent Yield”.
• Continuously compounded rate: r∞ (0, 0.25) = · · · = 3.5219%

Note that the effective 3-month rate, x : 0.991234 = 1/(1 + x) ⇒

x = 0.8844%, is not usually quoted.
1.4. Spot and Forward rates 14

Relation between rates with different compounding frequencies

Using the previous relations, we can easily move from one compounding
frequency to another.

Proposition 1.3.1: Relation between rates with compounding

frequencies n and s
 n×(T −t)  s×(T −t)
rn (t, T ) rs (t, T )
1+ = 1+
n s

Example 1.3.3. Continuing with r1 (0, 0.25) = 3.5846% from

the previous example, compute the APR with semi-annual com-
pounding.
r2 (0, 0.25) : . . . r2 (0, 0.25) = 3.5530%

1.4 Spot and Forward rates

1.4.1 Term structure of interest rates

Definition 1.4.1: Term structure of interest rates

The term structure of interest rates or yield curve is the set of rates
r(t, T ) for different maturities T . Each r(t, T ) represents the yield on
a ZCB with maturity T .

Remarks:

1. The discount curve is the set of discount factors: Z(t, T ), ∀T

2. A spot rate is a rate for an investment that starts today (t = 0). Hence,
the spot rate curve is the set of rates r(0, T ), ∀T

Example 1.4.1. Consider the following [Link]:

• 90-day maturity, P (0, 0.25) = 99.1234%

1.4. Spot and Forward rates 15

• 180-day maturity, P (0, 0.5) = 98.0199%

• 360-day maturity, P (0, 1) = 95.1229%

The yield curve (with cts comp.) is:

• r∞ (0, 0.25) = . . . = 3.5219%

• r∞ (0, 0.5) = . . .
• r∞ (0, 1) = . . .

1.4.2 Forward discount factors

The forward discount factor defines the time value of money between two
future dates.
Definition 1.4.2: Forward discount factor
Given a set of spot discount factors at current time t, the forward
discount factor between T1 and T2 , with t < T1 < T2 , is given by

Z(T1 , T2 ) : Z(t, T1 ) × Z(T1 , T2 ) = Z(t, T2 )

Use the alternative notation Z(t, T1 , T2 ) when it is necessary to be clear

that it is a discount factor computed at time t, to discount from T2 back to
T1 .

Example 1.4.2. Consider the following [Link]:

• 3-month maturity, P (0, 0.25) = 99.2583%

• 9-month maturity, P (0, 0.75) = 97.0733%

Check that Z(0.25, 0.75) = 0.977987

1.4.3 Forward interest rates

Definition

A forward rate applies to a time period in the future.

1.4. Spot and Forward rates 16

Definition 1.4.3: Forward interest rate

Given a set of spot rates at current time t, the forward rate between
T1 and T2 (t < T1 < T2 ), with compounding frequency n, denoted
rn (T1 , T2 ), is given by
 n.(T1 −t)  n.(T2 −T1 )  n.(T2 −t)
rn (t, T1 ) rn (T1 , T2 ) rn (t, T2 )
1+ 1+ = 1+
n n n

or with continuous compounding,

r∞ (T1 , T2 ) : er∞ (t,T1 ).(T1 −t) er∞ (T1 ,T2 ).(T2 −T1 ) = er∞ (t,T2 ).(T2 −t)

Alternative notations:

1. rn (t, T1 , T2 ), to stress that it is a forward rate computed at time t.

2. fn (T1 , T2 ) or fn (t, T1 , T2 ), sometimes together with sn (t, T ) or sn (t, t, T ),

to stress which ones are forward and which ones are spot rates.
But all this mess should be clear from the relation between t and T1 in
the rn (t, T1 , T2 ) notation: t = T1 means spot, t < T1 means forward.

Example 1.4.3. 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%
Check that r2 (0.25, 0.75) = 4.5018%

The forward rates can be interpreted, as a first approximation, as the

market expectations of future interest rates.4
4
But note that this is not totally correct because bond prices, and thus forward rates,
are also influenced by investors’ degree of risk aversion.
1.5. Fixed coupon bonds 17

Relation to discount factors

The forward rate and the forward discount factor are obviously related. With
discrete compounding,
1
Z(T1 , T2 ) = h in.(T2 −T1 )
rn (T1 ,T2 )
1+ n

and with continuous compounding,

Z(T1 , T2 ) = exp{−r∞ (T1 , T2 ).(T2 − T1 )}

Example 1.4.4. Check that r2 (0.25, 0.75) in example 1.4.3 matches

Z(0.25, 0.75) in example 1.4.2.

The forward curve

If we fix a given point in the future (T1 ) and compute all the forward rates
that start from that point, we get the forward curve for time T1 :

{rn (0, T1 , Ti ), i = 2, 3, . . .}

Example 1.4.5. Using the spot rates in example 1.4.3, compute

the forward curve that starts at T1 = 0.25:
r2 (0.25, 0.50) = . . .
r2 (0.25, 0.75) = 4.5018%
r2 (0.25, 1.00) = . . .

1.5 Fixed coupon bonds

1.5.1 Bond price

Fixed coupon bonds pay a regular constant interest, in addition to the prin-
cipal at maturity.
1.5. Fixed coupon bonds 18

1. Assume that ZCB for all maturities trade in the market.

2. The fixed-coupon bond can be replicated with a portfolio of ZCB.

3. If there are no arbitrage opportunities, the price of the fixed-coupon

bond must equal the price of the replicating portfolio.

Proposition 1.5.1: Price of fixed-coupon bond (B)

m
cX
B(t) = P (t, Ti ) + P (t, Tm )
n i=1
where:

• c is the annual coupon rate with compounding frequency n

• T1 , T2 , . . . , Tm = T are the coupon payment dates

• P (t, T ) = Z(t, T ) are the prices of ZCB

Alternative representations. Given the equivalence between ZCB prices

and interest rates, we can also write:

• Rates with continuous compounding:

m
c X −r∞ (t,Ti )(Ti −t)
B(t) = e + e−r∞ (t,Tm )(Tm −t)
n i=1

• Rates with compounding frequency s:

m
X c/n 1
B(t) = h is×(Ti −t) + h is×(Tm −t)
rs (t,Ti ) rs (t,Tm )
i=1 1+ s
1+ s

Special case. Treasury Notes and Bonds usually pay coupons semi-annually.
The stated coupon rate (c) is always annual with semi-annual compounding
(n = 2).
1.5. Fixed coupon bonds 19

Example 1.5.1. 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%
A 5% [Link] matures in 9 months. Its price is
B(0) = . . . = 101.98%

Note: this example shows why it is simpler to work with discount

factors or continuous rates.
Suggested Homework: convert the rates above to continuous com-
pounding or discount factors and check that you get the same B.

1.5.2 Yield to Maturity

Definition

Definition 1.5.1: Yield to Maturity (YTM)

The YTM is the constant discount rate that makes the present value
of the bond payments equal to its price.
m
X c/n 1
ys : B(t) = +
ys s×(Ti −t) ys s×(Tm −t)
  
i=1 1+ s
1+ s

where:

• ys is the annual YTM with compounding frequency s;

• c is the annual coupon rate with compounding frequency n;

• T1 , T2 , . . . , Tm = T are the coupon payment dates

We can only solve the YTM equation numerically. Use Excel (IRR, Goal
seek, or Solver), Matlab (fminsearch, fzero, etc), or other appropriate soft-
ware.
1.5. Fixed coupon bonds 20

Example 1.5.2. From the previous example, consider a 5% [Link]

with 9-month maturity, trading at B(0) = 101.98%.

1. Using fminsearch in Matlab, we get y1 = 0.040338 (EAR).

Check that this value is correct.
2. The yield on [Link] is most often quoted with the same
frequency as the coupon rate. But given y1 , we can easily
compute
y2 : . . . = 0.039939
3. Alternatively, we could have begun by computing y2 directly.
(Try it at home)

Interpretation of the Yield to Maturity

The YTM is just an alternative way of quoting the bond price.

Interpretation:

• YTM is an “average” of the spot rates.

• The YTM will be the realized yield if the investor:

1. Holds the bond until maturity;

2. Reinvests all coupons at the ytm.

Drawback: the YTM depends on the coupon rate (unfortunate bc the

artificial c is unrelated to the actual market rates r(t, T )).

Example 1.5.3. Consider a 8% [Link] that matures in 9 months.

1. Using the same spot rates as before, we get

B(0) = . . . = 104.93%
2. However, its yield is
y2 = 3.9825% 6= 3.9939%
1.6. Floating rate bonds 21

1.6 Floating rate bonds

1.6.1 Overview
Definition

Definition 1.6.1: Floating-rate bonds

Floating-rate bonds (or floaters) are securities that pay 100% at ma-
turity and make interest payments that are tied to some measure of
current market rates, that is, the coupon rate is defined as

Coupon Rate(c) = Market Index(I) + Spread(s)

While I varies through time, s is fixed.

Examples:

• Annual coupon at LIBOR 12m

• Semi-annual coupon at Euribor 6m + 0.5%

• Quarterly coupon at TBill 3m + 1%

Coupon details

The coupon rate is determined with the index value on the date of the last
coupon payment.

1. Suppose a bond pays annual coupons at LIBOR + 0.3% on the 15th/Jan

of each year.

2. If the LIBOR rate on 15/Jan/2013 equals 1%, then the coupon that
will be paid on 15/Jan/2014 will be 1.3%.
1.6. Floating rate bonds 22

Coupon payments

Let T1 , T2 , . . . , Tm denote the coupon payment (or reset) dates and

T0 the previous coupon date. At any date t before the next coupon,
T0 ≤ t < T1 ,

1. The next coupon in known: c(T1 ) = I(T0 ) + s

˜ i−1 ) + s,
2. All following coupons are unknown: c̃(Ti ) = I(T i=
2, . . . , m

1.6.2 Pricing
Assumptions. Assume the following:

1. The issuer has the same credit risk as the entities that determine the
index.
Examples:
(a) Treasury issues bond indexed to TBill 6m
(b) Bank A issues bond indexed to LIBOR 6m + 0.2%.
2. We have an interest rate curve or discount curve for the corresponding
risk level.
Examples:
(a) r(t, T ) or Z(t, T ) from Government debt
(b) r(t, T ) or Z(t, T ) from LIBOR rates

Case 1: Reset date, zero spread

Proposition 1.6.1: Price of floating rate bond with no spread

at reset date
Let t be any reset date. The ex-coupon price (one instant after the
coupon is paid) of a bond that pays no spread (s = 0) is

B(t) = 100%
1.6. Floating rate bonds 23

Example 1.6.1. Consider a 0.5-year floating rate bond with semi-

annual coupons at LIBOR 6m + 0%. The fixing today is LIBOR
6m = 4%.
B(0) = . . .

Example 1.6.2. Same bond as before, but 1 year to maturity.

Since the bond still has several periods to go, do a backward
induction:

1. Suppose we are at the next to last period, t = 0.5.

• I(0.5) is determined now. Suppose I(0.5) = 6% (any
other value will also work).
• c(1) = I(0.5) = 6% is known now.
• Hence,
1 + 6%/2
B(0.5) = = 100%
1.03
• The point is that the discount rate is the same as the
index!
2. Step back to t = 0.
• In one period we will have an asset worth 100% plus a
known c(0.5) = I(0)
• Hence,
1 + 4%/2
B(0) = = 100%
1.02

Case 2: Between coupon dates, zero spread

At any date t between coupon payments, T0 ≤ t < T1 , the floating rate bond
with c = I is the sum of 2 assets:

1. A floating rate bond paying coupons at dates T2 , T3 , . . . , Tm . This will

have an ex-coupon price of 100% at T1 (case 1).

2. A known coupon rate c(T1 ) = I(T0 ) to be received at T1 .

1.6. Floating rate bonds 24

Proposition 1.6.2: Price of floating rate bond with no spread

Consider a floater with coupon rate c(Ti ) = I(Ti−1 ), with frequency n.

The price at any date t before the next coupon, T0 ≤ t < T1 , is
 
I(T0 )
B(t) = 100% + Z(t, T1 )
n

Example 1.6.3. Consider a 1.5-year floating rate bond with an-

nual coupons at LIBOR 12m. The fixing 6 months ago was LI-
BOR 12m = 4%.
The term structure today (t=0) is
Ti (yr) r2 (0, Ti )
0.5 4.00%
1.0 4.50%
1.5 4.80%
B(0) = . . . = 101.9608%

Case 3: Between coupon dates, non-zero spread

At any date t between coupon payments, T0 ≤ t < T1 , the floating rate bond
with c = I + s is the sum of 3 assets:

1. A floating rate bond with coupon rate I. This will have an ex-coupon
price of 100% at T1 (case 1).
2. A known coupon rate c(T1 ) = I(T0 ) + s to be received at T1 .
3. A stream of fixed coupon rates s to be received at dates T2 , T3 , . . . , Tm .

Proposition 1.6.3: Price of floating rate bond (general case)

Consider a floater with coupon rate c(Ti ) = I(Ti−1 ) + s, with frequency

n. The price at any date t before the next coupon, T0 ≤ t < T1 , is
  m
I(T0 ) + s X s
B(t) = 100% + Z(t, T1 ) + Z(t, Ti )
n i=2
n
1.7. Exercises 25

Example 1.6.4. Consider a 1.25-year floating rate bond with

semi-annual coupons at LIBOR 6m + 1%. The fixing 3 months
ago was LIBOR 6m = 4%.
The term structure today (t=0) is
Ti (yr) r2 (0, Ti )
0.25 4.30%
0.75 4.70%
1.25 5.00%
B(0) = . . .

Remark: Note that if we are at the reset date (t = T0 ), then the index
equals the first discount rate, I(T0 ) = rn (t, T1 ), and the floater price is simply
m
X s
B(t) = 100% + Z(t, Ti )
i=1
n

(note that the summation now starts at T1 )

1.7 Exercises
Ex. 1 — Consider the following Treasury Bills:
[Link] Maturity (days) Discount
a 28 2.15%
b 150 2.81%
1. Compute the price of each [Link].
2. Compute the spot rates. Show both spot rates with continuous com-
pounding and spot rates with semi-annual compounding.
Note: convert from days to years by dividing the number of days by
360.

Ex. 2 — Consider the following rates:

T r2 (0, T ) T r2 (0, T ) T r2 (0, T )
0.25 3.00% 1.25 4.00% 2.25 4.70%
0.50 3.25% 1.50 4.20% 2.50 4.80%
0.75 3.50% 1.75 4.40% 2.75 4.90%
1.00 3.75% 2.00 4.60% 3.00 5.00%
Price the following securities:
1.7. Exercises 26

•0.5-year zero coupon bond.

•2.5-year coupon bond paying 6% annually.
•1.25-year coupon bond paying 4% quarterly.
•1.5-year floating rate bond with 20 basis-point spread with semiannual
payments.
•2.25-year floating rate bond with 30 basis-point spread with semian-
nual payments. The index was 2.50% (semiannual compounding) at
the previous reset date (t = −0.25).

Ex. 3 — Compute the continuously compounded Yield to Maturity for a

1.35-year coupon bond paying 5% semiannually, trading at 102.45%.

Ex. 4 — The term structure today (t=0) is

T (years) r2 (0, T )
0.25 3.00%
0.50 3.50%
0.75 4.00%
1.00 4.20%
Consider the following Treasury Bonds, whose prices are consistent with the
rates above:
[Link] Coupon Rate (c2 ) Maturity (years) Price YTM (y2 )
A 5% 0.75 101.98% 3.9917%
B 8% 0.75 104.93% 3.9870%

1. For each bond, assume that you hold it until maturity and that you
are able to reinvest the coupon at the YTM of that same bond. Check
that the realized rate of return at maturity (T = 0.75) does indeed
equal the YTM computed today.
2. Compute the forward rate f2 (0, 0.25, 0.75)
3. Now assume instead that you are able to reinvest the coupon at the
actual rate that will be available on the market in the future, and
that this rate will be equal to the forward rate computed today, that
is, f2 (0, 0.25, 0.75) = s2 (0.25, 0.25, 0.75). Compute the realized rate of
return for each bond.
Chapter 2

Yield curve fitting

2.1 Introduction
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 ).
• 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 i=1 n i=1

27
2.2. Bootstrap 28

2.2 Bootstrap

2.2.1 Method
Start from short-maturity ZCB and work your way up to long-term coupon
bonds.
Definition 2.2.1: 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 2.1.
2.2. Bootstrap 29

Figure 2.1: Interest rates vs discount factors

2.2.2 Application

Example 2.2.1. 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

2.2. Bootstrap 30

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

interpolate 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%
2.3. Nelson-Siegel Model 31

2.3 Nelson-Siegel Model

2.3.1 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 a convenient parametric function for the continuously compounded
spot interest rate:

Definition 2.3.1: 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.

2.3.2 Meaning of the parameters

The optimization procedure is sensitive to the initial guesses. It is thus
important to understand the meaning of the parameters.
2.3. Nelson-Siegel Model 32

• The instantaneous short rate is:

lim r(T ) = β1 + β2
T →0

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 2.2: β1 determines the level

β1 = see legend, β2 = 0, β3 = 0, λ = 0
2.3. Nelson-Siegel Model 33

Figure 2.3: β2 determines the slope

β1 = 0.05, β2 = see legend, β3 = 0, λ = 2

Figure 2.4: β3 determines the curvature

β1 = 0.05, β2 = 0, β3 = see legend, λ = 2
2.3. Nelson-Siegel Model 34

Figure 2.5: λ determines the position of the hump

β1 = 0.05, β2 = −0.02, β3 = 0.1, λ = see legend

2.3.3 Application

Example 2.3.1. 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%
2.4. Svensson Model 35

2.4 Svensson Model

2.4.1 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
[Link]

Definition 2.4.1: 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
 
−T /λ2
+ β4 −e
T /λ2

where β1 , β2 , β3 , β4 , λ1 , λ2 are parameters to be estimated.

2.4.2 Aplication

Example 2.4.1. 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.
2.5. Conclusion 36

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%

Figure 2.6: Svensson vs Nelson-Siegel vs Bootstrap vs YTM

2.5 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.

2.5. Conclusion 37

• Numerical optimization is sensitive to initial guesses for the pa-

rameters.

3. Svensson model

• More flexible function than NS.

• Even more parameters to fine-tune than NS.
Chapter 3

Bond portfolio management

3.1 The variation in interest rates

Go to the following site and watch the yield curve move through time.

• [Link]

Note how:

1. Over short periods of time, all yields move up or down roughly together.
We see small parallel shifts in the curve.

2. Over longer periods of time, there are additional patterns, namely, the
short rate varies more than the long rate (slope effect) and there are
also changes in the curvature.

3.2 Duration
Veronesi (2010) is a good reference for this section.

3.2.1 Definition
To simplify the notation, let r(t, T ) denote the continuously compounded
interest rate (aka, r∞ (t, T )). Decompose the term structure r(t, T ) into

r(t, T ) = r̂(t, T ) + r, ∀T (3.1)

38
3.2. Duration 39

where r is a given number, common to all maturities.

Duration is the sensitivity of a security price to a change in the level of
interest rates, i.e., to a change in the common quantity r.
Definition 3.2.1: Duration
The Duration of a security with price P is
1 dP
D := − (3.2)
P dr
where r is as defined in (3.1).

Interpretation. Consider a small (infinitesimal) parallel shift of size dr in

the whole term structure:
r(t, T ) → r ⋆ (t, T ) = r(t, T ) + dr = r̂(t, T ) + r + dr, ∀T
As a consequence of the shift in rates, the price of the security changes to
P → P ⋆ = P + dP
Given the duration, we can replace dP to get
P → P ⋆ = P − D × P × dr (3.3)
Remark. In some textbooks this duration is referred as “Modified” duration.

3.2.2 Duration of a zero coupon bond

Proposition 3.2.1: Duration of ZCB

The duration of a zero coupon bond with price P (t, T ) is

DZCB = T − t

Proof. The price is P (t, T ) = exp{−r(t, T )(T − t)} = exp{−[r̂(t, T ) + r](T −

t)}. Hence,
1 dP
DZCB = −
P dr
1
= − {−(T − t)}P
P
=T −t
3.2. Duration 40

Example 3.2.1. A portfolio manager purchased 3-year STRIPS

with a market value of 10 M$. By how much will the portfolio
value change if interest rates increase by 1 basis point?
∆P ≈ −D × P × ∆r = . . . = −3 000 $
Note that this equation applies to either P as the price of the
individual security (in percentage of face value) or to P replaced
by the market value of our total investment (like we did here).1

Note that for large (discrete) shifts, duration gives just an approximation
to the price change:

P → P ⋆ = P + ∆P ≈ P − D × P × ∆r

P⋆
✻

✲
r⋆

1
Formally, let N denote the quantity of the security and V the total market value.
∂(P N )
Then, dV = ∂V
∂r dr = ∂r dr = ∂P
∂r N dr = −DP N dr = −DV dr
3.2. Duration 41

3.2.3 Duration of a portfolio

Proposition 3.2.2: Duration of a portfolio

The duration of a portfolio of m securities is
m
X
Dp = w i Di (3.4)
i=1

where:

• Di is the duration of security i,

Ni P i
• wi = V
is the weight of security i

• Pi is the price of security i

• Ni is the quantity of security i (number of units or face value,

consistent with the units of P )
P
• V = i Ni Pi is the value of the porfolio

Proof. From the definition of duration,

" #
1 dV 1 d X
Dp = − =− Ni Pi
V dr V dr i
" #
1 X dPi 1 X X Ni Pi
=− Ni = Ni Pi Di = Di
V i
dr V i i
V

Example 3.2.2. A portfolio manager has 10 M$ invested in 3-

year STRIPS and 20 M$ invested in 7-year STRIPS. By how much
will the portfolio value change if interest rates increase by 1 basis
point?
First, compute the portfolio duration,
Dp = . . . = 5.(6)
Second, the approximate portfolio value change is:
∆V ≈ . . . = −17 000 $
3.2. Duration 42

3.2.4 Duration of a fixed coupon bond

Proposition 3.2.3: Duration of a fixed coupon bond

Consider a fixed coupon, or “Coupon Bearing Bond” (CBB), with price

B, paying a coupon rate c with compounding frequency n, at coupon
payment dates T1 , T2 , . . . , Tm . Its duration is:
m
X
DCBB = wi × (Ti − t)
i=1

where:
c Z(t,Ti )
• wi = n B
, for i = 1, . . . , m − 1,

• wm = (1 + nc ) Z(t,T
B
m)

• Z(t, Ti ) is the discount factor

Proof. A CCB is a portfolio of ZCB.

Example 3.2.3. A portfolio manager has 10 M$ invested in a

1-year Treasury Bond paying 4% semiannually. By how much
will the portfolio value change if interest rates increase by 1 basis
point?
The spot rates are r∞ (0, 0.5) = 1% and r∞ (0, 1) = 1.2%
First, compute the bond duration,
B = . . . = 1.027733
DCBB = . . . = 0.990318
Second, the approximate value change is:
∆B ≈ . . . = −990 $

3.2.5 Duration of a floating rate bond

3.2. Duration 43

Proposition 3.2.4: Duration of a floating rate bond without

spread

Consider a floating rate bond paying a coupon rate c(Ti ) = I(Ti−1 ),

with frequency n, at dates T1 , . . . , Tm , as a function of some index I.
Let T0 denote the last and T1 the next reset dates. Its duration at time
t, with T0 ≤ t < T1 is
DFRB = T1 − t

Proof. At time t, the floating rate bond is equivalent to a zero coupon bond
paying 100%+ I(Tn0 ) at time T1 . Hence, its duration is the same as the duration
of this ZCB.

Example 3.2.4. What is the duration of a 1.4-year floating rate

bond paying annual coupons at LIBOR 12m?
DFRB = . . .

Proposition 3.2.5: Duration of a floating rate bond with

spread

Consider a floating rate bond paying a coupon rate c(Ti ) = I(Ti−1 ) + s,

with frequency n, at dates T1 , . . . , Tm , as a function of some index I,
with a constant spread s. Let T0 denote the last and T1 the next reset
dates. Its duration at time t, with T0 ≤ t < T1 is
m
X
DFRB = w1 .(T1 − t) + wi (Ti − t)
i=2

where:
 
I(T0 )+s Z(t,T1 )
• w1 = 1 + n B

s

Z(t,Ti )
• wi = n B
, for i = 2, . . . , m

• Z(t, Ti ) is the discount factor

• B is the floater price at time t

3.2. Duration 44

Proof. At time t, the floating rate bond is equivalent

 to a portfolio of zero
I(T0 )+s
coupon bonds. The first ZCB pays 1 + n at time T1 . The other ZCB
s


pay n at times T2 , . . . , Tm . Hence, the duration of the floater equals the
duration of this portfolio of ZCB (proposition 3.2.2).
Alternatively, just for fun, we can check this formula from the basic defi-
nition. Start from the floater price:
  m
I(T0 ) + s X s
B(t) = 1 + Z(t, T1 ) + Z(t, Ti )
n i=2
n
 
I(T0 ) + s
= 1+ exp{−[r̂(t, T1 ) + r](T1 − t)}
n
m
X s
+ exp{−[r̂(t, Ti ) + r](Ti − t)}
i=2
n

Now compute the duration using the definition:

1 dB
DFRB = −
B "dr #
  m
1 I(T0 ) + s X s
=− 1+ (−(T1 − t)) exp{.} + (−(Ti − t)) exp{.}
B n i=2
n
  m
I(T0 ) + s Z(t, T1 ) X s Z(t, Ti )
= 1+ (T1 − t) + (Ti − t)
n B i=2
n B

which gives the expected result.

Example 3.2.5. Compute the duration of a 2-year floating rate

bond paying annual coupons at LIBOR 12m + 0.5%. Assume
that we are one instant after the last coupon was paid. Assume
that the term structure is flat at r1 (0, T ) = 4%, ∀T
First, compute the price,
B(0) = . . . = 100.94%
Next, compute the duration:
DFRB = . . . = 1.0046

The next example checks that this duration gives a good approximation
to the bond price.
3.2. Duration 45

Example 3.2.6. Continuing the previous example, assume that

interest rates decrease by 1% in the next instant, that is, r1 (0 +
dt, T ) = 3%, ∀T .
1) Estimate the price change with the duration formula:
DFRB ≈ − B1 ∆B
∆r
⇒ ∆B ≈ . . .
B(0 + dt) = B(0) + ∆B ≈ . . . = 101.95%

2) Check that this is a good approximation. Compute the true

price for the new rates:
B(0 + dt) = . . . = 101.93%

Remark. One traditional interpretation of duration is the average time of

payments. However, that interpretation only holds for bonds that pay fixed
coupons. As this section shows, a floating rate bond with, say, 5 years to
maturity can have a duration of nearly zero if we are 1 day before the next
coupon. For floating rate bonds, duration and average time of payments
may be very different quantities. Hence, we should keep with the more
useful interpretation of duration as the percentage sensitivity of a security
to changes in interest rates.

3.2.6 Traditional definitions of duration

In (3.2) we defined duration with the full term structure of interest rates
(with continuous compounding). An older approach is to define duration
against the bond’s own YTM (typically with semiannual compounding, y2 ).
By the definition of YTM, the price can be written as
m−1 c
X
n
1 + nc
B= +
i=1
(1 + y22 )2Ti (1 + y22 )2Tm

Then, it can be shown that the modified duration is

1 dB D Mc
D mod = − =
B dy2 1 + y22

where:
Pm
D Mc = i=1 wi Ti and is referred as MacCaulay Duration,
c
1
wi = n
y2 2T
(1+ 2 ) i B
, for i = 1, . . . , m − 1,
3.3. Active bond management 46

1+ nc 1
wm = y
(1+ 22 )2Tm B

We will rarely use MacCaulay’s duration for risk management purposes

due to the following issues with this measure:

1. It cannot be computed for securities like floating rate bonds where the
YTM is not well defined.

2. The YTM is bond specific and depends on the coupon rate (see the
examples in the previous chapter). Hence, the concept of a parallel
shift in yields to maturity is ambiguous. In contrast, a parallel shift in
spot rates is well defined and has a precise effect on the price of each
security.

3. It does not aggregate well: the MacCaulay duration of a portfolio has

to be computed from the cash flows and YTM of the portfolio itself.
There is no direct relation to the durations and YTM of the individual
bonds.

4. The formulas become more cumbersome.

3.3 Active bond management

If a manager is able to forecast unanticipated interest rate movements better
than the rest of the market, then equation (3.3) suggests an investment rule:

• Increase the portfolio duration if the manager expects interest rates to

decrease;

• Decrease the portfolio duration if the manager expects interest rates to

increase;

This rule works under the following assumptions:

1. Instantaneous parallel shifts in the term structure. In more general

setups — any type of interest rate changes at any point in time — the
best way to decide which bonds to buy is to simulate their total return
for the investment period.

2. The manager has expectations different from the forward rates. If rates
evolve according to forward rates, all bonds produce exactly the same
return.
3.4. Passive bond management: Immunization 47

3.4 Passive bond management: Immunization

Some financial institutions (e.g., insurance companies) sell contracts (e.g.,
annuities) that commit them to a stream of fixed payments in the future.
They 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 CBB leaves the firm exposed to price risk:

loose money if interest rates go up.
• Investing in short-term revolving deposits or FRB leaves the firm
exposed 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 elimi-
nates interest rate risk.
• Allows the institution to choose bonds with good liquidity and
lower transaction costs.

Immunization strategy
Goal: guarantee a given cash flow by a due date, regardless of what
happens to interest rates until then.
Implementation:

1. PV assets = PV liabilities

2. Duration assets = Duration liabilities

3.4. Passive bond management: Immunization 48

Example 3.4.1. 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

An immunization strategy will work under the following assumptions:

1. Interest rates shift in an infinitesimal parallel way:

r(0, T )⋆ = r(0, T ) + dr

2. This term structure shift from r(0, T ) to r(0, T )⋆ happens immediately

after setting up the portfolio.

3. Afterwards, the forward rates implied in r(0, T )⋆ will become the future
spot rates.
3.4. Passive bond management: Immunization 49

Example 3.4.2. Continuing the previous example, suppose that

interest 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.
Check that the terminal value of the assets exceeds the liability:
VA (2) = . . . = 1 000 008 > VL (2) = 1 000 000

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 invest-
ment grows at a faster rate in the future.
2. If interest rates decrease, we have a capital gain today, but the invest-
ment 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)
✻

✲
D t

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⋆
3.5. Duration of zero investment portfolios 50

Note that in this particular example the terminal value of the assets always
exceeds the liability. This is because the convexity of the assets exceeds the
convexity of the liabilities, as discussed in section 3.7.4.
Remark. In practice, we need to rebalance the portfolio frequently to adjust
for interest rate changes. There is a tradeoff between higher transaction costs
and a more effective immunization. (Example 3.10 in Veronesi (2010) shows
this dynamic immunization strategy).

3.5 Duration of zero investment portfolios

The duration of a portfolio defined in equation (3.4) is only applicable to
portfolios with non-zero prices (because the value of the portfolio is in the
denominator). To compute the duration of zero-investment long-short port-
folios, we need an alternative measure of duration.

Definition 3.5.1: Dollar Duration

The Dollar Duration of a security with price P is
dP
D $ := − (3.5)
dr
where r is as defined in (3.1).

Note that for a security with positive price, dollar duration is related to
the basic duration by
D$ = P × D (3.6)

The designation “Dollar” duration is a bit misleading when the price is

not quoted in dollars. The units are:

• Price quoted in dollars:

– D is in years
– D $ is in “dollars × years”

• Price quoted in (dimensionless) percentage of par value:

– D is in years
– D $ is in years
3.5. Duration of zero investment portfolios 51

Nevertheless, the portfolio formula below always aggregates to “dollars ×

years”.

Proposition 3.5.1: Dollar Duration of a portfolio

The Dollar duration of a portfolio of m securities is
m
X
Dp$ = Ni Di$ (3.7)
i=1

where:

• Di$ is the dollar duration of security i,

• Ni is the quantity of security i (number of units if prices are

quoted in $; face value in $ if prices are quoted in % of par)

P
Proof. Let V = i Ni Pi denote the value of the porfolio. From the definition
of dollar duration,
" #
dV d X X dPi X
Dp$ = − =− Ni Pi = − Ni = Ni Di$
dr dr i i
dr i

Remark. Note that when prices are quoted in % of FV, formula (3.7) is
m
X
Dp$ (in $.y) = Ni (in $) × Pi (in %) × Di (in y) (3.8)
i=1
m
X
= MarketValuei (in $) × Di (in y) (3.9)
i=1

Further note that an investment in one single bond is a portfolio with m = 1.

Example 3.5.1. (This is adapted from example 3.6 in Veronesi

(2010), though some values in the book are wrong.)
The term structure is flat at r2 (0, T ) = 4%, ∀T ⇔ r∞ (0, T ) =
3.96%, ∀T . An arbitrageur is considering the following trade:
1) LONG position.
3.6. Asset-Liability management 52

10-y [Link], paying 4% semiannually. Given the rates above we

can compute
Pcbb = 100%
Dcbb = 8.34 y
Therefore, using (3.6)
$
Dcbb = . . . = 8.34 y
2) SHORT position.
To fund the long position, the trader will use the bond as collat-
eral in a 6-month term repo (assume a zero haircut). Borrowing
through a 6-month repo is essentially equivalent to shorting a
6-month floating rate bond. This FRB has
Df rb = . . .
Pf rb = . . ., and thus
Df$rb = . . . = 0.5 y
3) Combined total portfolio
The portfolio has zero value at time 0. However, this portfolio is
not risk free. If interest rates go up, the loss in the long position
is higher than the gain in the short position.
Suppose the trader is considering buying 10 M$ of face value.
The dollar duration of the portfolio is:
Dp$ = . . . = 78.4 M$y
Hence, if interest rates go up by 1 bp, the portfolio value will
change by:
∆V ≈ −Dp$ ∆r = . . . = −$ 7 840

3.6 Asset-Liability management

Many financial institutions have a mismatch between their assets and lia-
bilities. For example, commercial banks typically have short-term liabilities
(e.g., deposits) and long-term assets (e.g., mortgage loans). The higher du-
ration of the assets relative to the duration of the liabilities leaves the bank’s
equity exposed to interest rate risk (equity drops if interest rates rise).
3.6. Asset-Liability management 53

Asset-Liability Management
Goal: minimize the exposure of the firm’s Equity to interest rate risk.
Implementation:

1. Look at the balance sheet as a long-short portfolio of Assets (A)

and Liabilities (L) to estimate the duration of Equity (E):

DA$ = DE$ + DL$ ⇒ DE$ = DA$ − DL$ (3.10)

2. Adjust the composition of the Assets and/or Liabilities to achieve

DE$ = 0 (3.11)

Example 3.6.1. (This is example 3.11 in Veronesi (2010)). Con-

sider the following balance sheet of a commercial bank. The basic
duration (D) of each item has already been computed.

Assets Mkt Value (M$) D (y) Liabilities Mkt Value (M$) D (y)
Cash 100 0 Deposits 600 0
S.T. Loans 300 0.8 S.T. Debt 400 0.5
M.T. Loans 500 3 M.T. Debt 400 4
L.T. Loans 1500 12 L.T. Debt 400 8
Total 2400 Total 1800

Equity 600

Estimate the Equity’s exposure to interest rate risk.

1. Use (3.9) to compute the Dollar Duration for each item.

$
For example, DST Loans = Mkt ValueST Loans ×DST Loans = . . .
...
...

And then use (3.7) to compute the total dollar duration of

A and L.
3.7. Convexity 54

The final result is:

Assets D $ (M$y) Liabilities D $ (M$y)
Cash Deposits
S.T. Loans S.T. Debt
M.T. Loans M.T. Debt
L.T. Loans L.T. Debt
Total 19 740 Total 5 000

2. Compute the dollar duration of Equity:

DE$ = DA$ − DL$ = . . .
3. Since DE$ > 0, the bank is exposed to interest rate risk. For
example, suppose that there is an upward parallel shift in
interest rates of 1% per year. The change in the value of
Equity is
DE$ ≈ − ∆E
∆r
⇒ ∆E ≈ . . . = −147.4 M$
A huge drop in Equity value relative to its current value of
600 M$!

To manage this risk, the bank could try to replace some of the
short-term debt with more long-term debt, in order to increase
the duration of the liabilities.

3.7 Convexity

3.7.1 Definition
Consider the same decomposition of the term structure as in (3.1).

Definition 3.7.1: Convexity

The Convexity of a security with price P is

1 d2 P
C := (3.12)
P dr 2
where r is as defined in (3.1).
3.7. Convexity 55

Convexity is useful to approximate price changes for large interest rate

movements. Consider a discrete parallel shift of size ∆r in the whole term
structure:
r(t, T ) → r ⋆ (t, T ) = r(t, T ) + ∆r = r̂(t, T ) + r + ∆r, ∀T
As a consequence of the shift in rates, the price of the security changes to
P → P ⋆ = P + ∆P

Since P is not linear in r, approximate the new price with a Taylor ex-
pansion around the current price:
dP 1 d2 P
P⋆ = P + ∆r + (∆r)2 + . . . (3.13)
dr 2 dr 2
∆P 1
⇒ ≈ −D∆r + C.(∆r)2 (3.14)
P 2

Remark. This expansion shows that:

1. Duration may have a positive or a negative effect on the value of our
portfolio, depending on whether rates move up or down;
2. Convexity, on the contrary, always contributes positively to the value
change of the portfolio, regardless of the direction of the rate change.
All else equal, we would always prefer more convexity!

3.7.2 Convexity of a zero coupon bond

Proposition 3.7.1: Convexity of ZCB

The Convexity of a zero coupon bond with price P (t, T ) is

CZCB = (T − t)2

Proof. The price is P (t, T ) = exp{−r(t, T )(T − t)} = exp{−[r̂(t, T ) + r](T −

t)}. Hence,
1 d dP
CZCB =
P dr dr
1 d
=− [(T − t)P ]
P dr
= (T − t)2
3.7. Convexity 56

Example 3.7.1. Consider a 20-year ZCB. Assume the spot curve

changes from 5% to 3% (cts compounding).
1) Estimate the new price using only duration.
P ⋆ ≈ P + ∆P = . . . = 0.515031
2) Estimate the new price using both duration and convexity.
P ⋆ ≈ P + ∆P = . . . = 0.544462
3) Compare these estimates with the new true price.
P true = . . . = 0.548812

3.7.3 Convexity of a portfolio

Proposition 3.7.2: Convexity of a portfolio

The convexity of a portfolio of m securities is
m
X
Cp = w i Ci (3.15)
i=1

where:

• Ci is the convexity of security i,

Ni P i
• wi = V
is the weight of security i

• Pi is the price of security i

• Ni is the quantity of security i (number of units or face value,

consistent with the units of P )
P
• V = i Ni Pi is the value of the porfolio

Proof. Start from the definition and take derivatives (similar to previous
proofs).

3.7.4 Convexity of a fixed coupon bond

3.7. Convexity 57

Proposition 3.7.3: Convexity of a fixed coupon bond

Consider a fixed coupon, or “Coupon Bearing Bond” (CBB), with price

B, paying a coupon rate c with compounding frequency n, at coupon
payment dates T1 , T2 , . . . , Tm . Its convexity is:
m
X
CCBB = wi (Ti − t)2
i=1

where:
c Z(t,Ti )
• wi = n B
, for i = 1, . . . , m − 1,

• wm = (1 + nc ) Z(t,T
B
m)

• Z(t, Ti ) is the discount factor

Proof. A CCB is a portfolio of ZCB.

3.7.5 Application to immunization

Immunization can work even for large discrete rate changes,

r(0, T )⋆ = r(0, T ) + ∆r

if the convexity of the assets is larger than the liabilities. Since the durations
are matched, the higher convexity of the assets guarantees a larger increase
for any interest rate change. Plot the portfolio values immediately after the
rate change to illustrate the effect of convexity:

V ⋆ (0)
✻

✲
r r⋆
3.7. Convexity 58

If there are no further interest rate changes, we are guaranteed to have

VA⋆ (D) > VL⋆ (D).

Example 3.7.2. Continuing with example 3.4.1,

1) Compare the convexities:
CL = . . .
CZCB = . . .
CCCB = . . .
Hence, CA = . . . = 4.9757 > CL = 4
2) Now suppose that there is a large rate change immediately
after we invest in the two bonds: spot rates increase by 3% and
then remain at r∞ (0, T ) = 8%, ∀T , for the next two years.
Check that the terminal value of the assets still exceeds the lia-
bility:
VA (2) = . . . = 1 000 439 > VL (2) = 1 000 000
As expected, this strategy still worked for this large discrete rate
change!
Chapter 4

Interest rate derivatives

4.1 Forward rate agreements

4.1.1 Definition

Definition 4.1.1: Forward Rate Agreement (FRA)

A Forward Rate Agreement (FRA) is an OTC contract between two

counterparties whereby:

1. One counterparty (“fixed rate payer”) agrees to pay a fixed rate

Fn (0, T1 , T2 ), on a notional N, over a future period from T1 to
T2 .

2. The other counterparty (“floating rate payer”) agrees to pay a

variable rate In (T1 , T2 ) over the same amount and period.

where:

• Fn (0, T1 , T2 ) is the fixed rate defined at the beginning of the con-

tract (t = 0)

• In (T1 , T2 ) is a money market index, typically the LIBOR, which

will be observed in the market at future time T1

• Both rates are expressed with a compounding frequency n cor-

responding to the length of the period in the contract, that is,
1/n = T2 − T1

59
4.1. Forward rate agreements 60

Settlement. Only the net value of the two legs exchanges hands. The
contract may specify that settlement happens:

• At T2 . The net payment is:

 
In (T1 , T2 ) Fn (0, T1 , T2 )
CF to fixed payer at T2 = N × −
n n

• At T1 (more common). The net payment is the same amount, but

discounted back to T1 :
h i
N × In (Tn1 ,T2 ) − Fn (0,Tn 1 ,T2 )
CF to fixed payer at T1 =
1 + In (Tn1 ,T2 )

These two alternatives are equivalent in the sense that, whatever the FRA
specifies, one of the counterparties can always do unilateral trades in the
market at T1 (borrowing or lending) to move the cash flow from T1 to T2 or
vice-versa.

Example 4.1.1. Consider the following FRA:

• Notional = 10 M$
• Start in 6 months, end in 9 months.
• Bank pays F4 (0, 0.5, 0.75) = 2%
• Firm pays LIBOR 3M
• Settlement in 6 months.

Suppose that in 6 months, LIBOR 3M is at 1.7%. Then,

CF to fixed payer (bank) at 0.5 = . . . = −$ 7 468

4.1.2 Hedging with a FRA

The goal of a FRA is to guarantee a certain interest rate for a future loan
or investment. The assumption underlying the FRA is that borrowing or
lending could alternatively be done at the (uncertain) LIBOR.
4.1. Forward rate agreements 61

Example 4.1.2. A firm has a receivable of 10 M$ in 6 months.

The firm wishes to invest this amount for an additional 3 months.
If the firm enters into the FRA of the previous example (as fixed-
rate receiver), the firm is guaranteed to get 2% on its investment.
Check:
1) Suppose again that in 6 months, LIBOR 3M is at 1.7%. From
the previous example, the firm will receive $ 7 468 from the bank.
2) The total amount that the firm has to invest at T1 is
V (0.5) = . . .
3) Assuming the firm can invest at LIBOR, the terminal value
will be
V (0.75) = . . .
4) The effective rate obtained by the firm is
V (0.75)−10 M $
10 M $
× 4 = . . . = 2%

4.1.3 The value of an existing FRA

At the initial moment t = 0, the price of the FRA is zero and no payments
are exchanged. However, as time passes and forward rates change in the
market, the value of the FRA also changes.

Proposition 4.1.1: Value of a FRA

At time t after inception, 0 < t < T1 , the value of a FRA for the
counterparty receiving the fixed rate is
 
FRA Fn (0, T1 , T2 )
[Link]. (t) = N × 1 + Z(t, T2 ) − NZ(t, T1 ) (4.1)
n

where:

• Fn (0, T1 , T2 ) is the fixed rate defined at the beginning of the con-

tract

• Z(t, T ) is the spot discount factor implied in the current money

market rates.

• N is the notional
4.1. Forward rate agreements 62

Proof. Being in a FRA as fixed rate receiver and floating rate payer is equiv-
alent to the sum of two positions:
h i
1. Long a fixed rate bond paying N × 1 + Fn (0,Tn 1 ,T2 ) at T2 . Its value at
t is  
fixed leg Fn (0, T1 , T2 )
V (t) = N × 1 + Z(t, T2 )
n
h i
2. Short a floating rate bond paying N × 1 + In (Tn1 ,T2 ) at T2 . Its value is
100% at T1 . Therefore,

V floating leg (t) = NZ(t, T1 )

The value of the FRA today is thus

FRA
[Link]. (t) = V fixed leg (t) − V floating leg (t)

Example 4.1.3. Continuing the previous example, 5 months have

passed and the LIBOR rates are now the following:

T r∞ (t, T )
1/12 3%
4/12 4%

The value of the FRA to the firm is

V FRA (t) = . . .
This means that if the firm no longer needs the FRA and wishes
to terminate it, the firm has to pay $58 142 to the bank to cancel
the FRA.
4.1. Forward rate agreements 63

4.1.4 Setting the fixed rate in a FRA

Proposition 4.1.2: The initial fixed rate in a FRA

Assume that a large financial institution can borrow and lend at the
same money market rates. This institution can quote a FRA with

Fn (0, T1 , T2 ) = rn (0, T1 , T2 )

where:

• Fn (0, T1 , T2 ) is the fixed rate quoted in the FRA.

• rn (0, T1 , T2 ) is the forward rate implied in the money market zero

curve (typically, the LIBOR/swap zero curve).

Proof. The FRA rate is set such that the FRA value is zero for both coun-
terparties at the start of the contract (ie, there is no initial cash flow). Using
(4.1),
 
FRA Fn (0, T1 , T2 ) Z(t, T1 )
[Link]. (t) = 0 ⇒ 1 + =
n Z(t, T2 )
 
Fn (0, T1 , T2 ) 1
⇒ 1+ =
n Z(t, T1 , T2 )
   
Fn (0, T1 , T2 ) rn (0, T1 , T2 )
⇒ 1+ = 1+
n n
Hence, the FRA rate should equal the forward LIBOR, rn (0, T1 , T2 ).

Example 4.1.4. The money market rates (mid quotes) are the
following:

Money market Rate

1 month (r12 (0, 1/12) 1.00%
3 months (r4 (0, 3/12)) 1.20%

The fair quote on a FRA 1x3 is

F6 (0, 1/12, 3/12) = . . . = 1.2989%
4.1. Forward rate agreements 64

Hedging a FRA quote

An instructive alternative way of determining the FRA rate is to analyze the

hedging strategy of the bank that is quoting the FRA (the market maker).
It is useful to think of the FRA as if the notional (N) is actually exchanged
at T1 and notional plus interest at T2 .
Case 1: the bank pays fixed in the FRA.

• The bank is committed to the following CF in the FRA:

CF to bank time t = 0 time T1  time T2 
Fn (0,T1 ,T2 )
Pays fixed 0 +N −N 1 + n
 
In (T1 ,T2 )
Receives floating 0 −N +N 1 + n

• Hedging strategy

– At t = 0. Define zero-cost trades to hedge the fixed leg:

1) Borrow NZ(t, T1 ) for T1 years.
2) Invest the same amount, NZ(t, T1 ), for T2 years.
– At T1 . Trade to hedge the floating leg:
3) Borrow N for T2 − T1 years.
Note: in reality, there is no exchange of principal in the FRA, so
we can think of this loan in (3) as a roll-over of the loan in (1).
– Alternative procedure: think of the trade that will hedge the float-
ing leg, and then work backwards, ie, do 3), 1), 2).

• Net outcome for the bank:

CF to bank time t = 0 time T1  time T2 
Fn (0,T1 ,T2 )
FRA fixed 0 +N −N 1 + n
(1) +NZ(t, T1 ) −N
(2) −NZ(t, T1 ) +NZ(t,  T1 )/Z(t, T2)
FRA floating 0 −N +N 1 + In (Tn1 ,T2 )
 
In (T1 ,T2 )
(3) +N −N 1 + n
 
Z(t,T1 ) Fn (0,T1 ,T2 )
Total 0 0 N Z(t,T2 ) − 1 − n
4.1. Forward rate agreements 65

The bank will be perfectly hedged if the total CF at T2 is zero:

Fn (0, T1 , T2 ) Z(t, T1 )
= −1
n Z(t, T2 )
1
= −1
Z(t, T1 , T2 )
 
rn (t, T1 , T2 )
= 1+ −1
n
rn (t, T1 , T2 )
=
n
• Hence, the bank will quote a FRA with a rate equal to the forward
LIBOR, rn (0, T1 , T2 ).

Case 2: the bank receives fixed in the FRA. Similar to case 1, changing
the hedging trades appropriately.

4.1.5 Exercises
Ex. 5 — Six months from now, a firm will need to borrow 50 million dollars
for 3 months. The firm is usually able to borrow at the Libor rate. The firms
wants to hedge the risk that interest rates go up, so it asks its bank for
a Forward Rate Agreement (FRA) quote. The bank quotes a FRA with a
fixed rate of 4.2% (with quarterly compounding) against a variable rate of
3-month Libor, for a notional of 50 M$.
1. To hedge its exposure, the firm should enter the FRA as the fixed-rate
payer or receiver?
2. If 6 months from now the 3-month Libor fixes at 4.8%, what would
be the cash flow to the firm at the end of the FRA?
3. Suppose that 3 months have passed since the firm entered the FRA.
The firm decides that it will no longer need to borrow the 50 million
dollars in the future, so it decides to cancel the FRA. The current
market rates are r∞ (0, 0.25) = 0.03 and r∞ (0, 0.5) = 0.0325. What is
the value that the firm needs to pay or receive to terminate the FRA?

Ex. 6 — The spot LIBOR rates (converted to continuous compounding)

are:
T r∞ (0, T )
1/12 2%
2/12 3%
3/12 4%
4.1. Forward rate agreements 66

Bank A quotes a FRA 2x3 at F12 (0, 2/12, 3/12) = 4%. Bank A clearly messed
up because this rate is obviously too low relative to the forward rate (even
accounting for compounding differences).
Bank B is going to trade against bank A to take this arbitrage opportunity.
Define the trades that bank B needs to do and compute its profit for a
Notional of 10 M$.

Ex. 7 — A firm will have to borrow 10 M$ two months from now, for a
period of 1 month. The manager wants to hedge the interest rate risk. A
bank quotes a FRA 2x3 at 2.50%.
1. The firm wants to pay or receive the fixed rate in the FRA?
2. The firm manager is not happy with this quote and decides to fig-
ure out how he could hedge the risk by himself with a “homemade
FRA”. He asks a different broker for current borrowing and lending
rates. The broker quotes the following money market spot rates (the
compounding frequency follows the usual money market convention,
i.e., each rate has a compounding frequency corresponding to its own
maturity):
T Bid Ask
2 months (r6 (0, 2/12)) 0.90% 1.10%
3 months (r4 (0, 3/12)) 1.40% 1.60%
For example, these rates mean that the firm can invest for 2 months
at 0.9%, but can only borrow for 2 months at 1.1%.
Using these spot rates, what is the rate that the manager can guar-
antee today for the loan to be done 2 months from now?

Ex. 8 — Exercise 5 from Veronesi’s book, end of chapter 5 (page 184).

Remarks:
•Assume a notional of 100 M$
•In (a), compute the value for the fixed rate receiver. I get $215381.
•In(b.i), I get $321891
•In ([Link]), compute the amount for the fixed rate receiver. My result is
that the fixed rate receiver gets +$330000
4.2. Interest Rate Swaps 67

4.2 Interest Rate Swaps

4.2.1 Definition
An IRS is a portfolio of FRA.

Definition 4.2.1: Interest Rate Swap (IRS)

A plain vanilla fixed-for-floating Interest Rate Swap (IRS) is a contract

between 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 cor-

responding to the length of the period in the contract, that is,
1/n = Ti − Ti−1

Example 4.2.1. Consider the following IRS:

• Notional = 10 M$
• 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.
4.2. Interest Rate Swaps 68

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.

Remark. For simplicity, we are assuming that the fixed and floating payments
always occur on the same dates. However, this does not need to be the case.
For example, one typical swap specification in the U.S. defines fixed payments
every 6 months against floating LIBOR payments every 3 months.

4.2.2 Hedging with an IRS

IRS are useful to transform existing liabilities (or assets) from fixed rate to
floating rate, or vice-versa.

Example 4.2.2. A firm has arranged to borrow 10 M$ at LIBOR

6M + 0.4%. 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 fixed-rate

debt at a rate of
Every 6 months, the firm will pay a net amount of $220 000.

4.2.3 The value of an existing IRS

At initiation the price of the IRS is zero and no payments are exchanged.
However, as time passes and rates change in the market, the value of the IRS
also changes.
4.2. Interest Rate Swaps 69

Proposition 4.2.1: The value of an existing IRS

At any time t after inception, the value of an IRS for the counterparty
receiving the fixed rate is
" m #  
X cn In (T0 , T1 )
IRS
[Link]. (t) = N Z(t, Ti ) + Z(t, Tm ) −N 1 + Z(t, T1 )
i=1
n n

where:

• cn is the fixed rate defined at the beginning of the contract

• T1 is the next payment date (t < T1 )

• In (T0 , T1 ) is the floating rate fixed at the last reset date (T0 ) that
will be paid at T1

• Z(t, T ) is the spot discount factor implied in the current swap

rates.

Proof. Being in a IRS as fixed rate receiver and floating rate payer is equiv-
alent to the sum of two positions:

1. Long a fixed-rate bond paying a cn coupon rate. Its value at t is

" m #
X cn
V fixed leg (t) = N × Z(t, Ti ) + Z(t, Tm )
i=1
n

2. Short a floating rate bond. Its value is 100% plus coupon at T1 . There-
fore,  
floating leg In (T0 , T1 )
V (t) = N 1 + Z(t, T1 )
n

The value of the IRS today is thus

V IRS (t) = V fixed leg (t) − V floating leg (t)

4.2. Interest Rate Swaps 70

Example 4.2.3. Continuing example 4.2.1, suppose that some

years have passed and termination is in 8 months. The swap
curve is currently flat at r2 (0, T ) = 3%, ∀T . LIBOR 6M was
3.2% at the last reset date.
Compute the value of the IRS for the firm:
V fixed leg (t) = . . .
V floating leg (t) = . . .
Hence, the value of the IRS for the firm is:
V IRS (t) = . . .
This means that if the firm wants to terminate the IRS now, it
needs to pay $ 88 819 to the bank.

4.2.4 Setting the fixed rate in an IRS

Proposition 4.2.2: The initial fixed rate in an IRS

At inception (t = 0), the fixed rate in an IRS (cc ) is set at the value
that satisfies the following equation:
m
cn X
cn : Z(0, Ti ) + Z(0, Tm ) = 1
n i=1

where:

• cn is the swap fixed rate with compounding frequency of n periods

per year

• The fixed payments happen at dates T1 , . . . , Tm

• Z(0, T ) is the currently required discount factor (which can be

estimated from other LIBOR/swap rates).

Proof. Since the contract specifies that there is no exchange of money at

inception, the fixed rate is set such that the two legs in the swap have the
same value. The floating leg is like a floating rate bond, so its value is 100%
at time zero (righ-hand side of the equation above).
4.2. Interest Rate Swaps 71

Example 4.2.4. A large bank can borrow and lend at the fol-
lowing rates:

T LIBOR
6 months 1.0%
12 months 2.0%

What rate would this bank quote on 1-year swap paying fixed
every 6 months?
Answer: c2 (0, 1) = 1.9852%

4.2.5 The swap curve

Definition

Definition 4.2.2: Swap curve

The swap curve is the set of swap rates for all maturities:

cn (t, T1 ), cn (t, T2 ), . . . , cn (t, Tm )

Since the swap rate represents the yield to maturity on a bond that is
priced at par, the swap rates are sometimes denoted swap par yields. Swap
rates are quoted daily by swap dealers for swaps contracts up to thirty years
to maturity.1

Extracting discount factors

Note that the initial swap rate represents the coupon rate of a fixed coupon
bond that is priced at par. Hence, we can use the usual procedures (e.g.,
bootstrap or Nelson-Siegel) to extract a set of implied swap discount factors
and the corresponding swap zero curve.
This zero curve is sometimes referred to as LIBOR/Swap zero curve be-
cause the floating leg in the swap pays LIBOR and also because traders
usually build the short end of the curve up to 1 year directly from LIBOR
rates.

1
Rates are available at [Link]
4.2. Interest Rate Swaps 72

Example 4.2.5. Consider the following market rates (published

on the Federal Reserve web site, 20/Feb/2015):

Rate c2 (0, T )
6-month Eurodollar deposit (London) 0.37%
1-year IRS 0.49%
2-year IRS 0.91%

These swaps pay fixed every 6 months.

Check that the implied swap discount factors, extracted with a
bootstrap procedure, and the corresponding swap zero curve are
the following:

T c2 (0, T ) Z(0, T ) r2 (0, T ) r∞ (0, T )

0.50 0.3700% 0.998153 0.3700% 0.3697%
1.00 0.4900% 0.995115 0.4902% 0.4896%
1.50 0.988540 0.7699% 0.7684%
2.00 0.9100% 0.981964 0.9121% 0.9100%

Note that we used interpolation to get Z(0, 1.5) and Z(0, 2) at

the same time.
Why is r2 (0, T ) > c2 (0, T )?

The credit risk level of the LIBOR/Swap curve

1. LIBOR is the rate at which banks lend to other AA-rated banks. While
this is not totally risk free, under normal market conditions the prob-
ability that a AA-rated financial institution will default during a short
period of time (e.g., 3 months) is very small. Hence, traders typically
use LIBOR as a proxy for the risk free rate over short maturities up to
1 year.2
2
However, after the financial crisis of 2007, many dealers now prefer to use the overnight
indexed swap (OIS) rate as a proxy for the risk-free rate. An OIS is a swap where a fixed
rate is exchanged for the geometric average of the overnight rates during a period. It
therefore allows overnight borrowing/lending to be swapped for borrowing/lending at a
fixed rate. See Hull (2012, sec 7.8) for further details.
4.2. Interest Rate Swaps 73

2. Similarly, even though swap rates are not absolutely risk-free, they are
close to that. For example, a financial institution can earn the 5-year
swap rate on a certain principal N by doing the following:
• Lend N for 6 months to a AA-rated borrower and then rollover this
loan for successive 6-month periods to other AA-rated borrowers;
• Enter into a 5-year swap with notional N to pay LIBOR and
receive fixed.
The net result is earning the 5-year swap rate in a loan with a credit
risk that is “continuously refreshed” to AA level at the beginning of
each 6-month period. This is much less risky than lending for 5 years
to the same counterparty (it is AA now, but may deteriorate over the
next 5 years).

Therefore, the LIBOR/swap curve actually provides a better proxy for the
true “risk-free rate” than Treasury bond rates. Treasury rates are artificially
too low due to tax and regulatory issues — see Hull (2012, sec 4.1).

4.2.6 Asset-Liability management with IRS

Interest rate swaps are very useful to manage the duration gap between assets
and liabilities at financial institutions.
Proposition 4.2.3: Dollar duration of a swap
At inception, the dollar duration of an interest rate swap for the coun-
terparty receiving the fixed rate is
$
DIRS ([Link].) = N × (Dfixed − Dfloating )

where

• N is the notional of the swap

• Dfixed is the standard duration of the fixed-rate leg

• Dfloating is the standard duration of the floating-rate leg

$
Proof. From formula (3.7) for the dollar duration of a portfolio, DIRS =
$ $ $
NDfixed − NDfloating . Since D = P D and at inception both prices are 100%,
the result follows.
4.2. Interest Rate Swaps 74

Example 4.2.6. Consider a 10-year swap paying c2 = 4% against

LIBOR every 6 months. Assume the term structure is flat at
r2 (0, T ) = 4%, ∀T . The swap duration is:
Dfixed = 8.34 y
Dfloating = . . .
Hence,
$
DIRS = (N $) × (7.84 y)
(Note that this is the duration for the counterparty receiving
fixed)

Asset-Liability Management with IRS

Trade a swap with notional N such that

DE$ = DA$ − DL$ + DIRS

$
=0

Example 4.2.7. (This is example 5.15 in Veronesi (2010)). In

example 3.6.1, we looked at a bank with

D $ (M$y) D $ (M$y)
Assets 19 740 Liabilities 5 000
Equity 14 740

which left the bank very exposed to increases in interest rates.

To manage this risk, the bank can enter into the swap of the
previous example. The required notional is
DE$ = 0 ⇒ . . .
This means that the bank should enter into a swap paying fixed
on a notional of 1 880 M$.
4.3. Forward contracts 75

4.2.7 Exercises
Ex. 9 — Exercise 9 from Veronesi’s book, end of chapter 5 (page 184). I
get swap rates that differ from the solutions by 1 or 2 bp.

Ex. 10 — Exercise 10 from Veronesi’s book, end of chapter 5 (page 184).

Remarks:
•The Libor curve in the table is in continuous compounding.
•For (c), assume a notional of 100 M$ and compute the value for the
floating rate receiver.

Ex. 11 — An insurance company has the following balance sheet:

Market value Duration Market value Duration
Assets (M$) (years) Liabilities (M$) (years)
S.T. Bonds 200 1 Life annuities 400 14
L.T. Bonds 300 8
Total 500 Total 400
Equity 100

1. Compute the dollar duration of the equity.

2. A 20-year interest rate swap is quoted at 4% (semiannual compound-
ing) against 6-month LIBOR. Given today’s interest rates, a 20-year
bond paying a fixed coupon c2 = 4% has a duration of 16 years.
Determine the notional amount of the swap that the firm should trade
to hedge its equity exposure.
3. To hedge its equity exposure, the firm should enter the Swap as the
fixed-rate payer or receiver?

4.3 Forward contracts

4.3.1 Definition
4.3. Forward contracts 76

Definition 4.3.1: Forward contract on a bond

A Forward contract is an OTC agreement between two counterparties
whereby:

1. One counterparty (“long forward”) agrees to buy N units of a

given bond, at a future date Tf , for a forward price F (0, Tf )
defined today

2. The other counterparty (“short forward”) agrees to sell the bond

under the same conditions.

Settlement. At the settlement date Tf , the payoff to the long forward

counterparty is

CF to long forward at Tf = N × [B(Tf ) − F (0, Tf )]

where

• F (0, Tf ) is the forward price defined at the beginning of the contract

(t = 0)

• Tf is the date when the forward contract is settled.

• B(Tf ) is the spot price of the underlying bond at the settlement date
Tf

• N is the face value or number of units of the bond specified in the

contract.

Payoff
✻

p ✲
F (0, Tf ) B(Tf )
4.3. Forward contracts 77

Example 4.3.1. 1) 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) = 0.995012.
2) A firm “buys the forward” or “buys the bond forward”, i.e.,
takes the long position in the forward.
3) Suppose that 6 months from now, r∞ (0, 0.25) = 3%. The
payoff to the firm is
CF to long forward at Tf = . . . = N × (−0.002484)

4.3.2 Hedging with a Forward contract

Suppose that we want to hedge an investment sometime in the future. To
decide whether to buy or sell the forward, we can think in two alternative
ways:

• Intuitively, investing is equivalent to buying a bond in the future, so

we want to go long the forward to guarantee the bond future purchase
price.
• Alternatively, note that to hedge the investment we need to make
money in the forward contract if interest rates go down in the mar-
ket (which means that bond prices go up). From the definition of a
forward contract, the long position makes money when the bond price,
B(Tf ), goes up.

Rule for hedging with a Forward contract

• To guarantee a fixed rate for an investment/loan at future date

Tf , the firm should enter into a long/short position in a Forward
contract to buy a Treasury Bill with a remaining maturity at Tf
equal to the length of the investment/loan.

• The Notional (N) in the Forward contract should be such that

NF (0, Tf ) equals the amount of the future investment/loan.

For a firm that can borrow or lend at the risk-free rates, this hedging
strategy guarantees that the rate on the future investment/loan will equal
the forward rate implied in the Forward bond price.
4.3. Forward contracts 78

Example 4.3.2. A firm has a receivable of 10 M$ in 6 months.

The firm wishes to invest this amount for an additional 3 months.
To guarantee the investment rate, the firm enters into the Forward
contract of example 4.3.1 with the following notional:
Investment = NF (0, Tf ) ⇒ N = . . .
The forward rate implied in the forward [Link] price is:
r∞ (0.5, 0.75) : . . .
This means that the firm is guaranteed to get 2% on its invest-
ment.
CHECK:

1. Assume that 6 months from now, the 3-month spot rate is

r∞ (0.5, 0.5, 0.75) = 3%.
2. The payoff from the Forward is:
CF to long forward at Tf = . . .
That is, the firm will have to pay -24 969 $ to the bank.
3. The total amount that the firm has to invest at Tf = 0.5 is
V (0.5) = . . .
4. Assuming the firm can invest at r∞ (0.5, 0.5, 0.75), the ter-
minal value will be
V (0.75) = . . .
5. The effective rate obtained by the firm is
k : V (0.75) = (10 M$)ek×0.25 ⇒ k = . . .

4.3.3 Setting the initial Forward price

4.3. Forward contracts 79

Proposition 4.3.1: The initial forward price on a Treasury

bond
Consider a forward contract on a Treasury bond that will pay a coupon
of cn /n at dates
T1 , T2 , . . . , Ts , Ts+1 , . . . , Tm
where Ts is the date of the last coupon that will be paid during the life
of the forward contract, that is, Ts ≤ Tf < Ts+1 . Assume that a large
financial institution can borrow and lend at the same risk-free rates.
This institution can quote a forward price today (t = 0) given by
m
X cn
F (0, Tf ) = Z(0, Tf , Ti ) + Z(0, Tf , Tm ) (4.2)
i=s+1
n

where

• Z(0, Tf , Ti ) is the risk-free forward discount factor.

Alternatively, the forward price is also given by

1
F (0, Tf ) = [B(0) − I] (4.3)
Z(0, Tf )

where

• B(0) is the current spot bond price.

• Z(0, Tf ) is the risk-free spot discount factor

• I = sj=1 cnn Z(0, Tj ) is the present value of the coupons received

P
during the life of the forward contract.

Proof. For (4.2), see equation 5.34 in Veronesi (2010). For (4.3), see equation
5.2 in Hull (2012). The proofs are essentially the same as we did for FRA.

Note that equation (4.2) makes intuitive sense as it is the same as the
typical bond pricing formula, but using the appropriate forward discount
factors that start off from the forward contract settlement date.

Example 4.3.3. The spot risk-free rates are the following:

4.4. Interest rate Futures 80

T r∞ (0, T )
0.25 1.20%
0.50 1.40%
0.75 1.60%
1.00 1.80%
1.25 2.00%
1.50 2.20%

Check that the Forward in example 4.3.1 is correctly priced.

4.3.4 Exercises
Ex. 12 — Show that the forward price in equation (4.2) is really the same
as in (4.3).

Ex. 13 — Using the rates in exercise 4.3.3, determine the 6-month forward
price for a 1.5-year 4% Treasury Bond. (This means that at the forward
settlement date, the bond will still have 1 year until maturity: Tf = 0.5,
Tm = 1.5. Recall that [Link] pay semiannual coupons.) Check that you
get the same value with equation (4.2) or (4.3).

4.4 Interest rate Futures

4.4.1 Overview
Definition
4.4. Interest rate Futures 81

Definition 4.4.1: Interest rate Futures contract

A Futures contract is an Exchange-traded contract between two coun-
terparties whereby:

1. One counterparty (“long futures”) agrees to buy N units of an

underlying asset, at a future date Tf , for a future price F (0, Tf )
defined today

2. The other counterparty (“short futures”) agrees to sell the asset

under the same conditions.

In some contracts the underlying asset is a standard bond, whereas in

others it is an artificial asset designed to represent an interest rate.

Differences between Futures and Forwards

Futures and forward contracts are very similar. However, there are some
important differences:

1. Futures are exchange-traded. If we enter into a Futures, our counter-

party is the exchange.

2. Futures are standardized. We can only trade the contracts for the un-
derlying securities, maturities, and quantities that the exchange spec-
ifies. 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 contracts today (i.e., enter
into the long position), our cash flow tomorrow is:

CF to long futures at day 1 = N × [F (1 day, Tf ) − F (0, Tf )]

This mark-to-market is repeated every day until we close the position.

4.4. Interest rate Futures 82

If we hold the futures until maturity,

CF to long position over life of the futures = N×

[F (1 day, Tf ) − F (0, Tf )
+ F (2 days, Tf ) − F (1 day, Tf )
+ ...
+ F (Tf , Tf ) − F (Tf − 1 day, Tf )]
= N × [B(Tf ) − F (0, Tf )]

The last line uses the fact that the futures price must converge to the
underlying security price at the futures maturity date: F (Tf , Tf ) =
B(Tf ).
Hence, the total payoff is similar to a forward contract, despite the
timing of the cash flows being different.
4. Margins. Traders need to post an initial amount of money (“initial
margin”) with the exchange and must replenish the account whenever
the daily losses drive the balance below a given “maintenance margin”.

Margins and daily mark-to-market limit the credit risk that the exchange
takes from each trader. Therefore, futures have much less credit risk than
forward contracts.

Typical futures contracts

Some important interest rate futures are:

• Futures on short-term interest rates. Mainly used to hedge exposures

to short-term rates. Typically, settlement is in cash. Examples:
– 3-Month Eurodollar
– 3-Month LIBOR
– 3-Month Euribor
– 30-day Federal Funds
• Futures on government bonds. Mainly used to hedge exposures to long-
term rates. Typically, settlement is physical. Examples:
– U.S. Treasury Bond
– 2-, 5-, and 10-year U.S. Treasury Note
4.4. Interest rate Futures 83

Approximate Futures price

Futures prices are closely related to forward prices.

Proposition 4.4.1: Relation between Forward and Futures

prices
Consider a futures and a forward contract on the same underlying
security. Assume that:

1. There is no daily mark-to-market in the futures

2. The futures and forward contracts payoffs occur on the same date
Tf

3. (For futures on [Link] also assume that both the cheapest-to-

deliver and the delivery date are known, as discussed in section
4.4.3)

Then,
Ffutures (0, Tf ) = Fforward (0, Tf ) (4.4)

Proof. See fact 6.2 in Veronesi (2010).

In reality, these assumptions are not true, so the futures is only approx-
imately equal to the forward price. The approximation will be better for
short maturities and when the volatility of interest rates is small.

4.4.2 Eurodollar futures

One of the most popular short-term interest rate futures contracts is the
Eurodollar futures contract traded by the CME Group. A eurodollar is a
dollar deposited outside the U.S. The underlying interest rate is the USD
3-month LIBOR.

Contract specifications

• Underlying unit: 3-month Eurodollar interbank deposit of 1 M$.

4.4. Interest rate Futures 84

• Price quote. The price quote is really an indirect quote of an interest

rate, that is, the price quote today is a number Q(t) such that it implies
a future rate:
Q(t)
f4e$ (t, Tf , Tf + 0.25) = 1 −
100
E.g., a price quote of 97.45 signifies a deposit rate of 2.55% per annum.
At the expiration of the contract (on the 3rd Wednesday of the contract
month) this rate converges to the actual 3-month LIBOR.

• Contract Months: Nearest 40 months (i.e., 10 years) in the March

Quarterly cycle (Mar, Jun, Sep, Dec) plus the nearest 4 “serial” months
not in the March Quarterly cycle.

Q
Mark-to-market. The purpose of the contract design of 100 = 1−f4e$ is to
make the futures quote move like a bond in response to interest rate changes:

f4e$ (t, Tf , Tf + 0.25) Q(t) Long Short

ր ց Loss Profit
ց ր Profit Loss

The total profit/loss for each day is:

Q(t) − Q(t − 1 day) 1

CF to long futures at t = N × × (4.5)
100 4
with N being the notional traded (number of contracts times 1 M$).3

Hedging with Eurodollar Futures

Given how the futures is designed, the rule is very similar to the one for a
Forward contract on a ZCB.
Rule for hedging with Eurodollar futures

To guarantee a fixed rate for a 3-month investment/loan at future date

Tf , the firm should enter into a long/short position in a Eurodollar
Futures contract with expiration as close as possible to Tf .

3
In particular, a change of 0.01% in the interest rate equals a change of 0.01 price
points in the futures quote, which means a profit or loss of (1 M$) × 0.0001/4 = $25 per
contract.
4.4. Interest rate Futures 85

For a firm that can borrow or lend at the 3M LIBOR rates, this hedging
strategy guarantees that the rate on the future investment/loan will be very
close to the rate implied in the current futures price. However, the hedge
may not be perfect due to the timing of the cash flows, as in the following
example.

Example 4.4.1. (This is example 6.3 in Hull (2012).) Its May

and an investor wants to lock in the interest rate of a 3-month
100 M$ investment beginning in September this year. The Septem-
ber Eurodollar futures quote is 96.50.
The implied rate that the investor can lock in is:
f4e$ = . . .
The investor hedges by (buying/selling) 100 con-
tracts.
CHECK:

1. Suppose that in September the 3m USD LIBOR turns out

to be 2.6%.
2. The final settlement in the futures is at Q(Tf ) = . . .
3. The total cash flow to the investor is:
CF to long futures at Tf = . . . = $225 000
4. The amount available for investment is
V (Tf ) = . . .
5. The final amount will be
V (Tf + 0.25) = . . .
6. The effective rate of return is thus
k4 = (V (Tf + 0.25)/(100 M$) − 1) × 4 = . . . = 3.5058%

The rate obtained is close, though not equal, to the initial 3.5%.
This deviation is due to the futures paying off at the beginning
of the investment period (September), rather than at the end
(December).4
4
The futures payoff of N × ∆Q × 0.25% that is received at Tf should, in fact, only
be received at Tf + 0.25 to be comparable to the standard return on a LIBOR deposit.
In other words, if we are receiving the payoff earlier, at time Tf , we would only need to
receive the smaller amount N1+LIBOR/4
×∆Q×0.25%
. However, there is no way to perfectly adjust
for this timing because we do not know the future value of LIBOR. One way to partially
adjust is to assume the initial 3.5% for the investment period and reduce the notional by
1/(1 + 0.035/4) = 0.9913, that is, to buy only 99 contracts (rather than 100).
4.4. Interest rate Futures 86

Furthermore, the futures is settled daily (not all at the end), so the evo-
lution of interest rates during the life of the futures will also have a small
impact on the performance of the hedge.

Extending the LIBOR zero curve

LIBOR rates only go out to 12 months. Since Eurodollar futures are more
liquid that Interest Rate Swaps, traders typically estimate the LIBOR/Swap
curve from the following:

1. LIBOR rates up to 1 year

2. Eurodollar futures from 1 to around 3 years

3. Interest Rate Swaps for longer maturities

A Eurodollar futures is very similar to a FRA, so equation (4.4) would

suggest that the rate implied in the futures is the same as the forward LIBOR
rate. However, the two rates are not exactly the same. The reason is that
the assumptions of equation (4.4) do not hold in reality, which causes the
futures rate to be too high, f4e$ (t, Tf , Tf + 0.25) > f4 (t, Tf , Tf + 0.25). The
following proposition, sometimes denoted convexity adjustment, provides a
simple correction.5
Proposition 4.4.2: Relation between forward and futures im-
plied rates

The forward LIBOR rate (with continuous compouding),

f∞ (t, Ti , Ti+1 ), can be extracted from Eurodollar futures through

e$ 1
f∞ (t, Ti , Ti+1 ) = f∞ (t, Ti , Ti+1 ) − σ 2 Ti Ti+1 (4.6)
2
where
e$
• f∞ (t, Ti , Ti+1 ) is the rate implied in the Eurodollar futures (with
cts comp.)

• σ is the annualized volatility of the underlying LIBOR rate. This

can be estimated from the annualized standard deviation of his-
torical changes in the LIBOR rate.

5
See Hull (2012, sec. 6.3) for details.
4.4. Interest rate Futures 87

Proof. See Veronesi (2010, fact 6.3) or Hull (2012, sec 6.3). Note that this
adjustment holds under a specific interest rate model (the Ho and Lee model),
which we do not cover in this course, so you are not responsible for knowing
this proof.

Example 4.4.2. Suppose the market is as follows:

Instrument Value
LIBOR 12M 3.00%
1-year Eurodollar Quote 96.5

Assume σ = 0.012
To estimate the LIBOR Zero curve, proceed as follows:

1. From the LIBOR,

r∞ (0, 1) = . . .
2. From the Eurodollar futures,
f4e$ (0, 1, 1.25) = . . .
and
e$
f∞ (0, 1, 1.25) = . . . = 3.4848%
3. Using the correction in (4.6), the forward LIBOR is
f∞ (0, 1, 1.25) = . . . = 3.4758%
4. From this f∞ (0, 1, 1.25) and r∞ (0, 1), we get:
r∞ (0, 1.25) : . . . ⇒ r∞ (0, 1.25) = 3.0599%

The estimated spot LIBOR Zero Curve is thus:

Rate Value
r∞ (0, 1) 2.9559%
r∞ (0, 1.25) 3.0599%

We could then proceed in the same way with longer Eurodollar

futures, and then switch to swaps at 3 years.
4.4. Interest rate Futures 88

4.4.3 Treasury bond futures

One of the most popular long-term interest rate futures contracts is the
Treasury bond futures contract traded by the CME Group.

Contract specifications

• Underlying unit: One U.S. Treasury bond having a face value at ma-
turity of $ 100 000.

• Deliverable grades: Bonds with remaining maturity of at least 15 years,

but less than 25 years, from the first day of the delivery month.

• Price quote: Points and 1/32 of a point. For example, 134-16 represents
134 16/32. One point represents $ 1 000 per contract unit.

• Contract Months: The first three consecutive contracts in the March,

June, September, and December quarterly cycle.

The party with the short position has the following delivery options:

• Quality option. The short position can choose which bond to deliver
within the maturity range specified in the contract. The short position
will therefore search for the cheapest-to-deliver bond as discussed below.

• Wild card option. Delivery can be made on any day during the delivery
month. While trading in the futures stops at 2 pm, the short can deliver
until 8 pm. If in a given day during the delivery month, the bond price
declines after 2 pm, the short can decide to deliver and purchase the
bond in the market to deliver at the higher 2 pm futures settlement
price. If the price does not decline, the short can keep the position
open and try the same strategy next day.

• End-of-month option. Trading in the futures stops 7 business days

before the last business day of the delivery month. However, delivery
can occur until the last business day of the delivery month.

These options to the short position reduce the futures price relative to
what it would be without those options.
4.4. Interest rate Futures 89

Cheapest to deliver

The short position decides which bond to deliver (under the specifications
of the contract). If no adjustment was made, every short trader would pick
the same bond — the one with the lowest market price. This would create
the opportunity for market manipulation by speculators, leading to liquidity
issues, and a potential failure of the futures market. Hence, the exchange
implements the following conversion factor system. The goal is to make all
eligible bonds equivalent, but in practice the system is imperfect and one
bond will emerge as the “cheapest to deliver”.

Conversion factor. The exchange determines a conversion factor for each

deliverable bond. The conversion factor is the quoted price that the bond
would have on the first day of the delivery month if its yield to maturity was
6% (with semiannual compounding).6

Example 4.4.3. Compute the conversion factor (C) for the fol-
lowing:

1. T-bond A paying 10% and with 20 years to maturity on the

first day of the delivery month:

40
X 0.05 1
C= i
+ = 1.462295
i=1
1.03 1.0340

This means that a 10% coupon bond is worth 146.2295% of

a 6% coupon bond.
2. T-bond B paying 3% and with 16 years to maturity on the
first day of the delivery month:

C = . . . = 0.694169

Bond price in the futures contract. When a particular bond is deliv-

ered, the price that the long position pays for the bond is:

Invoice price = Most recent Futures settlement price × Conversion factor

+ Accrued Interest on the bond
6
See Hull (2012, sec 6.2) for details on how CME adjusts for uneven maturities.
4.4. Interest rate Futures 90

Intuitively, we can think of the Futures as being on an underlying 6% coupon

bond, and therefore we need to correct the futures price according to the
actual coupon of the particular bond that is delivered.

Example 4.4.4. Continuing the previous example, assume that

the Futures closes at 98 on the first delivery day. Then, price for
each bond under the futures contract is:

Bond F*C AI Invoice

A 143.30% 0 143.30%
B 68.03% 0 68.03%

Cheapest to deliver. The short position will choose the bond that mini-
mizes the cost of delivery:

Cost of delivering bond i =

= Market price of bond i − Revenue from delivery in futures
=(Quoted price of bond i + AI) − (F × Ci + AI)
= Quoted price of bond i − F × Ci
≡ Basis of bond i

If the futures settles at 100% and all the yields are at 6%, all bonds cost
the same to deliver (zero). Otherwise, there will be some differences and one
bond will be the cheapest to deliver.

Example 4.4.5. Continuing the previous example, further as-

sume that both bonds are trading at a ytm of 6% on the first
delivery day. The cost to the short position of delivering each
bond is:

Bond Quoted price F*C Q - FC

A 146.23% 143.30% 2.92%
B 69.42% 68.03% 1.39%

The short would therefore choose to deliver Bond B.

For each futures contract, the short party would have to:
4.4. Interest rate Futures 91

1. Purchase $100,000 of face value of T-bond B on the market

at a total cost of
2. Deliver these bonds to the futures counterparty, receiving a
total of
3. The net cost would thus be $ 1 388.34 per contract.

Note that this net cost can never be negative during the delivery month.
Otherwise there would be an arbitrage opportunity: short the futures, buy
the bond, make immediate delivery.

Hedging with Treasury bond Futures

Rule for hedging with Treasury bond futures

To hedge the investment in a bond portfolio against a instantaneous
parallel shift in interest rates, the firm should short Treasury bond
futures contracts with
V p Dp
NF = − (4.7)
BCTD DCTD /CCTD

where

• NF is the total notional (in $) to be traded in the futures (the

minus sign means shorting)

• Vp is the current market value (in $) of the portfolio

• Dp is the current duration (in years) of the portfolio (as defined

in eqn (3.2))

• CTD denotes the bond that is expected to become the cheapest-

to-deliver.

• BCTD is the current full price (in %) of the CTD

• DCTD is the current duration of the CTD (as defined in eqn (3.2))

• CCTD is the conversion factor of the CTD

Proof. Let h denote the hedged portfolio (assets plus futures). We want its
4.4. Interest rate Futures 92

dollar duration (equation 3.7) to be zero:

Dh$ = NF DF$ + 1Dp$ = 0

where Dp$ is the dollar duration of the underlying total portfolio of assets to
be hedged (in $.y, already resulting from the aggregation of individual assets
through 3.7, and thus Np = 1). Note that from (3.5), Dp$ = Vp Dp .
Denote the CTD by i. Assume the CTD is known and its basis is constant:
Bi −(F Ci +AIi ) = const. Since only Bi depends on the current interest rates,
the dollar duration of the futures contract (equation 3.5) is DF$ := − dFdr
=
$ $
Di /Ci . Also, Di = Di Bi
Replacing in the previous equation,

Dh$ = 0 ⇒ NF Di Bi /Ci + Vp Dp = 0

See applications in Labuszewski, Kamradt, and Gibbs (2013). Rendleman

(1999) and Hull (2012, eqn 6.5) provide alternative rules.

Example 4.4.6. A fixed-income manager holds a portfolio on

which considerable capital gains have been earned. It is October
and the manager expects interest rates to go up in the coming few
weeks. The manager is reluctant to sell the portfolio and replace
it with lower duration bonds because that would result in large
transaction costs as well as the realization of capital gains for
tax purposes. The manager decides to hedge with the December
T-bond futures contract.
The current portfolio is worth 10 M$ and has a duration of 8.55
years.
The cheapest-to-deliver is estimated to be a bond with a current
price of 0.975938, a duration of 9.016 years, and a conversion
factor of 0.6867.
The total notional to trade is
N = ...
which represents shorting 67 contracts.
(These numbers are similar to the example in Labuszewski, Kam-
radt, and Gibbs (2013, p.16))
4.5. Interest rate Options 93

Remark. Over the life of the futures, its price tends to correlate more strongly
with the security that is expected to become the cheapest-to-deliver. Still,
the futures is not actually tied to any specific bond (the expected cheapest-
to-deliver may change throughout the life of the futures). This means that
there is basis risk, i.e., the price of our portfolio and the price of the futures
may move differently. In practice, hedging with T-bond futures is imperfect.

4.4.4 Exercises
Ex. 14 — Consider the following market prices:
Instrument Value
LIBOR 6M 1.0%
LIBOR 12M 1.5%
1-year Eurodollar Quote 96
1.25-year Eurodollar Quote 95
2-year Swap 3.0%
The swap pays fixed every six months against LIBOR 6M. The volatility of
the LIBOR rate underlying the Eurodollar futures is σ = 0.01
Compute the LIBOR zero curve with continuous compounding, r∞ (0, T ), for
maturities T = 0.5, 1, 1.25, 1.5, 2 years.

4.5 Interest rate Options

4.5.1 Definition
4.5. Interest rate Options 94

Definition 4.5.1: Financial option

An option is a contract between two counterparties, written on a vari-
able F , with maturity T , and strike price K. There are two types of
options:

• A call option gives its buyer (long call) the right, but not the obli-
gation, to ask the option seller for the payment of the following
amount:
Long call payoff = max(F (t) − K, 0)
The option seller (short call) has the obligation to pay this
amount when the counterparty decides to exercise the option.

• A put option gives its buyer (long put) the right, but not the obli-
gation, to ask the option seller for the payment of the following
amount:
Long put payoff = max(K − F (t), 0)
The option seller (short put) has the obligation to pay this
amount when the counterparty decides to exercise the option.

If the option can only be exercised at maturity (T ), then it is called

European. If it can be exercised at any time before maturity, it is called
American.

Payoff
✻

p ✲
K F (T )

There are options on different types of underlying variables (F ). The

most common interest rate options are the following:

1. Bond options: the underlying F is the price of a bond.

4.5. Interest rate Options 95

2. Caps and Floors: the underlying is an interest rate, e.g., 3-month LI-
BOR.

3. Futures options: the underlying is an interest rate futures, e.g., the

T-bond futures.

4. Swaptions: the underlying is a swap.

4.5.2 Bond options

Bond options trade OTC. In addition, bond options are embedded in several
contracts. For example:

• A callable bond is a bond that allows the issuing firm to buy back the
bond at a predetermined price sometime in the future. The holder of
a callable bond has in fact sold a call option to the issuer. The call
features tend to reduce the price and increase the yield of the bond.

• A 2-year fixed-rate deposit that can be redeemed at any time without

penalty contains an American put option on a bond. The deposit is
like a bond that the investor has the right to put back to the bank at
face value.

• A committed line of credit gives the client a put option on a bond.

Suppose a bank quotes a 6% rate on a 1-year loan and states that the
loan can start at any time during the next 6 months. The client thus
has an option to sell a ZCB to the bank within the next 6 months.

Example 4.5.1. (This is example 6.2 in Veronesi (2010)).

A firm will receive 100 M$ in 6 months and intends to invest it
for another 6 months. The firm is concerned that interest rates
decrease in the coming months.
HEDGE:

• Alternative 1. Long forward. A 6-m forward on a 6-m T-bill

is quoted at F (0, 0.5, 1) = 0.97938, which allows the firm to
fix a fwd rate of r2 (0, 0.5, 1) = 4.21%
4.5. Interest rate Options 96

• Alternative 2. Instead, the firm can purchase a call on a

6-m T-bill. Suppose that the price of a 6-m call with K =
0.97938 is

c = 0.2701% (of principal of T-bills)

The required option notional is:
N = . . . = 102.105414 M$
The total initial cost for the firm is:
Total option premium = ... = $ 275 787

VERIFICATION:

• SCENARIO 1. Rates do decrease and in 6 months r2 (0.5, 0.5, 1) =

2%.
1. The underlying variable is the price of the ZCB:
F (0.5) = . . . = 0.990099
2. The firm decides to exercise:
Payoff to firm = ... = 1.094469 M$
3. Total amount available to invest:
V (0.5) = . . .
4. Final value of investment:
V (1) = . . .
5. Total effective return on investment (excluding option
premium):
k2 : . . . ⇒ k2 = 4.21%
The firm guarantees the same rate as with the forward. How-
ever, if we also include the initial option premium, the firm
does a little worse than with the forward.
• SCENARIO 2. Rates increase to r2 (0.5, 0.5, 1) = 6%
1. The firm decides not to exercise.
2. Total effective return on investment (excluding option
premium) is k2 = 6%
The firm does better than with the forward.
To be more precise, we should include the option premium.
Suppose that the initial term structure was flat at r2 (0, T ) =
4.21%, ∀T .
4.5. Interest rate Options 97

1. The value of the option premium compounded to t=0.5

is:
call premium × . . .
2. Total amount available to invest:
V (0.5) = . . .
3. Final value of investment:
V (1) = . . .
4. Total effective return on investment (including option
premium):
k2 : . . . ⇒ k2 = 5.42%
Still better than the forward. Though note that options are
expensive!

“Everybody loves options until they see their cost” — popular

saying.

4.5.3 Caps and Floors

Caps and floors are OTC options on a underlying short-term rate (e.g., 3M
LIBOR). The strike price K is actually a strike “rate”. They are cash settled.

Caps

Definition 4.5.2: Caps

• A caplet with strike rate Kn on a floating rate In , for the period

between Ti−1 and Ti , is an option that pays the long position the
amount

Payoff at Ti = N × max(In (Ti−1 , Ti ) − Kn , 0) × (Ti − Ti−1 )

where the In (Ti−1 , Ti ) is the index observed at Ti−1 .

• A cap with maturity Tm , frequency n, and strike rate Kn , is a

sequence of caplets extending out to Tm .

Kn is denoted the “cap rate” and the period between reset dates, Ti −Ti−1
is the “tenor”. Caps are usually designed such that there is no payoff on the
first date T1 regardless of the value of the index today.
4.5. Interest rate Options 98

Caps are designed to insure floating-rate liabilities against the index in-
creasing above K.

Example 4.5.2. A firm has debt of 10 M$ at LIBOR 3m + 0.4%.

The firm manager expects interest rates to go up.
HEDGE:

• A bank offers a 2-year cap, with a cap rate of 4%, and 3

month tenor.
• If the firm purchases this cap with a notional of 10 M$ it
is guaranteed to never pay more than 4% (+ spread) in any
3-month period.

VERIFICATION:

1. Suppose that in 6 months, LIBOR 3m = 5%

2. 3 months later, at T3 = 0.75, the payoff from the cap is
Payoff at 0.75 years = . . . = $25, 000
3. The effective borrowing cost for the (0.5, 0.75) period (ex-
cluding the initial cap premium) is
k4 : . . . ⇒ k4 = 4.4%
Remark. Each caplet is equivalent to a long put option on a zero-coupon
bond. In both cases, the long position profits when the interest rate increases.
See Hull (2012, sec 28.2) for details.

Floors

Definition 4.5.3: Floors

• A floorlet with strike rate Kn on a floating rate In , for the period

between Ti−1 and Ti , is an option that pays the long position the
amount

Payoff at Ti = N × max(Kn − In (Ti−1 , Ti ), 0) × (Ti − Ti−1 )

where the In (Ti−1 , Ti ) is the index observed at Ti−1 .

• A floor with maturity Tm , frequency n, and strike rate Kn , is a

sequence of floorlets extending out to Tm .
4.5. Interest rate Options 99

A floor pays off whenever the LIBOR drops below the floor rate K.
Remark. Each floorlet is equivalent to a long call option on a zero-coupon
bond. In both cases, the long position profits when the interest rate decreases.

Collars

Definition 4.5.4: Collar

A collar is combination of a cap and a floor:

Long collar = long cap (K c ) + short floor (K p )

The purpose of a collar is to reduce the cost of hedging: the cost of the
cap is reduced by selling a floor. One typical configuration is to determine
the value of the strike rates, K p < K c , such that the initial cost of the collar
is exactly zero.
If a firm buys a collar to hedge a floating rate liability, it will never pay
more than K c , but it also never pay less than K p .

4.5.4 Futures options

In a Futures option the underlying is an interest rate futures, e.g., the Eu-
rodollar futures or the T-bond futures. The option gives the right, but not
the obligation to buy/sell the futures:

• The long call has the right to enter into a long futures position at the
strike price instead of the current futures price in the market.

• The long put has the right to enter into a short futures position at the
strike price.

Futures options are traded on the same exchange of the futures. Ex-
changes prefer to write options on an underlying whose price is easily ob-
servable (the futures), rather than on an illiquid bond.

4.5.5 Swaptions
In a swaption the underlying is a swap:
4.5. Interest rate Options 100

• The call option is denoted payer swaption. The long call has the right
to enter into a swap and pay the fixed strike rate instead of the current
market swap rate.

• The put option is denoted receiver swaption. The long put has the
right to enter into a swap and receive the fixed strike rate.
Chapter 5

Term structure dynamics:

Vasicek model

5.1 Model for the short rate

The starting point is a description of how the short end of the curve evolves
through time.
Definition 5.1.1: Vasicek model for the instantaneous short
rate
Vasicek assumes the following mean-reverting process (in the risk-
neutral measure):

drt = α(r̄ − rt ) dt + σ dWt (5.1)

where

• rt is the spot rate for the next instant (t, t + dt)

• α, r̄, σ are parameters (to be estimated)

• Wt is a standard Brownian motion: W0 = 0, Wt ∼ N(0, t),

independent increments.

The parameter α defines how fast rt converges to the long-term r̄. High
values of σ perturb this convergence.
To interpret this process think of its discrete time approximation:
√
r(t + ∆t) − r(t) = α(r̄ − r(t))∆t + σε(t) ∆t

101
5.2. Bond prices 102

where ε(t) ∼ i.i.d.N(0, 1).

Remark. One of the main criticisms of this process is that it allows interest
rates to be negative. More precisely, rT ∼ N(.) for some future time T
(details in Veronesi (2010, eqn 14.34)). However, on 9/March/2015 the rates
in the Euro area were the following:

Rate 9/Mar/2015
Eonia -0.064%
1-month Euribor -0.008%
1-year AAA-rated Bonds (ECB estimate) -0.221%
5-year AAA-rated Bonds (ECB estimate) -0.120%

5.2 Bond prices

The particular value of the state variable rt determines the whole term struc-
ture of interest rates at time t.
Proposition 5.2.1: Price of a Zero-Coupon Bond
In the Vasicek model, the price at time t of a ZCB that pays 1 at time
T is
Z(r, t; T ) = exp (A(t; T ) − B(t; T )rt ) (5.2)
with
1 − e−α(T −t)
B(t; T ) = (5.3)
α
σ2 σ 2 B(t; T )2
 
A(t; T ) = [B(t; T ) − (T − t)] r̄ − 2 − (5.4)
2α 4α

Proof. Veronesi (2010, fact 15.3)

The price of a coupon bond can then be computed in the usual way:
m
cn X
P (r, t; Tm , cn , n) = Z(r, t; Ti ) + Z(r, t; Tm ) (5.5)
n i=1
5.3. Estimating the parameters 103

5.3 Estimating the parameters

One possible way to estimate the parameters is the following:

1. Assume that rt equals the current overnight rate.

2. Search for values of the 3 parameters that make the model prices fit
bond prices observed in the market:1

I
X
2
minimize PiM arket − PiV asicek
α,r̄,σ
i=1

where PiV asicek is given by equation (5.5).

Example 5.3.1. Consider the following bond prices:

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%
All coupons are paid annually. The overnight rate is 2.4691%
Using the procedure described above, I get the following param-
eters:

α 0.83445
r̄ 0.05852
σ 0.01482

The resulting rates are in figure 5.1.

1
Alternatively, we can first estimate σ from historical data, and then search only over
r̄ and σ. This is ok because σ is the same in both the risk-neutral and the physical
measure. Note that r̄ and σ cannot be estimated from historical data because they are the
parameters of drift in the risk-neutral measure. See Veronesi (2010, sec 15.2.4) for details.
5.3. Estimating the parameters 104

Figure 5.1: Vasicek Term Structure

5.4. Option prices 105

5.4 Option prices

Proposition 5.4.1: Price of an option on a ZCB

Under the Vasicek model, the price today (t=0) of a European call
option with strike price K and maturity To on a zero coupon bond
maturing at Tb > To is

Call(r0 , 0) = Z(r0 , 0; Tb )N (d1) − KZ(r0 , 0; To )N (d2) (5.6)

and the price of a European put option is

P ut(r0 , 0) = KZ(r0 , 0; To)N (−d2 ) − Z(r0 , 0; Tb )N (−d1 ) (5.7)

where

• Z(.) is the discount factor defined in (5.2)

• N (.) is the standard Normal cdf

 
• d1 = S(To ) log KZ(r0 ,0;To ) + S(T2 o )
1 Z(r0 ,0;Tb )

• d2 = d1 − S(To )
h i1/2
σ2
• S(To ) = B(To ; Tb ) 2α
(1 − e−2αTo )

• B(.) is defined in (5.3)

Proof. See fact 15.5 in Veronesi (2010).

Example 5.4.1. (This is similar to example 15.2 in Veronesi

(2010)). Using the parameters estimated in the previous exam-
ple, determine the value of a 1-year call option on a 4-year ZCB
with a strike of K = 0.8.
We have To = 1 and Tb = 5. Using the formulas above,
Call(0.024691, 0) = 0.6757%

These formulas can be used to price standard derivatives, such as caps,

floors, and swaptions. See Veronesi (2010, ch 19) or Hull (2012) for details.
Solutions to exercises

Answer (Ex. 1) — .

T (years) Price r∞ (0, T ) r2 (0, T )

0.077778 99.8328% 2.1518% 2.1634%
0.416667 98.8292% 2.8266% 2.8466%

Answer (Ex. 2) — The prices are

•98.40%
•105.69%
•100.04%
•100.29%
•101.21%

Answer (Ex. 3) — y∞ = 3.6547%

Answer (Ex. 4) — .

2) The forward rate is r2 (0.25, 0.75) = 4.5018%.

3) For either bond, the realized rate of return is i2 = 4.00%, which equals
today’s spot rate! This shows that the difference between the initial YTM
of the two bonds is not reflective of a difference in expected returns, and
therefore illustrates this drawback of YTM.

106
5.4. Option prices 107

Answer (Ex. 5) — .
1) Payer
2) Net CF for fixed-rate payer: 50 × (0.048/4 − 0.042/4) = 0.075M$. The
firm would receive $75000.
3) Value of FRA for fixed-rate payer: 50×Z(0, 0.25)−50×(1+0.042/4)Z(0, 0.5) =
50 × e−0.03∗0.25 − 50 × (1 + 0.042/4)e−0.0325∗0.5 = −0.084201M$. Hence, the
firm needs to pay $84201.

Answer (Ex. 6) — .
1) Bank B will want to enter the FRA as fixed rate payer.
2) The hedges are:
i) At t = 0, borrow NZ(2/12) for 2/12 years.
ii) At t = 0, invest the same amount for 3/12 years.
iii) At 2/12, rollover the loan in i) until 3/12.
3) The net CF at 3/12 will be:

Z(0, 2/12)
N − N(1 + F12 /12) = N × 0.0016792
Z(0, 3/12)
4) For a notional of 10 M$, this represents a profit of $16 792

Answer (Ex. 7) — .
1) pay
2) The homemade FRA needs to generate a positive CF of +10 M$ at 2/12.
Hence, the trades are:
1
i) Invest 10 × 1+0.009/6 for 2 m.
ii) Borrow the same amount for 3 m. The final payment at 3/12 will be
1
10 × 1+0.009/6 × (1 + 0.016/4) = 10 × 1.002496
The effective rate is thus k12 : 1.002496 = 1 + k12 /12 ⇒ k12 = 2.9955%

Answer (Ex. 8) — See Student Solution Manual.

Answer (Ex. 9) — See Student Solution Manual.

Answer (Ex. 10) — .

a) 4.21%
b) 0
5.4. Option prices 108

c) For the first date, 1-Feb, I get −0.930877 M$. I did not check the other
dates.

Answer (Ex. 11) — .

1) DA$ = 200 + 300 ∗ 8 = 2600, DL$ = 400 ∗ 14 = 5600, DE$ = −3000 M$y
2) Dswap fixed receiver = 16 − 0.5 = 15.5 y. We want DE + NDswap fixed receiver =
0 ⇒ N = 194 M$.
3) receiver

Answer (Ex. 12) — Check whether the RHS are the same:
m
1 X cn
[B(0) − I] = d(0, Tf , Ti ) + d(0, Tf , Tm )
d(0, Tf ) i=s+1 n
" m #
X cn
⇒B(0) − I = d(0, Tf , Ti ) + d(0, Tf , Tm ) d(0, Tf )
i=s+1
n
m
X cn
⇒B(0) − I = d(0, Ti ) + d(0, Tm )
i=s+1
n
m s
X cn X cn
⇒B(0) = d(0, Ti ) + d(0, Tm ) + d(0, Tj )
i=s+1
n j=1
n
m
X cn
⇒B(0) = d(0, Ti ) + d(0, Tm)
i=1
n

which is clearly true as this is the typical bond pricing formula.

Answer (Ex. 13) — 101.3603%

Answer (Ex. 14) — The rates are:

•r∞ (0, 0.5) = 1/0.5 ∗ ln(1 + 0.01/2) = 0.9975%
•r∞ (0, 1) = 1/1 ∗ ln(1 + 0.015) = 1.4889%
•f4ed (1, 1.25) = 4% ⇒ f∞
ed
(1, 1.25) = ln(1+0.04/4)/0.25 = 0.039801. The
ed
corrected fwd rate is f∞ (1, 1.25) = f∞ (1, 1.25) − 0.5 ∗ 0.012 ∗ 1 ∗ 1.25 =
0.039739. Then, r∞ (0, 1.25) = (r(0, 1) ∗ 1 + f (1, 1.25) ∗ 0.25)/1.25 =
1.9859%
•f4ed (1.25, 1.5) = 5% ⇒ . . . r∞ (0, 1.50) = 2.4815%
5.4. Option prices 109

•1 = 0.015d(0, 0.5) + . . . + 1.015d(0, 2) ⇒ r∞ (0, 2) = 3.0024%

Bibliography

Hull, J. C., 2012, Options, Futures, and other Derivatives. Pearson Educa-
tion.

Labuszewski, J., M. Kamradt, and D. Gibbs, 2013, “Understanding Treasury

Futures,” CME Group. Available at [Link].

Rendleman, R. J., 1999, “Duration-based hedging with Treasury bond fu-

tures,” The Journal of Fixed Income, 9(1), 84–91.

Veronesi, P., 2010, Fixed income securities: valuation, risk, and risk man-
agement. John Wiley & Sons.

110