ADVISORY ALADDIN ANALYTICS INVESTMENT ACCOUNTING TRANSITION MGMT

Market Analytics Training Manual

Financial Concepts and Risk Analytics

Draft – for Internal Use Only

Table of Contents

1. Time Value of Money

2. Yield Curves

3. Duration and Convexity

4. Interest Rate Derivatives

5. Options

6. Corporate Credit and Credit Default Swaps

7. Interest Rate Models

8. Introduction to Mortgages

9. Structured Products

10. Value at Risk

Chapter 1 - Time Value of Money
'Time value of money' is a phrase that is heard day in and day out in this industry. Chapter One of this
training packet will provide insights into the different ways of quantifying this term by relating it to
more familiar concepts we come across in our daily lives.

Objectives of this chapter:

• Understand why a dollar today is not worth the same as a dollar a year from now.
• Serve as a refresher for basic financial concepts that you must already be familiar with.

The Value of a Dollar

Given the chance to receive $100 now or $100 a year from now, which would you choose? If you are a
rational market participant, for a variety of reasons, you would take the money now.

• Opportunity Cost: There is an inherent opportunity cost with almost any investment. In this
simple case if you take the money now, you can use it to make an alternate investment – for
example - in the equity markets, in the savings account at your bank, in your business that you
happen to own or a variety of other possibilities. This chance at earning extra income by taking
the money now and using it is the opportunity cost of receiving the money a year from now.
• Liquidity: At some point during the year, unexpected circumstances could put you in the
position of needing money on short notice. To have this money on hand, in case of any such
need arising, is also a benefit of taking the $100 now. This may seem very similar to an
opportunity cost but it is slightly different since it does not represent the ability to make
another investment, but the benefits of having liquid assets available.
• Inflation: The primary purpose of money is to make purchases. The price of the goods that we
buy is constantly in flux, but generally, increases over time. Given Inflation, the $100 will most
likely buy less real goods one year from now and is another cost to taking the delayed
payment.
• Credit Risk: In order to get your $100 a year from now, the party promising you payment needs
to have the funds available in a year’s time. Credit Risk refers to the uncertainty of receiving
your money in the future due to your counterpart going bankrupt.

These are just some basic examples and there are a multitude of other reasons to take the money now.
How much more though, would someone have to offer you to take a future payment instead of the
payment today? Put another way, how much less would you be willing to take, to have the money now?

Definition:

1. Present Value (PV) - The amount of money a future cash flow or future stream of
cash flows is worth today. Put another way, it is the dollar value that will make you
indifferent between receiving the future cash flow and receiving the PV today.
2. Discount Function (DF(t)) – The Discount Factor for a given time t is the PV of $1
received at some future time t. It is essentially the current price of a zero coupon
bond, with a face value of $1, and a maturity at time t.

To find the present value (or “price”) for any future stream of cash flows, all you need is the discount
function (t). In order to compute the PV, you just need to multiply each cash flow with the discount
factor corresponding to the time the cash flow is received.
n
PV = DF1 ⋅ CF1 + DF2 ⋅ CF2 + K = ∑ DFi ⋅ CFi
i =1

©2009 BlackRock, Inc. All Rights Reserved. 1

Chapter 1: Time Value of Money BlackRock Solutions

Return Measures
Simple Return and Yield
If you were to analyze a transaction in which you pay $80 today and receive $100 in two years, how
would you quantify and interpret that return? Let’s look at this trade from the perspective of
transactions you may be more familiar with.

• Investing money in bank - You can view this investment as putting money into a high yield
savings account in your local bank and having it grow to $100 at the end of the two year
period. Formulaically, the return (R) you will earn can be presented as

80(1+R) = 100
R = 25%

• Simple return / Expected Return (ER) - You could also view this investment in terms of the
amount of profit that you make on it. At the end of the investment time frame, how much
more money do you get back? i.e. What is your percentage profit?

(100-80)/80 = ER = 25%

• Yield - Lastly, the investment is analogous to buying a zero coupon bond. What then is the
yield ( y simple ) on this bond?
100
80 =
1 + y simple
y simple = 25%

Irrespective of how you view this transaction, all three interpretations result in the same return.

While the above formulas are good ways of quantifying the return of this investment, there is one
major element missing in all cases – TIME!

All the numbers above measure the return over a two-year horizon, the holding period of the
investment and, as such, are often referred to as holding period returns. Comparing two investments
with different holding periods using such a simplistic calculation is difficult. For example, both
investments in the table below have the same simple return, but of course a rational participant would
opt for the one that returns money earlier.

Payment Money Received in Year.. Simple

Paid Today One Two Return
Transaction 1 $80 $100 25%
Transaction 2 $80 $0 $100 25%

Scaled Returns and Compounded Rates

In order to make different investment decisions comparable, we need to normalize returns to the same
unit in time. The simplest way to introduce the concept of time to the numbers above is simply to
divide the return by the time elapsed. In our example, this would result in an “annual” return of 12.5%.

25%
R= = 12.5%
2

©2009 BlackRock, Inc. All Rights Reserved. 2

Chapter 1: Time Value of Money BlackRock Solutions

In general, after n periods, the value of your investment would be worth,

80(1 + nRsimple )
80(1 + 2 Rsimple ) = 100
Rsimple = 12.5%

where R is the simple annual return of the investment.

This is still not the best representation for returns though. A more popular format is to represent
returns as compounded returns. At the end of each interval we take what profit has been made and
place it back into the same investment, at the same rate of return. In this case, after n periods, the
investment grows to...

80(1 + RCompounded ) n
80(1 + RCompounded ) 2 = 100
RCompounded = 11.8%

Where R is the annually compounded return of the investment.

The annual compounding period we chose was arbitrary of course. We can choose a smaller
compounding interval, say every half a year. The semiannually compounded return for our example is
the value of R semi-compounded such that the equation below holds,

4
⎛ R semi-compounded ⎞
80⎜⎜1 + ⎟⎟ = 100
⎝ 2 ⎠
We divide the yield by two before compounding because there are now two compounding periods per
year.

Compounding Periods

Compounding does not have to be limited to two periods a year. It’s easiest to understand
compounded returns from the perspective of investing money in a bank. Given an investment
horizon of T years, the future value (FV) of the investment is related to the present value (PV)
invested in the bank and the compounding frequency (N) by the compounded return (R) as
described below:
T *N
⎛ R⎞
FV = PV ⎜1 + ⎟
⎝ N⎠

©2009 BlackRock, Inc. All Rights Reserved. 3

Chapter 1: Time Value of Money BlackRock Solutions

Now let’s look at the semiannually compounded yield for the following two transactions.

Payment Money Received in Year.. Simple Semiannual Compounded

Paid Today One Two Return Yield
Transaction 1 $80 $100 25% 23.60%
Transaction 2 $75 $0 $100 33% 14.91%

Here the simple return implies that the second transaction does better, while the compounded returns
show that that is not the case.

Spot rates
At this point, we can very briefly introduce one of the main building blocks of the yield curve, the spot
rate. Spot rates are nothing more than the yields of theoretical zero-coupon bonds. What is the spot
rate implied by the two-year transaction in our original example? The two year semi-annually
compounded spot rate can be calculated as follows...

100
80 = 2x2
⎛ Spot Rate 2yr ⎞
⎜⎜1 + ⎟⎟
⎝ 2 ⎠

Spot Rate 2yr = 11.47%

What is the two year discount factor associated with this spot rate? Per its definition, the two year
discount factor is the price of a zero coupon bond, with a face value of $1 and a maturity of two years.

1 1
DF2 yr = 4
= 4
= 0.8
⎛ Spot Rate ⎞ ⎛ 11.47% ⎞
⎜1 + ⎟ ⎜1 + ⎟
⎝ 2 ⎠ ⎝ 2 ⎠

What is the two year discount factor computed using the two year annually compounded spot rate
derived from the same theoretical bond? – More on this later.

Static Measures and Fixed Coupon Bonds

Fixed coupon bonds are securities that pay a constant percentage of the face value of the bond at
regular intervals. At maturity, the face value of the bond is returned to the investor, along with the
final coupon payment. How would you compute the present value (price) of this future stream of cash
flows? If you have the discount factors corresponding to the time each cash flow is received, it
becomes very simple to value this security.

An example is illustrated in the chart below. Here we have a bond with a face value of $100 that pays
four coupons, let’s say semi-annually, over a period of two years. Assuming we have a discount
function, we can compute its present value by using the formula below:

PV = DF6 m ⋅ C + DF1 yr ⋅ C + DF18 m ⋅ C + DF2 yr ⋅ (C + 100)

©2009 BlackRock, Inc. All Rights Reserved. 4

Chapter 1: Time Value of Money BlackRock Solutions

120
102.5
100

80

Cash Flow
60

40

20
2.5 2.5 2.5
0
1 2 3 4

We now turn to a couple of static measures that are commonly computed for securities. Theses are
static because the cash flows that we use to compute them are assumed to be fixed and the
computation of these analytics only depends on the market price of the security and its cash flows.

Yield to Maturity
The YTM is a single yield (that can be represented in any compounding convention) that, when applied
to all cash flows, returns the market price of the security. For example, using semi-annual
compounding,

C/2 C/2 C/2 C / 2 + 100

Market Price = + + +
(
1+ 2
y
) (
1 + 2y
2
) (
1 + 2y
3
)1 + 2y (
4
)
The yield is the value of y that equates the PV of the future stream of cash flows to its market price.

An important concept to touch upon here is the relationship between Par priced bonds and the yield of
the security. What happens to the price of a bond when its yield to maturity is equal to the coupon
rate? First, let’s take a look at a simple one-period bond, with one coupon payment ($2.50) along with
the principal payment ($100) at the end of the period. Assume you are pricing this security using a flat
yield curve at 2.5%.

Working out the math, we see that today, the price of the bond will be exactly par. The coupon
exactly compensates for the discounting as illustrated below.

2.5 + 100
PV = = 100
(1 + 0.025)

100 100

PV 2.5

PV Today Cash flow at T+1

©2009 BlackRock, Inc. All Rights Reserved. 5

Chapter 1: Time Value of Money BlackRock Solutions

Now let us analyze a multi-period bond where the coupon is equal to the yield being used to discount
future cash flows. Take the example of a bond that pays annually and matures three years from now.

PV? 100

2.5 2.5 2.5

Today Year 1 Year 2 Year 3

In order to price this bond, we can analyze the value of the final payment at year 3 exactly the same
way as the one period case from before. We can discount that final coupon and principal payment back
to the second year. This will give us the price of the year 3 cash flows as off year 2. From the first
exercise, we know the price of those two payments at year 2 to be $100. In this sense, we’ve reduced
the three year bond to a two year bond.

100 100 100

2.5 2.5 2.5

Today Year 1 Year 2 Year 3

We can keep moving backwards in time like this. The combination of this new $100 cash flow at year 2
and the original coupon payment in year 2 translates to a $100 payment in year 1. That again, in
combination with the original year 1 coupon payment translates to a PV of $100 today.

What does this mean? This tells us that when a bond’s yield to maturity is equal to its coupon, the
bond prices at par. Additionally, if the yield is higher than the coupon, then the bond will price at a
discount, and if it is less than the coupon, it will price at a premium.

In later chapters we will introduce the concept of duration and convexity which are used to capture
the sensitivity of a bonds price to changes in interest rates. Below we briefly introduce two static
measures which, like duration, provide investors with a sense of timing in terms of when they will
receive back their investment and which part of the yield curve they are exposed to.

Weighted Average Life

This number captures the average time to receipt of principal payments. Since WAL uses the principal
payments only, and not interest, it is not a measure of interest rate risk. WAL can be thought of as the
average number of years it takes for a dollar of unpaid principal to be returned to an investor.

©2009 BlackRock, Inc. All Rights Reserved. 6

Chapter 1: Time Value of Money BlackRock Solutions

∑ Pt i i
WAL = i =1

∑P i

where Pi is the principal payment at time ti .

Modified Duration
To understand the exposure of a security to interest payments, we can look at modified duration. This
measure is a scaled, cash flow weighted average of the time to receipt of cash flows. Modified duration
is an estimate of the interest rate risk of a bond and is static because we assume that the cash flows do
not change along with shifts in the yield curve. Because of this, modified duration is less reliable than
effective duration, which we will take a look at later, and which does take into account changing cash
flows. Given the number of coupon periods in the year (N) and the yield to maturity (y) of the security,
modified duration can be computed as follows:

∑ DF ⋅ CF ⋅ t
i =1
i i i

price
Modified Duration =
y
1+
N

©2009 BlackRock, Inc. All Rights Reserved. 7

Chapter 2 – Yield Curves
Markets express their views on the time value of money via the price they are willing to pay for
securities issued by different issuers. These prices correspond to yields that an investor can earn for
investing money and taking on risk. Building on our knowledge of discount factors, chapter two
introduces one of the fundamental building blocks of the fixed income market – the Yield Curve.

Objectives of this chapter:

• Understand the information embedded in a yield curve.

• Understand the difference between spot, forward and par rates.

Yield Curves and Discount Factors

We begin our understating of yield curves by revisiting the notion of a discount function. As we saw in
the previous chapter, the following equation represents the present value of a stream of future cash
flows...
n
PV = DF1 ⋅ CF1 + DF2 ⋅ CF2 + K = ∑ DFi ⋅ CFi
i =1

Since any bond can be represented as a collection of zero coupon bonds (i.e. as a stream of single cash
flows), in order to price a bond, all you need is a discount function. Discount factors embed in them
information on all the risk factors that determine the time value of money such as opportunity cost,
inflation and credit risk. Therefore we can have multiple sets of discount functions that represent
different risk profiles. For example, one could come up with a discount function derived from
government issued treasury bonds. This curve would essentially represent a risk free discount function
since government bonds have no default risk. Similarly one could derive a discount function based off
corporate debt issuances. Since investing in corporate debt has an element of credit risk, these
discount factors would be lower than the risk free discount factors i.e. investors will be willing to pay a
higher price for a treasury bond than an equivalent corporate.

So how do discount factors relate to the yield curve? While there are many ways of defining a yield
curve, one shouldn’t forget its primary function.

A Yield Curve is a function that defines a set of discount factors over time.

While a Yield Curve is defined by its discount factors, its representation in the market takes place in
yield or rate space. A given yield curve, representing a set of discount factors, is usually observed in
the form of Spot Rates, Forward Rates or Par Rates. Given a set of discount factors DF1 , DF2 , DF3
etc , we will walk through the definition and derivation of these different rates.

Spot Rates
A spot rate is the yield of a zero coupon bond i.e. it is the rate used to discount a single future cash
flow back to today. Given that discount factors represent the present value of a $1 received at a future
point in time, it’s no surprise that spot rates and discount factors are directly related to each other.

©2009 BlackRock, Inc. All Rights Reserved. 8

Chapter 2: Yield Curves BlackRock Solutions

Assuming that the two year discount factor DF2 is 0.8, the two year semi-annually compounded spot
rate can be calculated as follows...

1
0.8 = 2x2
⎛ Spot Rate 2yr ⎞
⎜⎜1 + ⎟⎟
⎝ 2 ⎠

Spot Rate 2yr = 11.47%

In general the relationship between compounded spot rates and discount factors can be summarized as
follows…
1
DFT = TxN
⎛ Spot Rate T ⎞
⎜1 + ⎟
⎝ N ⎠
where N represents the compounding frequency and T represents the time in years. The main take
away from the above equation is that spot rates and discount factors embed the exact same risk
information and are completely interchangeable. A discount function can be represented as series of
spot rates in any compounding convention we chose. Irrespective of whether the spot rate is quoted as
an annual rate, a semi-annual rate or a simple compounded rate – they all correspond to the same
discount factor and thus provide the same information on the time value of money.

Graphical View
Typically we observe discount factors decrease in value over time, i.e. the present value of a $1
decreases the further out into the future you receive it. Given the inverse relation of discount factors
to spot rates, the observed yield curve tends to be monotonically upward slopping.

5.5

4.5
SPOT RATES

3.5

2.5

1.5
1 3 5 7 10 12 14 16 18 20 22 24 26 28

Time

©2009 BlackRock, Inc. All Rights Reserved. 9

Chapter 2: Yield Curves BlackRock Solutions

Forward Rates
Forward rates are similar to spot rates but, as the name implies, tell us what spot rates are going to be
sometime in the future. A forward rate is used to discount a single cash flow that we will receive at a
future date to another date also in the future. When looking at a forward curve, we are generally
trying to answer one of two questions...

(a) What is the trajectory of a particular spot rate over time?

(b) What will the spot curve look like at some point in the future?

We first introduce some terminology to help us understand how forward rates are related to spot rates.
Let us define f x , y as the x-year spot rate y-years forward. For example, f 2,3 represents the 2 year
spot rate 3 years from today i.e. three years into the future the two year spot rate is f 2,3 . Now Assume
you invested in a zero coupon bond maturing in two years earning you a yield equivalent to the two
year spot rate S 2 . We also know that a three year zero coupon bond will earn you a yield equivalent to
the three year spot rate S 3 . So how much additional yield can you earn if your investment had a three
year horizon instead of a two year horizon? That’s what forward rates tell us!

Forward rates are breakeven rates that measure the marginal yield earned by lengthening the maturity
of a zero coupon bond by one year while spot rates measure the average yield from today to maturity.
Therefore spot rates are geometric averages of one or more forward rates. Going back to our example
above, investing your money for three years can be decomposed into making an initial investment for
two years and then investing the proceeds for one year after that. This can be mathematically
represented as
(1 + S3 )3 = (1 + S2 ) 2 (1 + f1,2 )1
where S t is the t year spot rate and f1, 2 is the one year spot rate two years forward. In general any
multi-period spot can be decomposed into the product of one year forward rates.

(1 + Sn )n = (1 + S1 ) ⋅ (1 + f1,1 ) ⋅ (1 + f1, 2 ) ⋅ (1 + f1,3 )..... ⋅ (1 + f1,n )

Graphical View
Given that spot rates are the geometric average of one or more forward rates, the shape of the
forward curve depends on the profile of discount factors (i.e. the shape of the spot curve). In a typical
upward slopping yield curve environment, the forward curve, represented below as the trajectory of
the three-month spot rate, will lie above the spot curve and will magnify movements in the spot curve.
If the average of a time series increases (decreases) as we add data points, it implies that every
additional data point must be significantly greater (lower) than the previous one in order to impact the
average.

©2009 BlackRock, Inc. All Rights Reserved. 10

Chapter 2: Yield Curves BlackRock Solutions

6 Forward Curve

5.5

4.5

4
Spot Curve
3.5

2.5

1.5
1 3 5 7 10 12 14 16 18 20 22 24 26 28

Time

Par Rates
A par rate is the single rate that can be used to discount all the cash flows of a par priced bond.
Similar to the relationship between spot and forward rates, par rates are averages of one or more spot
rates.

Assuming you had your spot yield curve (i.e. you know its discount factors), how would you determine
your par yield curve? Let us specifically try and solve for the three year par rate. To start off, we write
down the equation for pricing a three year fixed coupon bond. We can represent the price either as a
function of discount factors or spot rates.

PV = DF1 ⋅ C + DF2 ⋅ C + DF3 ⋅ (C + 100)

1 1 1
PV = ⋅C + ⋅C + ⋅ (C + 100)
(1 + S1 ) (1 + S 2 ) 2
(1 + S 3 ) 3

So how do we find the single rate that can be used to discount all cash flows of a three year par priced
bond? In other words, we want to solve for P3 in the equation below.
1 1 1
100 = ⋅C + ⋅C + ⋅ (C + 100)
(1 + P3 ) (1 + P3 ) 2
(1 + P3 ) 3

Since we have our term structure of discount factors, we can solve for the coupon C that will price
this three year bond at par.
100 = DF1 ⋅ C + DF2 ⋅ C + DF3 ⋅ (C + 100)

Great, so now we know what the fixed coupon needs to be to price a three year bond at par, but how
do find P3 ? We already know from the previous chapter that bonds price at par when the coupon
exactly compensates for the discounting i.e. when the coupon is equal to the yield. In other words, P3

©2009 BlackRock, Inc. All Rights Reserved. 11

Chapter 2: Yield Curves BlackRock Solutions

is equal to the coupon C ! Therefore it is easy to derive par rates from spot rates by solving for the
fixed coupon that would price the bond at par.

Graphical View
Par rates are averages of one or more spot rates and thus, in a typical upward slopping yield curve
environment, spot rates will lie above par rates.

6 Forward Curve

5.5
Spot Curve
5

4.5

4 Par Curve
3.5

2.5

1.5
1 3 5 7 10 12 14 16 18 20 22 24 26 28

Time

Shape of a Yield Curve

So far we have only looked at cases where the discount function is monotonically decreasing i.e. the
spot curve is upward slopping. However that doesn’t always have to be the case. If the spot curve is
inverted, the ordering of the curves gets reversed – the par curve is highest, the forward curve is
lowest and the spot curve is in the middle. In the case that the discount function is flat, all three
curves will lie horizontally on top of each other.

The curvature or shape of the yield curve is generally believed to be driven by two major factors:

Future Rate Expectations

It is clear that the market’s expectation of the level of future interest rates must impact the shape of
today’s yield curve. If markets expect interest rates to rise (fall) in a parallel fashion, today’s yield
curve will be linearly increasing (decreasing). If markets expect interest rates to rise and flatten,
today’s yield curve will be upward slopping and concave. And lastly, if markets expect rates to remain
unchanged, the yield curve will be horizontally flat.

Risk Premium
Investors tend to demand a liquidity premium for investing in securities with longer maturity. Not only
are they exposed to greater credit risk, but they also need to take into account the opportunity cost of
making a longer term investment. The positive risk premium that investors charge pushes the yield
curve to be upward slopping.

©2009 BlackRock, Inc. All Rights Reserved. 12

Chapter 3 – Duration and Convexity
Duration and convexity are two of the most important risk analytics that help investors understand how
sensitive their securities prices are to changes in interest rates. Even though the concepts themselves
are quite simple, understanding duration and convexity can quite quickly become confusing when
special features such as call/put options or prepayment options are thrown in.

Objectives of the chapter:

• Gain a solid mathematical as well as graphical understanding of duration and convexity.

• Understand how to estimate a securities change in price due to change in interest rates.

A Graphical View
Duration and Convexity are risk measures that capture the price sensitivity of a security to changes in
interest rates. Let us begin by first taking a graphical look at the price-yield relationship of fixed
income instruments. Plotting the price on the x-axis and the yield (or interest rate used for
discounting) on the y-axis could result in four common outcomes.

Duration
The duration of a security, at a given yield, is the first order price sensitivity to changes in yields i.e.
duration is the slope of the price-yield graph. A security has positive duration if its price increases as
yields decrease. A security has negative duration if its price increases when yields increase. We can
find the duration of an instrument by drawing a tangent line to the price-yield graph of that security. If
the slope is negative, it has positive duration and if the slope is positive, it has negative duration.

− slope ⇔ + Duration
+ slope ⇔ − Duration

The negative slopes of the top two graphs imply that they represent price-yield relationships of
securities with positive duration.

©2009 BlackRock, Inc. All Rights Reserved. 13

Chapter 3: Duration and Convexity BlackRock Solutions

The bottom two graphs have positive slopes, i.e. price and yield move in the same direction. These
graphs thus represent the price-yield relationship of securities with negative duration.

Convexity
Convexity measures the curvature or the second order change in price with respect to yields. A security
has positive convexity if its price increases more (decreases less) than duration would imply when
yields decrease (increase). Conversely a security has negative convexity if its price increases less
(decreases more) than duration would imply when yields decrease (increase). Graphically, positive
convexity can be observed If the price-yield curve is convex (the curve smiles) and negatively convex
securities have concave price-yield graphs (the curve frowns).

The two graphs on the left are positively convex while the two graphs on the right are negatively
convex. Drawing the tangent line we can see that positive convexity always helps the investor while
negative convexity hurts the investor. The graphs on the left are always above the tangent line while
the graphs on the right are always below it.

©2009 BlackRock, Inc. All Rights Reserved. 14

Chapter 3: Duration and Convexity BlackRock Solutions

Duration and Convexity graphs

Mathematical Derivation
We now derive duration and convexity mathematically using some basic derivative math and Taylor
series expansion. The objective of this section is to derive an equation that explains changes in price
as a function of duration and convexity.

Definitions:

• Duration: Percentage change in price for a 100 basis point change in rates.
Mathematically this is the first derivative of the price-yield relationship.
1 ∂P
Duration = −
P ∂Y
• Convexity: Convexity is the second derivative of the price-yield relationship.
1 ∂2P
Covexity =
P ∂Y 2
• Where P is the price of the bond and is a function of the yield Y.
N
CFt
P(Y ) = ∑ , t = 1,2,..., N
t =1 (1 + Y ) t

©2009 BlackRock, Inc. All Rights Reserved. 15

Chapter 3: Duration and Convexity BlackRock Solutions

Duration is the first derivative of the price function with respect to yield, which we can link to the
tangent line of the price-yield graph. Convexity is the second derivative of price with respect to yield,
which we linked to the curvature in the price-yield graph.

Taylor Series
Taylor series is a mathematical representation of a function as a series of infinite derivatives to a
single point. Given that the price of a fixed income security is a function of yield ( P = f (Y ) ) we can
use Taylor series expansion to represent the change in price ΔP as a function of its derivatives.

Let us start with the graph below that represents the price-yield relationship of a fixed income security
with negative duration and negative convexity. Assume that we know yield a and the corresponding
price f (a ) . Given the price function f ( y ) , we want to know how much the price will change if the
yields moves from a to x .

Using Taylor expansion, we can estimate the price change by representing the final price at the yield
of x ( f (x ) ) as a function of derivatives at a .

1 1 1
f (x) = f (a) + f ′(a) ⋅ ⋅ (x − a)1 + f ′′(a) ⋅ ⋅ (x − a)2 + ...+ f n (a) ⋅ ⋅ (x − a)n , n = ∞
1! 2! n!
Ignoring the higher order terms that do not contribute significantly to the price change, we can
rearrange and simplify the formula to
1
f ( x) − f (a) = f ′(a) ⋅ ( x − a) + f ′′(a) ⋅ ⋅ ( x − a) 2
2
Where ( x − a ) is the change in yield Δy and ( f ( x ) − f ( a ) ) represents the change in price ΔP .

1
ΔP = f ′(a ) ⋅ Δy + ⋅ f ′′(a ) ⋅ ( Δy ) 2
2

©2009 BlackRock, Inc. All Rights Reserved. 16

Chapter 3: Duration and Convexity BlackRock Solutions

f ′(a ) is 1st derivative of the price function with respect to yield and f ′′(a ) represents the second
derivative of the function with respect to yield.

∂P
f ′( a ) =
∂y y=a

∂2P
f ′′(a) = 2
∂y y =a

Dividing each side of our price change equation by the known price at the yield of a (or f (a ) ), our
equation reduces to

ΔP 1 ∂P 1 1 ∂2P
= ⋅ ⋅ Δy + ⋅ ⋅ 2 ⋅ (Δy ) 2
P P ∂y y =a 2 P ∂y y =a

Using the mathematical definition of duration and convexity discussed at the start of this session, we
can re-write the formula to estimate the price change of a fixed-income instrument due to movement
in interest rates.

ΔP 1
= − Duration ⋅ Δy + ⋅ Convexity ⋅ (Δy ) 2
P 2

1
ΔP = − Duration ⋅ P ⋅ Δy + ⋅ Convexity ⋅ P ⋅ (Δy ) 2
2

Duration Drift
We know that duration is the first derivative of price with respect to yield and that convexity is the
second derivative of price with respect to yield. So what is the relationship between duration and
convexity? Is convexity the first derivative of duration with respect to yield? Well... not quite.

While we won’t work through the math here, taking the derivative of duration with respect to yield
throws out an extra term besides convexity. The negative of the change in duration for a 100 basis
point change in interest rates is known as duration drift. A positive drift is a desirable feature in a
bond since its duration will increase as markets rally, thereby lending itself to further gains as markets
rally even further.

∂Dur
Duration Drift = − = Convexity − (Duration) 2
∂y

©2009 BlackRock, Inc. All Rights Reserved. 17

Chapter 4 – Interest Rate Derivatives
A derivative is a financial instrument whose value is derived from the price of an underlying asset. For
example, you could create a derivative instrument that pays the holder $100 every time it rains on a
Sunday. In general we can split up derivatives into two kinds, exchange-traded and over-the-counter.
Exchange-traded derivatives are standardized instruments that trade via a specialized exchange such as
the Chicago Mercantile Exchange. The exchange acts as an intermediary for all transaction thus
reducing counterparty risk and increasing liquidity in the instrument. In contrast, over-the-counter
(OTC) derivatives are privately negotiated contracts between parties and thus do not need to have
standardized terms. In this chapter we will introduce a number of common derivatives that directly or
indirectly depend on movements in interest rates.

Objectives of this chapter:

• Understand the mechanics of a few OTC instruments: FX contracts, swaps, caps and floors.
• Understand market conventions and the intuition behind exchange traded futures.

Pricing derivatives
Most of the derivatives we will cover in this chapter are structured in such a way that no money
changes hands at inception. This means that the expected value of the cash flows that the two parties
agreed to exchange during the life of the derivative is equal at the start of the contract. Both at
inception and during the life of the security, the value of a derivative is determined by finding its no-
arbitrage price.

Arbitrage is an opportunity to make money without taking on any risk. For example, if you notice that
you can buy XYZ company’s stock from party A for 10$ and immediately sell it to party B for 12$ you
spotted an opportunity for arbitrage. You can make 2$ per share without taking on any additional risk.
If indeed such an opportunity existed, by trying to take advantage of it, market participants would buy
as many shares as possible from party A thus driving up the quoted price. They would then try and sell
all of them to party B and by doing so would be driving the price that they could sell them for down.
This market activity would lead to both quoted prices settling down somewhere between $10 and $12.
Supply and demand would create a price that would be a no-arbitrage price – a “fair” price determined
by market forces.

FX Forwards
An FX forward is an OTC derivative where two parties agree to exchange, at some future point in time,
a fixed amount of one currency for another currency at an exchange rate that is fixed today. The
exchange rate specified in the transaction is called the FX forward rate. Two main reasons for entering
into an FX transaction would be to hedge out currency risk or to place an exchange rate bet. For
example, let’s assume that BlackRock anticipates receiving a payment of 80 GBP a year from now. In
order to hedge their exposure to currency risk, they could enter into a FX forward transaction with the
following terms: One year from now BLK will pay its counterparty 80 GBP in exchange for 120 USD. This
transaction implies that both BLK and its counterparty believe that the USD/GBP FX rate one year from
now is going to be 1.5 USD/GBP. One year from now, if the realized FX rate is different from 1.5, one
of the two parties will lose money on the transaction.

So how did BLK and the counterparty agree that 1.5 was a fair one-year FX forward rate? Forward FX
rates are observable in the market and are dictated by supply and demand as well as the prevailing
interest rates in the two currencies. FX rates must move in sync with interest rates in order to prevent
arbitrage. This is best understood looking at an example. Let’s assume that the spot USD/GBP
exchange rate today is 1.54 USD per GBP i.e. an investor would be indifferent between owning 100 USD
or 65 GBP today. We also assume that the prevailing one-year interest rate available to investors is 5%

©2009 BlackRock, Inc. All Rights Reserved. 18

Chapter 4: Interest Rate Derivatives BlackRock Solutions

for GBP deposits and 3% for USD deposits. Given this information, we can come up with a no-arbitrage
implied forward FX rate by looking at how much cash we would have at the end of one year if we held
the cash in deposit accounts.

Today One Year

100 USD 3% 100 (1+3%) = 103 USD

65 GBP 5% 65 (1+5%) = 68.25 USD

Today an investor with a one-year horizon could invest 100 USD at 3% for a year or alternatively could
immediately exchange 100 USD into 65 GBP at the spot exchange rate and invest 65 GBP at 5% for a
year. The spot FX rate and the interest rate markets in the two currencies imply that, one year from
now an investor should be indifferent between owning 103 USD or 68.25 GBP. Hence the implied
forward one-year FX rate is 1.51 (= 103/68.25) USD/GBP.

The above example illustrates the relationship between interest rates and FX rates called the interest
rate parity.

Interest Rate Swaps

A swap is an agreement between two parties to exchange cash flow streams at defined intervals in the
future. Swaps are OTC instruments and the details pertaining to how the cash flow payments on each
leg are computed are specified in a document called a term sheet.

Since swaps are over-the-counter instruments, the terms of swaps are not regulated by an exchange.
However, markets do have a notion of the most commonly traded type of swap referred to as a
“vanilla” interest rate swap. In the USD market, a vanilla interest rate swap refers to a swap where
one party receives a fixed rate semi-annually and pays a floating rate quarterly. The floating rate in a
vanilla USD swap is 3 month USD LIBOR. The amount on which interest payments are based is called the
notional principal (“notional” because it’s not real, doesn’t really change hands) and the time period
during which the cash flow exchanges take place is called the swap’s tenor.

London Inter-Bank Offered Rate

The most common rate used as an index for the floating leg of interest rate swaps is
LIBOR. LIBOR stands for London Inter-Bank Offered Rate and it is a reference rate based
on the interest rates at which banks (in London) offer to lend funds (in various
currencies) to other banks. LIBOR rates for different maturities (1M, 3M, 6M etc) are
published on a daily basis by the British Bankers Association (BBA) based on rates
received from sixteen contributing banks. The actual reported value is the average of
the middle eight contributions. LIBOR rates are computed for a number of currencies
such as USD, GBP, and JPY etc.

©2009 BlackRock, Inc. All Rights Reserved. 19

Chapter 4: Interest Rate Derivatives BlackRock Solutions

Let us analyze a plain vanilla USD swap that pays a fixed coupon of 4% on $1,000 notional in exchange
for 3M LIBOR.

Receive Leg
Swap Cash Flows
Pay Leg
25
20
15
10
5
0
-5
-10
-15
-20
3 6 9 12 15 18 21 24
Time in Months

Receive Leg
The receive leg of an interest rate swap makes semi-annual payments at a fixed rate. In our example,
each semi-annual payment is computed as follows:

Fixed Coupon
Notional *
2
4%
1,000 * = $20
2
Technically, there is no exchange of notional principal at the end of a vanilla swap term since the
amount is the same for both parties (unlike, say, for currency swaps). But, if one were to visualize the
notional being exchanged at the end, then the receiving leg cash flow profile would be simply that of a
fixed-coupon bond. The receive leg has positive duration because the price of fixed-coupon bond falls
as interest rates rise.

Pay Leg
The pay leg of an interest rate swap makes quarterly floating rate payments. Each payment of a vanilla
USD swap is based on a coupon that is determined by what 3M USD Libor is at the coupon
determination date corresponding to the coupon payment. In order to determine when and how much
the floating leg pays, we need to be aware of a number of different dates: coupon determination date,
coupon reset date and coupon payment date. For a plain vanilla swap, coupon reset date is 3M prior to
the coupon payment date, i.e. reset date = previous coupon payment date. In turn, the relationship
between the coupon determination date and coupon reset date is as follows:

Coupon determination date = coupon reset date – lookback

The look back in the USD market is 2 business days.

For example, assume that our swap just made a floating rate payment on 7/1/2008. Let us compute
the next cash flow payment due on 10/1/2008

©2009 BlackRock, Inc. All Rights Reserved. 20

Chapter 4: Interest Rate Derivatives BlackRock Solutions

Last coupon payment date: 07/01/2008

Next coupon payment date: 10/01/2008
Reset date for next coupon payment: 07/01/2008
Coupon determination date: 06/27/2008

Assuming that 3M Libor on 06/27/2008 was 3.5%, the coupon payment on 10/1/08 is computed as
follows:

Floating Coupon
Notional *
4

3.5%
1,000 * = $8.75
4

Floating rate for time T + 3mo is determined at time T

9
8
Projected 3M Libor

7
6
5
4
3 Coupon
2 Coupon
1
0
T T+3mo T+6mo

In order to estimate the future cash flows of the pay leg, we first need to derive the 3M Libor forward
curve i.e. we need to have a projection for the trajectory of 3M Libor over time. The forward curve
used is derived from a yield curve constructed from at-the-money swaps trading in the market.

What is the value of the floating leg of the swap? In chapter 1 we proved that a bond prices at par
when its yield is equal to its coupon payment. Or one could say that a bond prices at par if the interest
rate used to determine the discount factor is the same as the interest rate used to determine the
coupon payment. The floating leg of a swap, being discounted using a curve constructed from LIBOR
rates, will price at par for the same reason.

Swap Rates
The value of a swap is the difference in price of the receive leg and pay leg. As discussed before, swaps
are usually initiated at-the-money i.e. no money is exchanged upfront. This means that the value of
the fixed leg must equal the value of the floating leg at inception. Given that the floating leg is priced
at par, the fixed coupon is chosen so that the fixed leg also prices at par (assuming we discount using a
LIBOR curve) giving the swap a zero net present value.

©2009 BlackRock, Inc. All Rights Reserved. 21

Chapter 4: Interest Rate Derivatives BlackRock Solutions

A Swap rate is the fixed rate of a vanilla interest rate swap that causes the swap to have a value of
zero. Swap rates form the basis of the swap curve (usually referring to a par curve). The swap curve is
a yield curve constructed based on marked quoted ATM swap rates that would serve as an indicator for
the returns available by investing in swaps. The market swap curve (also know as the LIBOR curve) is
used for projecting different LIBOR rates into the future as well as discounting the cash flows of
interest rate swaps back to today.

In addition to plain vanilla fixed-for-float swaps, there are a number of other commonly traded swaps
in the market. A basis swap is an interest rate swap where both legs reference a floating index. It
derives its name from the notion of basis risk, which in this context is the risk that results from being
exposed to two different interest rates which are not perfectly correlated. A currency swap is a swap
where interest and principal in one currency are exchanged for interest and principal in another
currency. It requires that a notional amount be specified in each of the two currencies. Usually, the
principal amounts are chosen such that the value of the swap is zero at the initiation point. An
inflation linked swap is a swap where one party pays a fixed rate in exchange for receiving the return
on inflation.

Interest Rate Caps and Floors

An interest rate cap provides insurance against the underlying floating interest rate rising above a
certain level while an interest rate floor protects against falling interest rates. Caps (floors) have a
fixed strike which defines the level above (below) which the cap (floor) gets paid. Unlike the previous
derivatives we discussed, caps and floors fall under the category of options that counterparties can
buy/sell to each other in order to make speculative bets or hedge against moving interest rates. Like
all options, caps and floors do not price at zero at inception.

Cap
One can think of the payment mechanics of a cap as being similar to the floating leg of a vanilla
interest rate swap. The concept of coupon determination dates, reset dates and lookback hold true for
caps. The main difference is that the coupon paid is a function of the strike level. For example, assume
you bought a 1,000$ 3M Libor cap with a strike of 4%. Assuming that 3M Libor on the coupon
determination date was 4.5%, the cash flow paid on the coupon payment date is defined as follows:

Max(reset _ rate − strike,0)

Notional *
4
Max(4.5% − 4%,0)
1,000 * = $1.25
4
Note that if 3M Libor on the coupon determination date is less than the strike, then the cap doesn’t
pay out a cash flow for that coupon payment period. An interest rate cap can be analyzed as a series
of call options on interest rates which exist for each period the cap agreement is in existence. Each
such option is called a caplet. For example a 10 year cap with a 3 month tenor would have 39 caplets.
What happened to the 40th caplet?

Floor
Floors are similar in structure to caps except that payments are received when the reset index falls
below the strike level.
Max( strike − reset _ rate,0)
Notional *
4

©2009 BlackRock, Inc. All Rights Reserved. 22

Chapter 4: Interest Rate Derivatives BlackRock Solutions

Forwards and Futures

A forward contract is an OTC agreement to buy a security (or commodity) in the future at a pre-
specified price. The FX forward contract discussed at the start of the chapter is an example of a
forward contract. An important point to remember with forwards is that there is usually only one
payment at the end of the contract. For example, a farmer can enter into a forward contract to
purchase fertilizers one year from now at a price fixed today. The only payment that happens is one
year from now when the actual purchase is made. Depending on the actual price of fertilizers a year
from now, the farmer will have made or lost some money by entering into the contract.

Futures contracts are similar in concept to forward contracts with a few key differences:

• Exchange Traded: Futures contracts are exchange-traded i.e. an exchange acts as an

intermediary for all transactions thus reducing counterparty credit risk.
• Standardization: As a by-product of being exchange-traded, futures contracts have standard
terms and conditions (contract sizes, minimum price movements, etc) that all market
participants have to adhere to. A standard naming convention is also followed for all exchange-
traded futures.
• Mark to Market: Unlike forwards, where a single previously agreed upon payment is made at
the end of the contract, in a future’s contract, money exchanges hands on a daily basis as the
underlying changes in value. Hence, while the PnL of a forward is only realized at the
termination of the contract, a future’s PnL is settled on a daily basis thus reducing
counterparty default risk.

All the above characteristics lead to increased liquidity that make the futures market a good gauge of
market sentiment.

Naming Convention
Exchange traded futures have a standard four-character code to identify each contract.

Character Description
1st,2nd Ticker
3rd Expiration Month
4th Expiration Year

The two-letter ticker identifies the underlying that the future is referencing. Below are a few
commonly traded USD futures tickers:

Ticker Description
TU 2-year USD Treasury Bond Future
YR 3-year USD Treasury Bond Future
FV 5-year USD Treasury Bond Future
TY 10-year USD Treasury Bond Future
US 30-year USD Treasury Bond Future
ED EuroDollar Future
FF Fed Funds Future

©2009 BlackRock, Inc. All Rights Reserved. 23

Chapter 4: Interest Rate Derivatives BlackRock Solutions

The third character, represented by a letter, defines the month in which the futures contract expires.
While there is a letter to represent every month of the year, the most common expirations are:

Letter Expiration Month

H March
M June
U September
Z December

The last character, which is a digit, represents the futures’ expiration year within the current decade.
For example 8 refers to 2008, 0 refers to 2010.

For example TYZ8 is the ticker for the 10-year USD Treasury Bond Future that expires in December
2008. Once we pass the expiration date for this future TYZ8 will start referring to the future that
matures in December 2018. We will now cover the mechanics of two of the most commonly traded
futures: Eurodollar futures and Treasury Bond Futures.

Eurodollar futures
Eurodollar contracts are the most popular futures contracts on short-term interest rates. A Eurodollar
future is an instrument that allows investors to bet on (or hedge against) what 3M Libor will be at some
point in the future. The market expresses the forward 3M Libor rate (on which the future is set) via its
price: (100 – quoted market price) is the implied 3M Libor rate prevailing in the market. The contract
size of the future is $1,000,000.

Let’s look at the example of EDM9 to understand the mechanics of this future. EDM9 is a Eurodollar
future expiring on June 15th 2009. Let us assume its current market price is 98. The price of 98 implies
that currently the market thinks that 3M Libor on June 15th 2009 will be at (100-98) or 2%. The
counterparty that is long the future is essentially committing to lend $1,000,000 for 90 days from the
expiration of the futures contract at 2% simple interest. While no actual money is lent/borrowed at
the end of the futures contract, an EDF is structured to capture the same economic value of the
transaction described above.

Now let’s say that the day after the contract is bought, the market price of EDM9 moves up to 99 (i.e.
the implied forward 3M Libor rate is now 1%). From the perspective of the person who is long the
contract, how much does the economic value of the contract change by? Given that there was a
previous commitment to lend $1,000,000 for 90 days at 2%, and the going market rate has now dropped
to 1%, the counterparty who is long the contract must have made money. The amount of “profit” made
can be derived as the difference in the value of interest received by lending the money for 90 days at
the two different rates.

(99 − 98) 90
1,000,000 * * = $2,500
100 360
So for every basis point move in price (equivalently a basis point move in implied 3M Libor rate), a
mark to market of 25$ needs to be paid from one counterparty to another.

Fed Funds futures are similar in concept to Eurodollar futures with the following differences:

• The underlying rate is the simple average of the overnight Fed Funds effective rate in the
delivery month (so one can think of the underlying as a 30-day loan). The Fed Funds rate is the

©2009 BlackRock, Inc. All Rights Reserved. 24

Chapter 4: Interest Rate Derivatives BlackRock Solutions

overnight rate at which banks lend cash to each other at the Federal Reserve to meet reserve
requirements set by the Fed.

• The notional contract size is $5,000,000

• For every percentage point move in price, the mark to market is

1 30
5,000,000 * * = $4,167
100 360

Treasury Bond Futures

Unlike Eurodollar and Fed Funds futures, Bond futures have an actual deliverable. A Bond future is a
contract where the seller of the contract, or the “short”, agrees to deliver a treasury bond to the
buyer of the contract (“long”) at a formula-based price during the expiration month of the contract.

Mechanics of Treasury bond Futures

• Bond futures do not have one underlying. Instead there is a basket of underlying bonds (defined
by a set of maturity-based rules) and at maturity, the counterparty that is short the contract
has the option to deliver any bond from the basket. This option is often referred to as the
delivery option.

• The seller of the contract is obligated to deliver a particular quantity of a bond from the
contract’s basket anytime during the delivery month. The option to deliver anytime during the
month is know as the timing option.

• The price paid by the long for the delivered bond is defined by the following formula:

Delivery Price = Futures Market Price * Conversion Factor

• Conversion factors are selected in order to adjust for the fact that different bonds in the
delivery basket have different coupons and thus different economic values. Conversion factors
are computed by the futures exchange and are set so that all the underlying bonds have an
equal probability of delivery at the time the future starts trading. The higher the coupon of the
bond, the higher the conversion factor. Without the conversion factor, all things being equal,
the short will want to deliver the bond with the lowest coupon (i.e. the cheapest price).
However the higher conversion factor gives the short more incentive to deliver a higher coupon
bond as he will get paid a higher price in return.

• Conversion factors and delivery baskets are designed to increase liquidity in the future by
preventing market participants from putting a squeeze on the delivery bond. For example, if
there was only one bond that could be delivered, a big market participant could go long the
future and at the same time buy up large quantities of that bond’s issuance. At the time of the
future’s expiration, market participants who sold that futures contract would be scrambling to
buy the bond in order to deliver them in to the contract. This would allow the long participants
to sell both their futures and their bonds at prices that are well above fair market value.

For example, the seller of one contract of TYZ9 (10-year USD Treasury Bond Future) is obligated to
deliver $100,000 of a US treasury note maturing at least 6.5 yrs but not more than 10 yrs from
December 1st 2009. The actual delivery can be made anytime in the month of December 2009.

©2009 BlackRock, Inc. All Rights Reserved. 25

Chapter 4: Interest Rate Derivatives BlackRock Solutions

Cheapest to Deliver

In order analyze the risk characteristics of bond futures, we need a way to estimate which bond is
going to be delivered. While any bond in the delivery basket can be delivered, economically a rational
investor will deliver the bond that has the lowest net cost to him. The cost of delivery measures how
much it costs the seller of the futures contract to deliver the bond. It will differ from bond to bond
and the bond with the lowest cost of delivery is called cheapest-to-deliver.

Let’s analyze the cost-benefit equation of the contract seller. The counterparty that is short the
contract has to purchase the deliverable bond from the open market and then deliver it into the
future. So the cost associated with this transaction for the short is

Cost of Delivery = Bond Market Price – Price paid by long counterparty

= Bond Market Price – (Futures Market Price * Conversion Factor)

This cost of delivery is called the Gross Basis. For example, below are 3 bonds out of TYH9’s basket of
deliverables. According to this, the 5.125 bond is the cheapest-to-deliver. As we can see, even if
someone decided to squeeze the market by buying up 5.125s, the most they could extract is 3.196 -
2.946 = 0.25 before people would move on to buy the next cheapest-to-deliver issue, i.e. 4.625s. In
the absence of conversion factors we can see that among these three issues, 4.625s would be the
cheapest to deliver and the difference between costs of delivery would be much bigger.

Coupon Maturity Bond Price Futures Price Conversion Factor Gross Basis
5.125 5/15/16 110.527 113.171875 0.9506 2.946
4.625 2/15/17 106.827 113.171875 0.9157 3.196
4.5 11/15/15 107.479 113.171875 0.9202 3.338

While the Gross Basis is a good first order measure of which bond is going to be delivered, the short
really cares about the price of the bonds at the time of delivery. Using the carry of the bond, we can
estimate what the Forward Bond Price will be at the time of delivery. The cost of delivery using the
forward bond price is called the Net Basis.

Net Basis = Forward Bond Price – (Futures Market Price * Conversion Factor)

The bond with the smallest Net Basis is the cheapest-to-deliver!

©2009 BlackRock, Inc. All Rights Reserved. 26

Chapter 5 – Options
Freedoms and rights are privileges, so much so that in the financial world, we pay for them. Welcome
to the world of options, contracts that offer one the right to enter into a transaction in the future on
some underlying instrument, without the obligation of doing so. Option contracts exist on a wide range
of both equity and fixed-income underlying securities.

Objectives of this chapter:

• Understand the concept and uses of both equity and fixed-income options.
• Understand the main variables that contribute to the pricing of options.

Option Basics
When one buys an option, he or she is purchasing the right to buy or sell an underlying asset (for
example, a stock) on some future date or dates at a pre-specified price. There are two broad
categories of options: calls and puts. A call option gives the holder the right to buy the underlying
instrument at a given strike price on the exercise date. While a put option gives the option holder the
right to sell the underlying at the strike price on the exercise date.

Option contract terminology

• Strike Price: This is the price the holder of a call (put) option will pay (receive) to buy
(sell) the underlying security on an exercise date.
• Expiration date: This is the maturity of the option contract. All optionality terminates
after this date.
• Exercise date: This is the date(s) on which the option can be exercised.
• Exercise type: The exercise type, along with the expiration date, determines when the
option may be exercised. The three most common exercise types are
o European - These options have one exercise date, at the expiry of the option
contract.
o American - These options can be exercised at any point in time up to the
expiration date.
o Bermudan – These options can be exercised on pre-specified exercise dates before
the expiration of the options contract. Usually the dates are at regular intervals,
quarterly for example, from option issuance to expiration.
• Moneyness:
o At-the-money (ATM) – An at-the-money option is one where the strike is equal to
the underlying price.
o In-the-money (ITM) – An option which, if exercised today, would result in a non-
zero payoff to the option holder i.e. the intrinsic value is greater than 0.
o Out-of-the-money (OTM) – An out-of-the-money option is one which, if exercised
today, would result in zero payoff to the option holder i.e. the intrinsic value is 0.

At this point we need to point out a key difference between fixed income and equity options. For the
former, since the underlying is a fixed income instrument, such as a swap or a bond, we need to also
take into account the maturity of the underlying. European fixed income options give the holder the
right to enter into the underlying on a specific date before the maturity of the underlying. In this case
the option expiration date is the same as the exercise date. American and Bermudan fixed income
options can typically be exercised any time after the exercise date up to the maturity of the

©2009 BlackRock, Inc. All Rights Reserved. 27

Chapter 5: Options BlackRock Solutions

underlying. In this case the option expiration date is the maturity of the underlying. The longer an
investor waits to exercise an American fixed income option, the shorter the length of the underlying he
enters into.

Equity options are simpler in nature. The option expiration date and the exercise date are the same for
European equity options. American and Bermudan equity options can be exercised any time from the
issuance up to the expiration date. While the basic concepts are the same, we will focus on analyzing
equity options for the majority of the chapter since they are easier to understand. In order to
understand the value proposition of buying an option we need to first take a look at option payoff
functions upon exercise. Let us consider simple European equity call and put options.

Call Option
A call option gives the holder the right to buy the underlying at the strike price K at expiration. The
investor will, of course, only exercise the option if the market price S for the underlying is higher than
the strike, earning him the difference between the two price levels. The payoff is thus floored at 0 and
can be expressed as

Payoff of Call = (S − K ) = max (S − K ,0 )

+
Call Option Payoff

0),
-K
(S
ax
m

K
0

Stock Price S

Put Option
The complement to the call option is the put option, which gives the holder the right to sell at the
strike price and is only profitable if the market price of the underlying is lower than the strike price.
The payoff is also floored at zero and can be formulated as

Payoff of Put = (K − S ) = max (K − S ,0 )

+
Put Option Payoff

m
ax
(K
-S
, 0)

K
0

Stock Price S
©2009 BlackRock, Inc. All Rights Reserved. 28
Chapter 5: Options BlackRock Solutions

The part of the payoff diagrams where the option is not exercised, i.e. the zero payoff scenarios, is
marked slightly below the 0 line in both graphs. This is done to reflect the price that the investor pays
to purchase the option. These graphs are useful for visualizing option strategies, in which three or
more options of different strikes are traded simultaneously to express a specific view.

Option Valuation
With the basics covered, we can begin to try to understand how an investor would decide how much to
pay for an option. In order to understand the valuation, let us take the example of a simple European
equity call option with a strike price of 100 and an exercise date one year from now. Let’s assume that
the underlying stock is currently trading at $103.

The value of an option has two components: Intrinsic Value and Time Value.

Intrinsic Value
The intrinsic value of an option is the payoff of the option assuming it was exercised today. This is the
value of the option assuming that the price of the underlying stock does not change i.e. assuming zero
volatility in the stock.

In our example, assuming that the underlying stock is currently trading at $103, we can compute the
intrinsic value as follows

Option Expiration Date: T + 1Y

Strike Price: 100
Underlying Price: 103
Intrinsic Value: max(103-100,0) = $3

Time Value
The second component of an options price is time value. Time value captures the uncertainty in the
price of the underlying stock and the effect that uncertainty has on the value of the option. Time value
is computed as the difference between the options market price and its intrinsic value. While the
intrinsic value of our option today is $3, when deciding how much to pay for this option, an investor
must attempt to estimate what the value of the underlying will be on the exercise date a year from
now. In other words, the investor has to account for volatility of the underlying. If an option is not in-
the-money today, it may be in-the-money by the time the option expires.

Volatility
Volatility is defined as the annualized standard deviation of underlying stock price i.e. it is the
dispersion of the price of the underlying around its mean. The simple one period illustration below
shows us how the volatility of the underlying impacts the value of the option. Volatility is a measure of
possible dispersion and this example attempts to capture that essence.

The tree to the left represents today’s stock price and two possible values it could take on the exercise
date. Assume, for example, that we start out with the equity priced at $103. In the next time step, one
of two possible scenarios occurs, either the price increases, or it decreases. The corresponding tree to
the right represents the value of exercising the option in these two different scenarios. On the
expiration date, the option has no time value left and its price is its intrinsic value at that time.

©2009 BlackRock, Inc. All Rights Reserved. 29

Chapter 5: Options BlackRock Solutions

Stock Price Tree Option Value Tree

T T +1Y T T +1Y

$110 max(110-100,0) = $10

$103 ?

$96 max(96-100,0) = $0

From the above trees we could conclude that in the up-scenario the option will be worth $10 and in the
down-scenario the option will be worth $0. Assuming that either move has a 50% probability, we could
estimate the value of the option today to be the expected discounted value of the payoffs in the two
scenarios

Expected discounted value of option = (50% * $10 + 50% * $0) * DF1 year

Undiscounted value of option today = (50% * $10 + 50% * $0) = $5

The time value of the option is the difference between the fair market price of the option and its
intrinsic value. Assuming no discounting and that the underlying stock moves as described in the trees
above (i.e. given our volatility assumption) the time value of our option is $2 (=$5-$3).

Now let’s take a look at what the trees would look like if we increased the volatility of the underlying
stock.

Stock Price Tree Option Value Tree

T T +1Y T T +1Y

$118 max(118-100,0) = $18

$103 $9

$88 max(88-100,0) = $0

There are a number of important observations we need to make here

• Assuming that today’s stock price embeds all the markets information about the company,
we can view the current stock price as the expected stock price across all the different
scenarios that the stock tree might take. In this case, the undiscounted value of the option

©2009 BlackRock, Inc. All Rights Reserved. 30

Chapter 5: Options BlackRock Solutions

today cannot be less than today’s intrinsic value. We think of the intrinsic value as the
minimum amount the investor should pay to enter the option.
• As volatility increases, the expected stock price a year from now in the two scenarios is
more dispersed. The up-scenario stock price goes higher and the down-scenario stock price
goes lower.
• While the stock price in both scenarios changes as volatility increases, the option value at
expiration is impacted only in the up-scenario. The option still expires worthless in the
down scenario i.e. decreasing the stock price further below the strike does not impact the
option price in that scenario. The option price in the up-scenario increases linearly with
increases in stock price.
• As volatility increases, the expected discounted value of the option increases due to the
one sided nature of the security. The downside of buying an option is limited since you
cannot lose more than the option premium while upside is unlimited. The more volatile the
value of the underlying is, the more valuable the option is.

Volatility and time to expiration go hand in hand. The greater the time to expiration, the more impact
volatility is going to have on the option value. An investor who owns a deep out-of-the-money option
expiring tomorrow will value his option at its intrinsic value. Because no matter how volatile the
underlying is, the probability of the option going back in-the-money is very low.

Black-Scholes
What other factors determine the value of an option? The table below summarizes the impact of
increasing some of these factors on the value of the option.

Value of Option...
Increasing Factor Call Put
Volatility Increase Increase
Underlying Price Increase Decrease
Strike Decrease Increase
Time to expiry Increase Increase
Risk free rate ? ?

This brings us to the Black-Scholes formula. The Black-Scholes formula is a closed form solution that
has become the market standard formula used to value European equity options. The formula makes
assumptions about the distribution of the underlying stock price over time and allows investors to put a
value on their options. The Black-Scholes formula takes in five inputs: underlying price (S), strike (K),
time-to-expiration (T), volatility (V) and risk free interest rate (r).

BS(S0,T,K,V,r) = Option Value

Note that all the variables above are either directly observable in the market place or are defined in
the option contract except for the volatility. Indeed options trading is essentially volatility trading and
is often referred to as such.

©2009 BlackRock, Inc. All Rights Reserved. 31

Chapter 5: Options BlackRock Solutions

Historical vs. Implied Volatility

Volatility that is quoted in the market place can come from two different sources. In order to use
the volatility number correctly in a model, it is important to fully understand how the number
was derived.

Historical volatility refers to a volatility number computed using historical data. For example,
one could look at the time series of historical price or rate changes over the past year and
determine what the standard deviation of those moves is. While historical volatility is a useful
number from a risk management perspective, it is more common for investors to quote implied
volatility in the market place.

Implied volatility is the volatility number that, when inputted into a market standard formula
such as a Black-Scholes, will return a price that matches the market price of the option. Thus it is
the volatility implied by the market price.

Option Greeks
Similar to how an investor would use duration and convexity to gauge the risk of a fixed income
instrument to interest rates, the “greeks” refer to metrics that capture the sensitivities of the price of
an option to the various variables that enter the Black-Scholes equation.

Delta
Delta represents the sensitivity of the option price to the price of the underlying asset, or the first
order derivative of the Black-Scholes equation with respect to the stock price. This is one of the more
commonly hedged risk parameters. An option portfolio that is delta-hedged is supposed to be insulated
against small movements in the underlying price.

As an approximation, for a given movement in the underlying price, delta can be thought off as

Change in price of option

Delta =
Change in price of underlying

In extreme cases of deep in-the-money options, the option’s price will move in a 1:1 ratio with the
price of the underlying. In deep out-of-the money cases, the option’s price will not move as the price
of the underlying changes. Thus, the delta of a call option is bounded between 0 and 1 and the delta of
a put option is bounded by 0 and -1. Plotting delta versus moneyness (i.e. varying underlying price)
results in a typical S shaped curve with a delta of approximately 0.5 when the underlying stock price is
equal to the strike price K (at-the-money).

Delta of Call Option

0.9

0.8

0.7

0.6
Delta

0.5

0.4

0.3

0.2

0.1

0
K

Current Market Price (S)

©2009 BlackRock, Inc. All Rights Reserved. 32

Chapter 5: Options BlackRock Solutions

Gamma
Gamma is the second order derivative of the option price with respect to the underlying stock price.
Gamma tells the option holder how sensitive the delta is to changes in the underlying stock price

Change in Delta
Gamma ≈
Change in price of underlying

Since the delta itself changes as the underlying price changes, a delta hedged portfolio will have to be
continually rebalanced. Gamma-hedging helps in this regard since portfolios will be more protected
between the times when the asset moves and the delta-hedge is adjusted. Also, a higher gamma means
that the option is harder to delta-hedge and will need more frequent rebalancing.

Vega
This measure captures the sensitivity to volatility and is also known as kappa. Technically, vega is not a
letter in the Greek alphabet.

Theta
Theta measures the option’s sensitivity to the passage of time - it represents the time-decay of the
option price. As the option expiration approaches, there is less opportunity for the volatility to
increase the payoff of the option.

Rho
Rho measures the sensitivity to the underlying risk free rate.

Swaptions
We now move away from equity options and introduce fixed income options. The bread and butter of
fixed income options and fixed income volatility trading lies in the universe of swaptions, or options on
vanilla interest rate swaps. The volatility data from this market is an important input into a lot of our
risk models, including our interest rate lattices. Swaptions are traded over-the-counter and are used to
hedge convexity as well as to take a view on volatility.

Let us begin with some swaptions basics:

• A swaption is an option to enter into an interest rate swap at some point in the future.
Typically swaptions exercise into vanilla interest rate swaps and are all European in nature.

• The strike defined in the swaptions contract is the fixed rate on the underlying swap.

• The time frame of a swaption is defined by the option term and the tenor of the underlying
swap. For example, a 3x10 swaption is an option that expires 3 years from now and gives the
holder the right to enter into a 10 year swap. Thus the underlying swap will mature 13 years
from now.

• Similar to equity space, swaptions also have calls and puts.

o Call swaptions are referred to as receiver swaptions or right-to-receive (RTR)
swaptions. If exercised, the option holder enters into a vanilla interest rate swap
where he receives the fixed rate (i.e. the option strike) and pays the floating rate.

©2009 BlackRock, Inc. All Rights Reserved. 33

Chapter 5: Options BlackRock Solutions

o Put swaptions are referred to as payer swaptions or right-to-pay (RTP) swaptions. If

exercised, the option holder enters into a vanilla interest rate swap where he pays
fixed and receives float.

As a simple illustration, imagine an entity that knows that it will need a four-year interest rate swap
three years in the future. This could be because, three years from now, they anticipate receiving a
floating stream of 3M Libor cash flows that they want to swap to a fixed rate. In order to do this, they
will need to enter into a vanilla swap paying float and receiving fixed. One option would be for them to
enter into a forward-starting swap and lock in the fixed coupon today; let’s say at 5%. However, if the
market swap rate three years from now ends up being higher than 5%, then the forward swap
agreement would have been a poor choice as they could have been receiving a higher rate. And on the
flip side if the market rate ends up being lower than 5% then it would have been a good choice.

Whether they make money off it or not, entering into the forward locks them into the swap contract
for four years. If they do not wish to take this risk, as an alternative, they can enter into a 3x4 receiver
swaption struck at 5%. This gives them the option to enter into the swap 3 years from now if the strike
is higher than the market rate.

Swaption Valuation
Much like for equity options, the only uncertain parameter in swaption pricing is the volatility of the
underlying forward rate. All the other variables are either observable in the market, or are part of the
option contract. To determine the moneyness of our swaptions we can project forward what the ATM
rate might be on the exercise date. But it’s the uncertainty (volatility) in the forward ATM rate that
gives the swaptions their value. At expiry the ATM rate will become observable and the price of the
underlying swap will be fully determined. Prior to expiry, the option price depends largely on the
probability distribution of that ATM rate, and the volatility of that distribution. The former is a
component of whatever model you happen to be using to price the swaption, and the latter is often
implied from market data.

©2009 BlackRock, Inc. All Rights Reserved. 34

 Chapter 5: Options BlackRock Solutions

Normal vs. Lognormal (Black) Volatility

Volatility can be quoted in two different forms. Normal volatility represents volatility in absolute form
whereas lognormal volatility refers to percentage volatility. Let us look at a simple distribution of rates using
a normal and lognormal volatility assumption.

The graph below represents the evolution of an interest rate assuming a normal volatility of 10 basis points
with an initial value of 5%. In each subsequent time step the level of rates increases and decreases by exactly
10 basis points.

T T+1 T+2

5.2 %

5.1

5% 5%

4.9

4.8 %

Now contrast that with the tree below which represents the evolution of interest rates assuming a lognormal
volatility of 10% and the same initial value. In each subsequent time step the levels of rates increases and
decreases by 10% of current level in that state.

T T+1 T+2

5.5(1.1) = 6.05 %

5.5

5% 5.5 (.9) = 4.5 (1.1) = 4.95 %

4.5

4.5(.9) = 4.05 %

©2009 BlackRock, Inc. All Rights Reserved. 35

Chapter 6 – Corporate Credit and Credit Default Swaps
As with any investment, the potential reward is proportional to the risk the investor is willing to take.
Treasuries provide investors with risk free interest on their investments. Credit securities by contrast
are those securities that are exposed to credit risk, i.e. a risk that an issuer will not be able to fulfill
its obligations due to default, restructuring etc. The two most prevalent types of credit securities are
corporate bonds and credit default swaps (CDS).

Objectives of this chapter:

• Define what corporate bonds are and understand the mechanics of CDS
• Understand how to think about relative value when it comes to credit securities

Corporate Bonds
A corporate bond is a bond issued by a corporation as opposed to bonds issued by federal or local
governments, agencies, foreign governments, supranationals, etc. Corporate bonds typically have
maturities longer than a year. Corporate debt that has maturity less than 270 days at issuance is called
commercial paper. Bonds and equities are capital market securities - bonds represent a claim on assets
as opposed to equities which represent a claim on profits. The reason the corporate bond market
exists is because they offer investors higher returns than treasuries to compensate them for taking on
more risk.

Corporate Bond terminology

• Issuer: This is the corporation that is borrowing the money

• Capital Structure: Corporate debt and equity holder both have claim on the assets of the company.
Capital structure tiers define the seniority structure of the claims in the event of default. The
capital structure in the order of decreasing seniority is typically as follows:

o Senior Secured Debt - guaranteed by a specific asset

o Senior Unsecured Debt – most corporate bonds
o Junior Subordinate Debt
o Preferred Equity - in-between bonds and pure equities
o Equity Holders

• Ratings: Provides a notion of credit worthiness and issued by the big 3 ratings agencies – Moody’s,
S&P, Fitch
• Optionality – Majority of corporate bonds don’t have any embedded options. Those that do are
classified as:
o Callable: Bond can be called by the issuer prior to maturity (usually at par)
o Puttable: Bond can be put back to the issuer by the investor prior to maturity (usually at
par)
o Convertible: Bond where the investor has the option to convert debt into equity stake in
the company.

Now let us assume that a corporate bond and a Treasury bond have exactly the same coupon, maturity,
etc. Which of these should have a lower price? The market price of the corporate bond will, of course,
be less since it has credit risk. How do you measure how muck riskier a corporate bond is compared to
a Treasury? There are several related measures of this risk, the simplest of which is static spread.

©2009 BlackRock, Inc. All Rights Reserved. 36

Chapter 6: Corporate Credit and Credit Default Swaps BlackRock Solutions

Suppose we have a bond that matures a year from now and pays a semi-annual coupon C. The bond’s
current price Px is observable in the market. Then static spread ss is the single spread over yield curve
that makes its discounted cash flows equal to the market price of a bond:

C/2 C/2 100

Px = + +
1 + y1 + ss (1 + y 2 + ss ) (1 + y 2 + ss ) 2
2

Assuming price Px is observable, coupon C is given, and we discount using the treasury yield curve, we
can solve for static spread ss for the bond using the equation above.
The bigger the static spread, the riskier the bond. This spread is a measure of how much investors need
to be compensated for taking on additional risk by investing in corporates compared to investing in
Treasuires.

Credit Default Swaps (CDS)

Credit Default Swaps were invented relatively recently in the late 90s. Since then the CDS market has
grown very rapidly and today they are the most popular type of credit derivative instruments. In fact
CDS exposures have played a major role in the downfall of many institutions during the recent
economic crisis.

Mechanics of CDS
CDS were originally devised as instruments that provide insurance or protection against a particular
company defaulting. All CDS trade over-the-counter and there are two parties to the contract – a
protection buyer and a protection seller.

Let’s walk through a simple example to understand the mechanics of the contract. Suppose investor A
owns 10,000,000 notional of a bond issued by company XYZ and would like to protect himself from the
possibility of that company defaulting. He decides to take out a 5 year “insurance policy” on the
company by entering into a CDS contract with counterparty B. Let us assume that the market 5 year
CDS premium is 100 basis points per year. The contract they enter into essentially stipulates that:

• For the next 5 years (or until an event of default) investor A agrees to pay investor B
(100bps/4) * 10,000,000 = $25,000 every quarter. This is effectively the insurance premium.
• In the event of a default, investor B makes a payment to covers all losses suffered by investor
A. The CDS contract terminates after the event of default.

CDS premium paid periodically

Protection Buyer Protection Seller
“Short” Payment only if credit event happens “Long”

In the event of default, depending on how the contract is structured, there are two ways investor B can
make investor A whole:

• Physical Settlement: Investor A delivers the defaulted asset to his counterparty B and gets
10,000,000 in return. Note that the defaulted asset will have some recovery value that investor
B will be able to recoup. If the recovery rate is R, Investor B will be able to reclaim (R *
10,000,000), thus the total loss for protection seller in this case is (10,000,000 – R * 10,000,000)
= (1-R) * 10,000,000.

©2009 BlackRock, Inc. All Rights Reserved. 37

Chapter 6: Corporate Credit and Credit Default Swaps BlackRock Solutions

• Cash Settlement: Investor B makes a single cash payment to compensate investor A for the
defaulted value i.e. Investor A receives (1-R) * 10,000,000.

CDS terminology

• CDS Type: The example we looked at was the most common type of CDS know as a “single-
name” CDS. There are a number of other variations of the single-name structure that trade.

o Single-Name CDS: Provides protection against one entity’s default

o Basket CDS (CDX): Provides protection against defaults for any name in a pre-defined
basket of entity names.
o Tranche Baskets: Provides protection against a slice of names defined in the basket. For
example, one can get protection on the riskiest 3% of names in a certain basket (the 0-
3% tranche) or the safest 70% (30-100%) or somewhere in between (10-15%)
o Nth to default: For example, for 2nd-to-default, the payment occurs only when the
second name from a certain basket defaults

• Reference entity: This refers to the entity/company that protection is being bought/sold on
e.g. Verizon, Republic of Italy
• CDS spread: This term refers to the insurance premium (in annualized basis points) paid by the
protection buyer. Note that the term “spread” is a little misleading here as it refers to a
coupon payment and not additional discounting.
• Doc Clause: This field specifies exactly what corporate actions qualify as credit events that
would trigger the protection seller to make a payment.
• Tier: This represents the seniority of the deliverable bond in the event of default (e.g. Senior,
sub etc)
• Reference Obligation: Reference entities issue different types of debt along the capital
structure. The reference Obligation specified in the contract is a cusip that represents the most
junior bond that can be delivered in the event of default.

Apart from hedging corporate bond exposures, the CDS market has given investors a valuable tool to
play around in the credit markets. Going long a CDS provides investors with a way of owning exposure
to an entity without actually owning the security (i.e. without having to fund the position). It also gives
investors a way to isolate and quantify default risk embedded in a corporation. CDS also allow investors
to express views on the credit term structure (e.g. sell 2yr, buy 5yr CDS on the same name) or make
Bond-CDS arbitrage plays as we will discuss later.

Valuation of a CDS
Similar to valuing an interest rate swap, pricing a CDS requires us to project forward cash flows on
each “leg” of the CDS and then discount them back to today. The complication arises from the fact
that we cannot predict how long the CDS contract will last because we do not have a way of telling
when the underlying will default with 100% certainty. Therefore in order to project CDS cash flows, we
need to take into account the probability of default in each cash flow period.

©2009 BlackRock, Inc. All Rights Reserved. 38

Chapter 6: Corporate Credit and Credit Default Swaps BlackRock Solutions

Let’s take the example of a hypothetical one-period CDS on a $100 notional that pays annually. Let us
define

S CDS spread
R Recovery rate
p One period probability of default

So the protection seller is guaranteed to be paid 100*s as long as the company does not default while
the protection buyer will be paid 100*(1-R) only in the event that the company defaults. Given the
probability of default from Time 0 to Time 1, the expected cash flows of the CDS are as follows:

PROTECTION SELLER PROTECTION BUYER

Time 0 0 0
Time 1 100 *s * (1 − p) 100 *(1 − R) *p +0 *(1 − p)

(1-p) represents the one period survival probability. Discounting the cash flows using a yield curve (for
example the LIBOR (swap) curve), the price equation for the CDS, from the perspective of the
protection seller, can be written as follows

100 *s *(1 − p ) 100 * (1 − R)* p

Price (NPV) = −
1+ y 1+ y

Now the obvious question arises, where do we get the probability of default from? This brings us to the
Reduced Form Model.

Reduced Form Model: Calibration of market data to determine default probabilities

On a daily basis we collect CDS spread curves for almost the entire universe of CDS trading in the
market. For a specific entity and a given swap term, these spreads represent the quarterly premium
that one would have to pay to buy protection on the reference entity. In essence these are fair market
spreads that represent the markets view on the credit risk of the entity (based on supply and demand).

For a given name we receive not just a single spread, but a spread corresponding to CDS swaps of
different tenors i.e. we receive an entire spread curve for each issuer. Using these spreads we then use
the Reduced Form Model to estimate a curve of one-period default probabilities. Let us walk through
simple example to demonstrate how this is done. For a particular entity, assuming we know s1 and s2
representing the CDS premium one would pay to enter into a 1-year and 2-year ATM CDS contract
respectively. And we want to determine p1, representing the probability of default from Time 0 to Time
1, and p2 representing the probability of default between Time 1 to Time 2.

We know that these spreads quotes represent fair market spreads, i.e. the NPV of the CDS contracts at
these spreads is zero. Given s1, we can solve for p1 equating the expected discounted cash flows of the
protection buyer and protection seller.

100 *s1 *(1 − p1 ) 100 * (1 − R)* p1

=
1+ y 1+ y

©2009 BlackRock, Inc. All Rights Reserved. 39

Chapter 6: Corporate Credit and Credit Default Swaps BlackRock Solutions

Once we know p1, we can use it to estimate p2 by pricing a two period fair market CDS contract.
Equating the value of the two legs will provide us with the market implied probability of default from
Time 1 to Time 2.

PROTECTION SELLER PROTECTION BUYER

Time 1 100 *s 2 * (1 − p1 ) 100 *(1 − R ) *p1
Time 2 100 *s * (1 − p ) * (1 − p ) 100 *(1 − R ) *(1 − p1 ) * p 2
2 1 2

Once we imply our probabilities of defaults from the market quotes, we can use them to price any CDS
on that reference entity as described before.

Credit Relative Value

Both CDS and bonds are products that reflect the markets view on the default risk of a particular
issuer. The bond market represents this credit risk via its static spread
C + 100
Market Price of Bond =
1 + y +ss
While the CDS market expresses its view by charging different levels of CDS spread/premium. Investors,
trying to take advantage of market dislocations, will want to know if bonds and CDS are trading
relatively cheap/rich versus the other. Now we will briefly introduce two measures that have been
devised to allow them to make such a comparison.

CDS Implied Price

CDS spreads give us an alternative way to take into account that there is a possibility of default. We
can take the probabilities of default p extracted from CDS market and use them to theoretically price
the corporate bond. In order to do this we would adjust the bonds cash flows by the default
probabilities and discount those cash flows using a risk free yield curve. A price obtained by
discounting these default-adjusted cash flows is called the CDS Implied Price. In the case of a one
period bond, the pricing equation would look like this:

(C + 100) *(1 − p1 ) + 100 * R* p1

CDS Implied Price of Bond =
1+ y

If Bond Price < CDS Implied Price, that implies that the bond market is pricing securities credit risk
cheaper than the CDS market.

Default Adjusted Spread (DAS)

An alternative way of looking at relative value between bonds and CDS is via the measure called
default adjusted spread (DAS). DAS is defined as the additional spread or additional discounting that
needs to be added in order to make the CDS Implied Price of a bond equal to its market price.

(C + 100) *(1 − p1 ) +100 * R* p1

Market Price of Bond =
1 + y + DAS

If DAS >0, then bond is cheap relative to its CDS i.e. there is value in the bonds static spread, even
when its default component, as implied by CDS spreads, is stripped away.

If DAS < 0, then bond is rich vs CDS.

©2009 BlackRock, Inc. All Rights Reserved. 40

Chapter 7 – Interest Rate Models
To value fixed-income securities we need a methodology for projecting their future cash flows as well
as a method for discounting them back. Securities without embedded options can usually be valued of
today’s yield curve. For example, we have already discussed in the previous chapter how the cash
flows for vanilla interest rate swaps are projected using the forward curve and discounted using today’s
Libor curve. However when securities have optionality, forward curve valuation is no longer sufficient.

Objectives of this chapter:

• Understand the concept of backward induction and the definition of an interest rate lattice.
• Understand why certain securities need to be valued using backward induction while others
require Monte Carlo simulation.

Interest Rate Models

Investors use a variety of models to analyze options in their portfolios. To see why we need valuation
methods that go beyond using a forward curve let us first recall how callable bonds work. Corporations
often issue debt with call options that give them the right to call back the security if the market price
ever goes above par. As expected an investor would not be willing to pay as much for such a security
compared to one without optionality since the investors upside is limited. But how does one quantify
how much to pay for such a security? Given a set of discount factors we know how to price a fixed or
floating rate bond but how do we capture the optionality embedded in the cash flows?

In the case of equity options, valuation can be done using a closed-form solution such as the Black-
Scholes formula. But valuing options embedded in fixed income securities tends to be more
complicated.

The value of an option has two components: Intrinsic Value and Time Value. Previously we had
discussed how time value is directly related to the volatility in the underlying security value. Therefore
in order to value the embedded option in a callable bond, we need a method to capture the
uncertainty in the bond’s price. In order to achieve this, we need a way of projecting different interest
rate scenarios that can be used to generate different bond prices. An interest rate model is a function
that defines how interest rates move over time and can be used to create an interest rate lattice
which is simply a representation of those various interest rate scenarios.

We can point out three main categories of fixed-income securities that require different techniques for
valuation.

Security Type Valuation Methodology

Securities with no optionality: Forward curve
Securities with path-independent optionality: Backward induction on an interest rate lattice
Securities with path-dependent optionality: Monte-Carlo simulation on an interest rate lattice

Valuing a bond with no optionality

In order to value a security with no optionality, we need to first project out a set of cash flows (which
can either be fixed based on the bonds coupon or floating off the forward curve) and then discount
them back to today using the discount factors implied by today’s yield curve.

Assume we have a three-year fixed coupon bond with an annual coupon of 9%. Since we have our
discount factors, pricing this security is simple. However, we want to point out two different
methodologies for applying the discount factors to the cash flows that will result in the same price.

©2009 BlackRock, Inc. All Rights Reserved. 41

Chapter 7: Interest Rate Models BlackRock Solutions

Discount all cash flows back to time 0

Each cash flow can be treated as a standalone zero-coupon bond and we can compute a price by
applying corresponding discount factors to each cash flow.

PV = DF1 * 9 + DF2 * 9 +DF3 * (9 +100)

100

9 9 9

0 1 2 3

Backward Induction
Another way of computing such a security’s value is to think about how much it would be worth in the
future, and then work our way back to today. How much would we be willing to pay for this security
right before maturity? We know that at maturity T the bond has just one cash flow left (final coupon +
principal) and since there is no discounting, the price of the bond would be worth 109.

How much is it worth at time T-1? In order to answer this question we will need a discount factor
between T-1 and T denoted D1,T-1. The forward curve provides us the forward discount rate that can be
applied to discount cash flows from time T to T-1. Recall the formula that links spot rates to forward
rates from the chapter on Yield Curves:

(1 + S T ) T = (1 + S T -1 ) T −1 (1 + f 1,T −1 ) 1

This can be re-written in discount factor space as:

DFT = DFT −1 *DF1,T −1

100

9 9 9

0 T-2 T-1 T

©2009 BlackRock, Inc. All Rights Reserved. 42

Chapter 7: Interest Rate Models BlackRock Solutions

Once we have the value of the security at time T-1, we can then discount that value back to T-2 (again
using the forward curve). In the same manner we can work our way back through time until we get the
security’s value today. Since we are using the forward curve to discount (which is derived from the
spot curve), the resulting bond value will be identical to discounting each cash flow directly back to
today.

This valuation technique is known as backward induction and will be used again in the next section.

Valuing a bond with path-independent optionality

What if cash flows at a particular point in time are dependent on the prevailing interest rate
environment as is the case for callable bonds? As we mentioned at the start of the chapter, in order to
capture the time value of the embedded option, we need a way to project the underlying security’s
value in different interest rate scenarios.

An interest rate lattice is a representation of those different interest rate scenarios. One of the
simplest forms of a lattice is a recombining binomial tree. At each timestep, we assume that the next
move of the variable can only take on two values, a “high” value or a “low” value. Let us take a look
at a simple example.

Since we have our spot yield curve, at time T=0 we know the discount rate from T=1 to T=0 which is
the one year spot rate or the “short rate”. If the short rate is r= 3.80% at time T=0, at time T=1 we
assume that the one year spot rate can take on either a “high” value of 4.01% or a “low” value of
3.99%. For simplicity, we can further assume that the probability that the short rate moves up (p) and
the probability of a down move (1-p=q) are equal at 50%. From each of those two nodes, it in turn can
move along one of two paths. Below is the graphical representation of such a tree.

T T+1 T+2

4.22 %
N(2,2)

4.01
N(1,1)

3.8 % 4.20 %
N(0,0) N(2,1)

3.99
N(1,0)

4.18 %
N(2,0)

The actual values at each node N(x,y) and the probabilities associated with moving from one node to
another depend on our choice of interest rate model. This tree is said to be recombining since a down
move followed by an up move gets you to the same place as an up move followed by a down move.

©2009 BlackRock, Inc. All Rights Reserved. 43

Chapter 7: Interest Rate Models BlackRock Solutions

So how could we use this interest rate lattice to price a security? Remember, that the lattice was
introduced to help with pricing of callable securities. But let us first start with an example of how we
would price a regular non-callable fixed-coupon bond from our previous example: a three-year fixed
coupon bond with an annual coupon of 9%. Given an interest rate lattice, the next step is to plot out
the corresponding price-lattice representing the sum of the future discounted cash flows and the
coupon received in that period.
109

P(2,2)
4.22

P(1,1) 109
4.01

P(0,0) P(2,1)
3.8% 4.2

P(1,0) 109
3.99

P(2,0)
4.18

109

T T+1 T+2 T+3

In order to fill in the security’s price at each node, we use the same backward induction methodology
introduced earlier. Similar to the case with no optionality, we know that the bond is worth 109 at
maturity. We can now work backwards in time to solve for the price of the bond at node N(2,2)

P(2,2) = Coupon received at time T+2 + Expected discounted value of future cash flows
1
=9+ (0.5 × 109 + 0.5 × 109) = 113.59
1 + 4.22%
Note that the short rate of 4.22 at node N(2,2) represents the discount rate from time T=3 to T=2.
Similarly, we can calculate the bond’s price at the two other nodes at time T=2 to obtain a tree that
looks like this: 109

113.59
4.22

P(1,1) 109
4.01

P(0,0) 113.61
3.8% 4.2

P(1,0) 109
3.99

113.63
4.18
©2009 BlackRock, Inc. All Rights Reserved. 44

109
Chapter 7: Interest Rate Models BlackRock Solutions

Now that we have the prices at time T=2, we can proceed with calculating prices at T=1 by continuing
to walk backwards along the tree:
1
P(1,1) = 9 + (0.5 × 113.59 + 0.5 × 113.61) = 118.22
1 + 4.01%
Similarly, P(1,0) = 118.26 and finally, the price at time T=0 is P(0,0) = 113.91.

Now that we are experts in pricing non-callable bonds using a tree, how would we adjust the above
calculations if we knew that the same bond we’ve been working with has a European call at time T=2
with a strike price of 113.60. That is, the issuer has the right to call back the security at 113.60 in
two years. As soon as we reach the call date, it is advantageous to the issuer to call it back if the price
of the bond at that time is greater or equal to 113.60. Therefore the price would never be allowed to
go above 113.60 at T=2. We can capture the value of the option by capping our price values at 113.6 as
we work backwards through the lattice.
109

113.59
4.22

P(1,1) 109
4.01

P(0,0) 113.60
3.8% 4.2

P(1,0) 109
3.99

113.60
4.18

109

We can now re-calculate our values at nodes T=1 and T=0 using the adjusted values at T=2 to arrive at
a new price P(0,0) = 113.90. Since the investor is short the option, as expected, the price of a callable
bond is lower than the price of a non-callable bond that is otherwise equivalent.

Valuing a bond with path-dependent optionality

The future cash flows of a fixed rate (or floating rate) vanilla bond do not depend on how interest
rates looked like in the past. But there are some securities that cannot be priced using backward
induction on a lattice since the embedded option not only depends on the current level of interest
rates, but also depends on the historical interest rate path we used to get here. An example of such a
security is a Mortgage Backed Security or MBS. We will discuss the properties of such securities in more
detail over the next two chapters.

As you remember, the lattice “doesn’t know” if you reached node N(2,1) by going through node N(1,1)
or N(1,0). The price at node N(2,1) is identical whether you came from node N(1,1) or N(1,0).
However, this assumption is incorrect for path-dependent options such as the prepayment option of a

©2009 BlackRock, Inc. All Rights Reserved. 45

Chapter 7: Interest Rate Models BlackRock Solutions

MBS security. The path dependency for MBS arises because future prepayment speeds depend not only
on the current level of interest rates, but also on historical prepayments that have already taken
place.

In order to value such securities, we use an alternative valuation method know as Monte-Carlo
Simulation. In concept the methodology is quite simple and consists of the following steps

(1) Generate multiple interest rate scenarios by randomly sampling an interest rate lattice.
(2) Generate option adjusted cash flows for each interest rate scenario.
(3) Discount those cash flows back to today to get a security price for each scenario.
(4) The final value of the security is the average price across all paths.

N(2,2) N(2,2)

N(1,1) N(1,1)
N(0,0) N(2,1) N(0,0) N(2,1)

N(1,0) N(1,0)
N(2,0) N(2,0)

T=0 T=1 T=2 T=0 T=1 T=2

Path 1 Path 2

For example, in a two period binomial lattice we could have the two paths shown above. Each path
would result in a different price at node N(2,1) and thus a different price at node N(0,0). The average
price across multiple such paths would provide us with our best estimate of the true value for this
security.

©2009 BlackRock, Inc. All Rights Reserved. 46

Chapter 8 – Introduction to Mortgages
Owing a home is the key to long term wealth – or so people thought until the housing bubble burst in
2008. The journey of a mortgage starts with the dreams of millions to own their own property and ends
up sliced into complicated structured products on the books of companies that do not understand their
risk. In this chapter we walk through the life of a mortgage from inception to securitization and pay
special focus to their prepayment risk profile.

Objectives of this chapter:

• Understand how a typical mortgage works

• Understand how a mortgage gets securitized into a Mortgage Backed Security.

The Mortgage
A mortgage begins with the desire of an individual to own a home. The value of the house they are
interested in is often well beyond what they could afford to pay with their savings alone. And so, the
buyer borrows money and becomes a homeowner. A mortgage is a secured loan in which a borrower
pledges his property as collateral for a loan from a lender. In other words, if the borrower fails to pay
back the mortgage in a timely manner, the bank can sell the property in order to recoup the money
lent. Since the mortgage is a non-recourse loan, in the event that the borrower does not pay, the bank
cannot go after other assets (his car, savings account, etc.) that the individual might have.

The loan contract specifies the loan amount, interest rate, loan term and frequency of payments.
Typical lenders include commercial banks, credit unions, savings and loan banks, community banks,
and mortgage bankers. Mortgages can take a variety of shapes and forms and can typically be
characterized into the following types based on the property that serves as collateral for the loan:

• Residential real estate

o Single Family Mortgages: One to four-family homes including both houses and apartments.
o Multi-Family Mortgages: More than four-family homes, includes condominiums and co-ops
• Non-residential real estate
o Commercial Mortgages: Shopping centers, hotels, office buildings, warehouses etc
o Farm properties

The characteristics of the loan are negotiated between the lender and the borrower and depend on
creditworthiness of the homeowner. A typical homeowner will make a 10% to 20% down payment at
the time of securing the loan. The down payment is the borrower’s equity in the home with the rest of
the home value financed with borrowed money. Before we dive into the life of a mortgage, we briefly
introduce some key characteristics that define a mortgage contract.

Principal or Loan Amount

Due to the initial down payment, the loan amount is usually lower than the value of the property used
as collateral. The ratio of the loan amount to the value of the underlying collateral is called the loan
to value ratio (LTV). A lower ratio is better because the lender has more protection due to the higher
property value securing a smaller loan. In the case of default, the lender has a better chance of
recovering the entire loan amount upon foreclosure and sale.

Loan Term
The loan term defines how long the mortgage will remain outstanding. Common terms for residential
mortgages are 15 and 30 years. All else being equal, for loans with standard flat payment structures,

©2009 BlackRock, Inc. All Rights Reserved. 47

Chapter 8: Introduction to Mortgages BlackRock Solutions

longer loan terms correspond to smaller periodic payments and more total interest paid over the life of
the loan.

Mortgage Rate
Mortgage rates are quoted in annual terms, but the corresponding mortgage payments are usually
monthly. Rates can be fixed or they can float based on an index and a spread. They can also be fixed
for a set period in the beginning of the loan and then start to float. Floating rate mortgages typically
have life caps as well as periodic caps that set a limit on homeowner payments. The mortgage rate
offered to a borrower depends both on the type of mortgage taken out as well as the creditworthiness
of the homeowner. Higher FICO scores, lower LTV ratios and lower debt-to-income ratios all lead to
lower mortgage rates for the homeowner. Since mortgage lenders are in the business of making money,
another key factor that determines mortgage rates is the level of current interest rates in the market.

Payment Schedule
Most mortgages require the home owner to make monthly payments. Each scheduled monthly payment
comprises of an interest payment on the outstanding loan as well as a “scheduled” principal payment
that pays down the balance of the loan. The payment schedule is determined at origination and is
expressed explicitly in the contract. The most common payment schedule is such that the net monthly
payments are flat throughout the life of the mortgage. The flat payment is first used to satisfy the
interest portion of the monthly payment and then any remaining portion is used to pay down the
principal of the loan. During the early life of the mortgage most of the periodic payments comprise
purely of interest payments, while towards the end almost the entire payment amount is used to pay
down principal.

All these characteristics together define the mortgage. A typical single-family mortgage would be a
fixed rate, 30-year mortgage with monthly payments and flat amortization.

The Prepayment Option

While the loan term determines the contractual schedule of payments, the actual payments received
by the lender can vary month to month. Most mortgages give the borrower the right to reduce principal
ahead of schedule, which is called a prepayment option. Any payments made by the borrower in
excess of scheduled payments are called prepayments. There are a number of reasons why a lender
might receive unscheduled payments. For example,

• Cash proceeds realized from default and foreclosure sale of property.

• If the homeowner refinances into a new mortgage with a lower interest rate.
• If the homeowner sells the property to relocate or a move to a new house.
• Partial prepayments (curtailments) or full prepayment made by borrower.

Life of a Mortgage Application

The market place for mortgages has always been quite competitive with a large number of lenders
offering products to attract homeowners. Even though the number of lenders has decreased, shopping
for a home loan is still a lengthy process and requires co-ordination of several parties. The steps that
are usually followed in order to secure an application are described below:

• Initial application and pre-approval – A borrower will initially conduct some research on rates
and mortgage products. After comparing rates from lenders, an application is submitted to
obtain pre-approval and receive a mortgage rate quote. This rate is locked in and the lender
usually allows a 2-3 month period to close the loan.

©2009 BlackRock, Inc. All Rights Reserved. 48

Chapter 8: Introduction to Mortgages BlackRock Solutions

• Home appraisal, credit verification – A complete application with all necessary credit checks
regarding the borrower’s assets, liabilities, employment, income, and other relevant
information to gauge creditworthiness follows. The lender reviews these documents and also
performs checks on the underlying collateral by performing title searches and home appraisals.
• Closing and Funding – Home inspections and legal formalities follow and then a final closing
date needs to be decided between all concerned parties after which the loan is considered
closed and funded.

So what happens next? Of all the types of consumer loans originated each year, mortgages are the
largest and home ownership remains an important political focus. To continue to originate mortgages
at the pace consumers need, additional sources of funding are necessary. At the same time banks
originating the loans do not necessarily want to hold on to the credit risk associated with the
underlying borrower. The capital markets serve as both the funding engine and help with the risk
dissemination needs of the mortgage market.

Mortgage Securitization - Mortgage Backed Securities (MBS)

Imagine you are a large institutional fixed income investor which wants to gain exposure to the large
US mortgage debt market. It is difficult to understand the local and often idiosyncratic nature of
individual mortgages. It would be cumbersome and time consuming to evaluate the credit risk of each
loan and if you have several million dollars to invest, liquidity would also be an issue.

The securitization of mortgage loans into Mortgage-Backed Securities (MBS) solved this problem by
diversifying mortgage credit risk, standardizing cash flows, creating new disclosures, adding
transparency in interest rate risk and pricing as well as improving liquidity. Securitization of mortgages
into MBS revolutionized the mortgage market by allowing banks to get rid of mortgages from their
balance sheets and at the same time feed the capital markets hunger for yield. This also helped the
homeowner as the resulting increase in capital attracted to this portion of the market reduced the cost
of mortgage loans.

The most basic MBS structure is a pass-through. As the name implies, the monthly principal and
interest payments from a pool of mortgage loans are passed through to investors in the secondary
market. A pass-through issuer first acquires mortgages either by originating them or by purchasing
them from a bank. Mortgages with similar characteristics are then collected into a pool and the
ownership interests in the pool are sold as pass-through certificates. The certificate entitles the owner
of the security to a pro rata share of all cash flow payments resulting from the pool of mortgage loans.

The MBS is issued with a face amount equal to the original balance of the mortgages. The MBS assumes
the same characteristics as the collateral that secures the principal and interest payments. For
example:

• Bonds that are based on collateral comprising of fixed rate mortgage loans are called Fixed
Rate MBS.
• Bonds that are based on floating rate mortgage loans are called Adjustable Rate Mortgage
backed securities (ARMs)
• Bonds that are based on collateral with an initial fixed rate period followed by a floating rate
period are called Hybrid ARMs. A 3/1 hybrid ARM is a mortgage that has a fixed rate for three
years after which it resets annually off an index.

The Agencies : Fannie, Freddie and Ginnie

The growth of the housing market in the United States can be attributed to the creation of three
government sponsored entities (GSE) know as Freddie Mac (FHLMC), Fannie Mae (FNMA) and Ginnie
Mae (GNMA). While the specific circumstances of their birth differ, the three agencies were all

©2009 BlackRock, Inc. All Rights Reserved. 49

Chapter 8: Introduction to Mortgages BlackRock Solutions

chartered by Congress to serve one purpose – develop an active secondary market for mortgages to
reduce mortgage rates in the primary market. They achieved this goal by purchasing mortgage loans,
securitizing them into MBS and then reselling them into the secondary market with an insurance wrap.
That is, the principal and interest payments from Agency MBS were guaranteed by the GSEs. Investors
in Agency MBS thus did not have to worry about the credit risk of the loans backing the MBS deal and
instead their only concern would be its prepayment characteristics.

Not all mortgage loans are eligible to be used as collateral for Agency MBS. Freddie Mac and Fannie
Mae only guarantee loans that meet size criteria, satisfy strict underwriting guidelines and cater to
single family homes. Mortgages that meet the guidelines set forth by Fannie Mae and Freddie Mac are
known as conventional conforming loans.

Non-agency MBS are securitized mortgages backed by mortgage loans that are excluded from agency
guarantees and typically fall into three major buckets:

• Jumbo Mortgages: Loans with balances that exceed a limit set annually by the US government.
• Alt-A Mortgages: Loans with insufficient documentation.
• Subprime: Loans where the borrowers have credit history problems.

In order to achieve a higher credit rating and wider appeal, non-agency MBS are structured with special
credit enhancement features. This can be achieved through credit features such as subordination,
over-collateralization and excess spread which will be discussed in more detail in the next chapter.

The process of Securitization

Mortgage-backed securities are created by the process of securitization where illiquid loans are
converted into a tradable liquid security. Securitization of loans can lower risk, add liquidity, and
improve economic efficiency. Securitization involves a “true sale” of the underlying loans from the
balance sheet of the original lender. This is why a separate legal entity called a special purpose
vehicle (SPV) is created for the purpose of securitization. The loans being securitized are first sold to
the SPV which then acts as the issuer of the bonds.

MBS Creation

Mortgage Loans Agencies MBS Investors

Lenders

Cash Movement

Monthly
Payment
Home Owner Servicer Agencies Investors

Servicing Fee Guarantee Fee

©2009 BlackRock, Inc. All Rights Reserved. 50

Chapter 8: Introduction to Mortgages BlackRock Solutions

The process for creating an Agency Mortgage Backed Security is as follows:

1. Agency approved lenders create loans that satisfy agency underwriting guidelines.
2. Lenders submit groups of similar mortgage loans to an agency for securitization. Lenders retain
a servicing fee to cover the cost of servicing the loans. Typical servicing fees are around 50bp.
3. The agency creates an MBS which simply passes on principal and interest cash flows from the
underlying loans to investors less the originators servicing fee and roughly 25 bps as a
guarantee fee. The guarantee fee compensates the agency for bearing all the credit risk of the
underlying loans.
4. The agency sells the MBS in the capital markets to banks, insurance companies, mutual funds or
any other investor wishing to gain exposure to the housing market.

Securitization has played a vital role in the growth of the US mortgage market. The growth of the
market skyrocketed as it satisfied the needs of both mortgage originators and secondary market
investors. It provides secondary market investors with the opportunity to earn higher yields while
taking on limited credit risk. At the same time it allows loan originators to manage their liabilities and
move out excess risk from their portfolios while earning a servicing fee.

The TBA Market

In addition to securitizing mortgages using a consistent set of guidelines, further steps have been taken
to improve the liquidity in trading mortgages. Market participants believe that MBS guaranteed by the
same agency, same coupon and same original term to maturity have the same risk and reward profile.
This assumption helps liquidity by producing a large set of interchangeable securities. For example, it
is common to hear investors to talk about MBS pools based on agency and collateral such as

• FHLMC 30 year 6.0% pools

• FNMA 30 year 6.5% pools
• GNMA 15 year 5.0% pools

This convention allows such pools to trade in the to-be-announced or TBA market. The TBA market is
essentially a forward market. Investors agree to buy or sell mortgage pools that satisfy certain criteria
at some time in the future. Buyers and sellers need to agree on a balance amount, agency type,
coupon, future settlement date and a price before trading. Only as settlement date approaches does
the seller need to specify which pools will be delivered to the buyer.

The TBA market has a couple of major advantages:

• Market participants don’t need to worry about a broad range of other pool characteristics that
many investors consider interchangeable.

• It allows mortgage originators to hedge their interest rate risk. As an originator works on
closing loans and securitizing them into pools through one of the agencies, they can
simultaneously sell them for forward settlement into the TBA market. This protects the
originator from suffering a loss due to declining rates on the originated pool.

MBS Cash Flows Characteristics

MBS investors receive all principal and interest cash flows from the underlying loans minus a servicing
and guarantee fee. A typical cash flow profile of an MBS pass-through is shown below. Initially the
interest component accounts for the majority of the periodic payments and then, as the balance pays
down, principal payments contribute to the bulk of the cash flow going to the investor.

©2009 BlackRock, Inc. All Rights Reserved. 51

Chapter 8: Introduction to Mortgages BlackRock Solutions

Post-processing, investors receive

Homeowner’s Aug payment due pro-rata share of Aug payment

FHLMC
& GNMA FNMA
15th 25th

Aug 1st Sept 1st Oct 1st

30 days in 15 to 25 days
arrears processing

45 to 55 days payment delay

Source: JPMorgan

If you are a homeowner with a mortgage, you will send a check at the end of every month to pay
interest on the outstanding balance of your loan and a scheduled amount of principal that will pay
down the mortgage. The loan servicer then collects these payments and passes them on to the relevant
agency. As an owner of Agency MBS, a secondary market investor will receive his cash flows either on
the 15th (Freddie) or the 25th (Fannie) of the following month. These securities are said to have a 45- or
55-day payment delay because the interest component of the MBS payment covers the cost of principal
outstanding at the beginning of the previous month.

©2009 BlackRock, Inc. All Rights Reserved. 52

 Chapter 8: Introduction to Mortgages BlackRock Solutions

Key Terminology

The characteristics of a mortgage pool are defined by a set of common metrics that give investors insight into
the characteristics of the underlying mortgage loans.

Agency Government agency or sponsored enterprise guaranteeing the MBS

Coupon or Net Coupon Annualized interest rate paid to the investor in the MBS
Term Total years from issuance until the final scheduled payment of the MBS
Original Face (O-Face) Dollar amount of the MBS originally issued
Current Face/Outstanding Dollar amount of the MBS currently outstanding (not yet repaid to the investor)
Balance
Factor Ratio of the Current Face to the Original Face; fraction of the dollars of a pool
that have not yet been repaid to the investor
Payment Delay Number of days between the date when a scheduled payment is owed to the
investor and when it is scheduled to be paid
Amortization Term Number of years over which the underlying mortgages amortize
Balloon Number of years after which the underlying mortgages are required to be paid
off in full
WAC "Weighted Average Coupon" - Weighted average interest rate paid by the
mortgages in the pool
WAM/WARM "Weighted Average Maturity/Weighted Average Remaining Maturity" - Weighted
average time (in months) until the last scheduled payment of the underlying
mortgages
WALA "Weighted Average Loan Age" - Weighted average time (in months) since the
origination of the underlying mortgages
ALS "Average Loan Size" - The average current dollar amount of the underlying
mortgages
AOLS "Average Original Loan Size" - The average original dollar amount of the
underlying mortgages
LTV "Loan-To-Value" - The average ratio of the current dollar value of the loan to
the current value of the home for the underlying mortgages
FICO "Fair and Isaac's Company" - The average credit score (calculated by Fair and
Isaac's) of the holders of the underlying mortgages
%Refi / %Purchase Percent of the dollar amount of the underlying mortgages which are from
refinancing or purchase loans, respectively
%Single / %Multi Percent of the dollar amount of the underlying mortgages which back single
family or multi-family homes, respectively
%Owner / %Second / Percent of the dollar amount of the underlying mortgages which back owner-
%Investor occupied homes, second homes, or investor properties, respectively

©2009 BlackRock, Inc. All Rights Reserved. 53

Chapter 8: Introduction to Mortgages BlackRock Solutions

Prepayment Risk
As discussed earlier, while the loan term determines the contractual schedule of payments, the actual
payments received by the lender can vary month to month. This variability in cash flow is then passed
on to the investor of an MBS. Any payments made by the borrower in excess of scheduled payments,
whether due to exercise of the prepayment option, proceeds from foreclosure or refinancing etc., are
generally called prepayments.

In general, prepayments can be categorized into a number of broad categories that we will look at in
detail.

• Rate Refinancing: Refinancing into a new mortgage with a lower interest rate.
• Cash-Out Refinancing: Refinancing done to take advantage of home value appreciation.
• Housing Turnover: Sale of property because of relocation or a move to a new house.
• Defaults
• Partial Prepayments (curtailments)

Rate Refinancing
This type of refinancing is sensitive to interest rate changes. Imagine you have a $300,000 fixed rate
mortgage with a rate of 7%. Your flat monthly payment is roughly $2,000 a month. If rates fall to 5%,
you can refinance your mortgage to the lower rate reducing your monthly payments to roughly $1,600 a
month. Taking into account the fixed costs associated with a new mortgage, you will have an obvious
incentive to refinance if the savings from lower monthly payments exceed your costs. Refinancing can
be done by taking out a new mortgage at the prevailing market rate and paying back any outstanding
balance left on the previous loan. On the flip side, if mortgage rates increase, there is very little
incentive to refinance and this effect is muted.

Cash-Out Refinancing
This type of refinancing is sensitive to increase in home prices. Imagine your home is worth $400,000
and you purchase it by taking out a $300,000 fixed rate mortgage. Your LTV is 75 [(loan amount) /
(home value)] giving you a 25% equity stake of $100,000 [(home value) – (loan amount)] in the
property. If the value of your home increases to $600,000, your current LTV decreases to 50 and your
equity in the house jumps to $300,000. However, this increase in equity is only an unrealized gain and
you would need to sell your home to take advantage of it.

You can, however, take out a new mortgage for $450,000 with an LTV of 75, pay back the original
$300,000 loan and pocket the $150,000 difference in loan amount. You would still have $150,000 equity
in the home. Even if mortgage rates increase above what you are currently paying for the existing loan,
the incentive to cash-out refinance can be bigger than the desire to keep paying a lower mortgage
rate.

Housing Turnover
Most mortgages require loan repayment when the underlying property is sold. This gives rise to a type
of prepayment called housing turnover. Some factors that result in housing turnover are:

• If home prices increase, you may want to sell your current home which has appreciated in
value and upgrade to a more expensive one. Low mortgage rates can add to the turnover
incentive.
• When the economy does well, your income increases and there are more job opportunities
available. Increased mobility leads to greater housing turnover.

©2009 BlackRock, Inc. All Rights Reserved. 54

Chapter 8: Introduction to Mortgages BlackRock Solutions

• There is a distinct seasonal pattern to home sales. As you would expect, it is higher in late
spring and summer when most people prefer to move as compared to fall and winter.

Defaults
If you fail to make the contractual payments on your mortgage, the lender can take possession of your
property and begin foreclosure proceedings. The mortgage is then said to be in default. The lender
would then sell the seized property and pass on the proceeds to the agency guaranteeing the loan in
case of agency backed loans. Since the agency guarantees the MBS, it takes the loss and passes on a
full principal repayment to the investor. From the point of view of the investor this repayment is the
same as any other form of prepayment.

From an investor’s perspective, managing prepayment risk is the most significant issue when investing
in an MBS security. Prepayments can lead to both extension and curtailment of cash flows beyond what
an investor would expect. Higher than projected prepayments result in earlier cash flows than
expected. This will benefit (hurt) investors who purchased the MBS below par (above par) as their
principal is returned earlier leading to an instant gain (loss). Conversely lower than expected
prepayments result in slower cash flows. This will benefit (hurt) investors of premium (discount) MBS as
they will receive above (below) market coupon payments for a longer period of time.

Generally prepayments are higher in lower interest rate scenarios. This gives rise to the risk of having
to reinvest unexpected cash flows in a lower yielding environment.

Measuring Prepayments
Prepayments are usually the most important factor when valuing MBS. Hence, we need standard
measures to quantify the level of prepayments which would enable us to make more consistent
comparisons among different MBS securities.

Single Monthly Mortality (SMM)

It refers to the percentage of outstanding principal at the beginning of the month that is prepaid during
the month. One needs to subtract the scheduled principal from the balance before calculating SMM.

SMM = [(Prepayment) / (Outstanding Principal)]

Conditional Prepayment Rate (CPR)

CPR is the annualized form of SMM. It can be defined as the cumulative prepayment rate over one year
if we assume a constant SMM every month.

CPR = 100 * [1 – (1 – SMM)12]

Public Securities Association (PSA) Convention

PSA differs from SMM/CPR since it recognizes that prepayment assumptions change over the life of a
mortgage. PSA is a CPR model that ramps up initially to account for seasoning of the loans since the
incentive to prepay is usually quite low immediately after origination. The ramp starts with a CPR of
0.2% in the first month and increases by 0.2% every month thereafter till month 30. By month 30 the
CPR levels out at 6% and is held constant going forward. Other PSAs are simply expressed as a multiple
of this base assumption. For example, 200 PSA will have CPRs which are 200% of the base PSA model
and 50 PSA corresponds to 50% of base mode as shown below.

©2009 BlackRock, Inc. All Rights Reserved. 55

Chapter 8: Introduction to Mortgages BlackRock Solutions

14.00%

12.00%

10.00%

CPR (%) 8.00%

6.00%

4.00%

2.00%

0.00%
0 50 100 150 200 250

Time (months)

100% PSA 50% PSA 200% PSA

As expected, taking prepayments into account significantly changes the cash flow profile of MBS
securities and impacts the duration of portfolios holding these securities. At a 100 PSA the cash flow
profile of a standard MBS security has the typical shape shown below.

Unlike a regular bond that has positive duration and positive convexity, the prepayment option
embedded in a Mortgage Backed Security gives it negative convexity. In order to analyze the risk
profile of an MBS, we need to understand the interaction between the discounting effect (that gives
regular bonds positive convexity) and the prepayment effect of owning a mortgage. Remember that as
interest rates in the market increase, the prepayment option that the homeowner is long (and the MBS
investor is short) is less valuable since he cannot refinance into a cheaper mortgage. Thus, as interest
rates increase, prepayments will decrease and vice versa. Investors who hold MBS that price above par
are essentially short an option that is in-the-money. These investors will be hurt by prepayments as
they lose premium when money gets paid back to them at par. Additionally, prepayments tend to
happen when interest rates fall thus giving investors their money back at precisely the worst time since
they will be forced to reinvest at lower rates. Hence falling interest rates have two offsetting effects
on a mortgage. On one hand, the rally leads to the value of an MBS increasing due to the discounting
effect, but, on the other hand, it decreases due to the prepayment effect. The dominating prepayment
effect is what gives an MBS its negative convexity.

©2009 BlackRock, Inc. All Rights Reserved. 56

Chapter 9 – Structured Products
While the Mortgage Backed Securities we covered in the previous chapter have become one of the most
widely held fixed income products in the world, their risk profile does not always meet the needs of all
investors looking for exposure into the housing market. A particular investor may not like the
prepayment risk profile of a MBS while another might be willing to take on more risk in return for a
higher yield. For some the duration of a pass-through might be too short, for others - too long. This
need for customization gave rise to the structured products market.

Objectives of this chapter:

• Understand the cash flow mechanics of Collateralized Mortgage Obligations and Asset Backed
Securities.
• Gain a broad understanding of the risk profile of each tranche for the various products
discussed.

Structured Products
A structured product is an instrument created by redistributing the principal and interest cash flows of
an underlying pool of assets, known as the collateral, to investors based on a pre-defined set of rules.
The underlying collateral can be used to create multiple securities or tranches such that the total cash
flow going to all the tranches matches the cash flow coming from the collateral. It is important to
remember that while each tranche may have a duration and convexity risk profile that is very different
from the parent collateral, the overall risk of the collateral must match the overall risk of all the
associated tranches. The structured product market takes the risk of the underlying collateral, breaks
it into different components and distributes each piece to investors most inclined to take them on.

It is not surprising that structured products can become quite complicated to analyze. The rapid
market growth of these products, and the losses that investors owning them have experienced, have
contributed significantly to the economic crisis that started in 2008. High market demand for these
products can be attributed to many different reasons. Firstly, on the investor demand side, these
products appeal to a much broader class of investors than the underlying collateral would be able to
attract on its own since they allow investors to achieve specific risk objectives more easily. From the
originators’ perspective, they have a huge incentive to create these structures as investor demand
allows them to sell the individual tranches for more than the collateral is worth. The supply of
collateral required to create these products come from banks that are looking for ways to move these
assets off their balance sheet.

In this chapter we are going to focus on two broad categories of structured products, Collateralized
Mortgage Obligations (CMO) and Asset Backed Securities (ABS). Within each of these broad
categories, cash flows can be redistributed to tranches based on rules defined along one of these three
avenues:

• Timing of cash flow

• Type of cash flow (interest or principal)
• Losses of expected cash flows

Before we continue let us quickly recap the typical cash flow profile of a pass-through mortgage. Recall
that, in a typical agency MBS, interest payments make up most of the cash flow at the beginning and
then, as the balance pays down, scheduled principal and prepayments contribute to the bulk of the
cash flow going to the MBS investor. The cash flow graph below represents a typical 30 year 5.5 coupon
FNMA mortgage run at 100 PSA.

©2009 BlackRock, Inc. All Rights Reserved. 57

Chapter 9: Structured Products BlackRock Solutions

Let us now take a deep dive into understanding the mechanics and risk profile of a number of
commonly seen structured products in the market.

Agency Collateralized Mortgage Obligation

Collateralized Mortgage Obligations (CMO) can be divided into two broad categories based on the
underlying collateral backing the tranches: Agency CMOs and Whole Loan CMOs. Asset Backed Securities
and Whole Loan CMOs are similar in structure and will be covered together in the next section.

As the name implies, Agency CMOs are collateralized mortgage obligations that are backed by agency
collateral. The underlying securities are MBS pools issued by the agencies and thus the principal and
interest are guaranteed by Freddie Mac, Fannie Mae or Ginnie Mae. Since the collateral has no credit
risk, the only two dimensions along which we can split its cash flows to form tranches are either by
timing of cashflow or type of cash flow. In this section we are going to go through some of the most
common CMO structures available in the market.
Agency CMO tranche
Timing of Cash Flow: Sequential, PAC/Support, Z-bond
Type of Cash Flow: Interest Only/Principal Only, Floater/Inverse

Sequential
Sequentials are one of the simplest CMO structures created that divide the collateral’s cash flow into
multiple classes with different durations and weighted average lives. The structure consists of three or
more tranches and distributes principal from the collateral sequentially. Both scheduled and
unscheduled principal from the collateral flows to the first tranche until it gets paid down completely.
Then principal starts paying down the second tranche and similarly we move through the entire
structure. Each tranche receives interest payments in accordance with the risk profile of the tranche.

Let us analyze this structure by looking at a simple example. We assume that the agency collateral
underlying the deal is $10,000,000 worth of Fannie Mae 5.5 30-year pass-through securities. We use this
collateral to create four sequential tranches defined by different notional amounts.

Tranche Notional Amount

A 3,000,000
B 1,500,000
C 3,000,000
D 2,500,000

©2009 BlackRock, Inc. All Rights Reserved. 58

Chapter 9: Structured Products BlackRock Solutions

In this hypothetical structure, tranche A gets the first 3M of principal payments, irrespective of
whether they come from prepayments or scheduled principal payments. Tranche B then gets the next
1.5M in principal payments. After that, once tranche B has paid down completely, tranche C gets the
next 3M in principal. Finally, tranche D receives the remaining 2.5M in principal once all the other
tranches have paid down.

Since each tranche is paid off sequentially, all but the first tranche have a principal lock-out period.
Thus all tranches have different weighted average life profiles with tranche A having the shortest
duration and tranche D having the longest. However, it is important to remember that these tranches
still have prepayment risk. If prepayments increase, the average life of all tranches will shorten as
each class gets wiped out faster.

Investors use sequentials to match the average life or duration of their MBS investment to their
liabilities. Banks and money managers looking for short or intermediate exposure to the yield curve
tend to purchase the first few tranches. Insurance companies, which usually have longer liabilities,
tend to invest in the last class of the structure in order to get a better duration match. Depending on
their specific needs, the principal lockout feature allows investors to take on greater or less
prepayment risk than the underlying MBS.

Planned Amortization Class (PAC)

While sequentials do a good job in providing investors with different weighted average life profiles,
they do not adequately allow for distribution of prepayment risk from the underlying MBS. Planned
Amortization Class (PAC) securities not only provide investors with a range of average lives but, in
exchange for lower yields, also keep those average lives stable over a range of prepayment rates. Since
the collateral cash flows are being redistributed to form the CMO, i.e. the total collateral cash flows
for the deal are still the same, the stable PAC tranche must be offset by a much more unstable
tranche. A support or companion tranche is created to absorb the undesirable volatility in the
collateral cash flows.

A PAC is created by first defining two amortization schedules that will make up the PAC Band. Let us
walk through a simple example with the same underlying collateral we used before: $10,000,000 worth
of Fannie Mae 5.5 30-year pass-through securities. Let us take this collateral and create a PAC with a
PAC Band between 100 PSA and 300 PSA. The amortization schedule of the underlying collateral
corresponding to the two PSA levels are represented in the graph below.

©2009 BlackRock, Inc. All Rights Reserved. 59

Chapter 9: Structured Products BlackRock Solutions

180,000.00

160,000.00

Scheduled Principal and prepayments

300 PSA 100 PSA
140,000.00

120,000.00

100,000.00

80,000.00

60,000.00

40,000.00

20,000.00

0.00
Apr-09

Apr-11

Apr-13

Apr-15

Apr-17

Apr-19

Apr-21

Apr-23

Apr-25

Apr-27

Apr-29

Apr-31

Apr-33

Apr-35

Apr-37
We can use these two amortization schedules to create a tranche with a stable average life as long as
prepayments stay within the PAC band of 100 to 300 PSA. The rules that define the PAC are as follows:

• The PAC schedule is based on the minimum of the two amortization schedules.
• Any principal received first goes to paying the PAC bond according to its schedule.
• In a particular month, principal in excess of what is needed to pay the PAC tranche is used to
pay down the companion or support tranche.
• If principal received in a particular month falls short of meeting the PAC schedule, then excess
prepayments from future months will flow to the PAC until it gets back on schedule. This is
known as the catch-up feature of the PAC.
• These features assure stability in the PAC’s cash flows. If prepayments fall near the lower
band, the collateral will have enough principal to meet PAC requirements in the early years. If
prepayments fall at the upper band, the collateral will still have enough prepayments to meet
the PAC schedule in later years.

©2009 BlackRock, Inc. All Rights Reserved. 60

Chapter 9: Structured Products BlackRock Solutions

The companion/support tranche absorbs a large part of the prepayment risk. If prepayments come in
faster than the upper bound of the PAC band, the support tranche gets eroded quickly and thus the
structure loses its ability to maintain the PAC tranche. If prepayments come in below the lower bound
of the PAC band, the support tranche increases in size.

The PAC and support tranches can be further broken up into tranches that pay sequentially thus
allowing investors to meet very specific duration targets. PACs attract many of the same investors as
sequentials but work better for portfolios that require a more predictable stream of payments. Their
stability allows investment managers to use them as substitutes for treasury and agency debt and
allows them to target specific points on the yield curve.

Due to the stability of the PAC structure, investors receive a lower yield (coupon) compared to owning
the pass-through. Some investors though would be willing to give up some of that stability in exchange
for a higher coupon payment. A Targeted Amortization Class or TAC is similar in structure to a PAC
but only provides protection against faster prepayments i.e. there is no catch-up feature in the
structure.

Z-Bond
A Z-Bond caters to investors, such as insurance companies, looking to take on much longer durations
than those provided within a sequential or PAC structure. A Z-bond refers to a specific tranche within
a CMO structure with two distinct payment phases:

• Accrual Period: During this phase the Z-tranche does not receive any interest payments.
Instead the interest payments get redirected to pay off the other tranches in the structure. In
each period the outstanding balance of the tranche is increased by the foregone interest
payment amount. Thus the Z-bond’s principal balance grows over time at the coupon rate of
the bond. This phase continues until all other tranches have paid off.
• Payment Period: Once all the other tranches have paid off, the Z-Tranche begins to receive
both principal and interest payments until it has paid down.

Restructuring the last sequential tranche into a Z-Bond shortens the average life of all other tranches
in the structure. This happens because the interest given up by the Z-Bond results in the other tranches
was being paid down faster.

Interest-only (IO) and Principal-only (PO)

One of the oldest and most common ways of splitting up a collateral’s cash flow is by breaking out its
interest and principal components. While some other types of CMO structures can also have interest-
only components, a Trust IO/PO refers to a structure consisting of only two tranches derived from the
cash flows of a pass-through: an Interest only (IO) piece and a Principal only (PO) piece.

In order to analyze the risk profile of an IO and a PO, we need to first remind ourselves of the impact
of interest rates on prepayment speeds. Remember that as interest rates in the market increase, the
prepayment option that the home owner is long is less valuable since he cannot refinance into a
cheaper mortgage. Thus as interest rates increase prepayments will decrease and vice versa. A PO
prices below par, hence it benefits from prepayments since you are getting money back at 100. At the
same time an IO gets hurt by prepayments since they not only reduce the outstanding balance of the
loan on which interest is computed but they also reduce the average life of the collateral thus reducing
the amount of time over which one can receive interest payments.

A PO tends to have positive interest rate duration and positive convexity. As interest rates increase the
price of a PO drops both due to an increased discounting effect and due to reduced prepayments. A
reduction in interest rates has the exact opposite effect giving the PO positive duration. Thus POs are
used to add duration or convexity to portfolios.

©2009 BlackRock, Inc. All Rights Reserved. 61

Chapter 9: Structured Products BlackRock Solutions

An IO by contrast has negative duration which can be explained by the interaction of the prepayment
effect and the discounting effect of shocking interest rates. As interest rate increase, the price of an
IO drops due to the discounting effect but increases due to reduced prepayments. Similarly as interest
rate go lower, the price of an IO increases due to the discounting effect but reduces due to the
increase in prepayments. The prepayment effect on an IO has more of an impact on its valuation than
the discounting effect, leaving the IO with negative duration. IOs are used to shorten portfolio
durations or to add to its yield.

Floater and Inverse Floaters

Even if all of the underlying collateral of a CMO structure consists purely of fixed coupon securities,
that CMO can still have a tranche that floats off an index – a Floater tranche. For example, it is
common to have tranches where the coupon is based off Libor + a spread. Remember that the
collateral is fixed-rate, and the total cash flows of the structure must match the collateral’s principal
and interest. This means that having a floating rate tranche in a structure requires us to also have an
offsetting Inverse-Floater tranche. A typical inverse floater will have a coupon payment defined as
(Fixed Coupon - Floating Rate).

Asset Backed Securities (ABS) and Whole Loan CMOs

Unlike Agency CMOs, where the redistribution of prepayment risk drives the CMO structure, Asset
Backed Securities and Whole loan CMOs are securitized products where credit or default risk is
allocated across different tranches.

Whole Loan CMOs

In addition to agency securitized mortgages, banks also have a large number of non-agency Jumbo or
Alt-A mortgages that they need to get off their balance sheets. In order to compete with agency CMO
market, originators need to create CMO structures that have a similar credit profile to the agency
market. They do this by creating dedicated tranches for absorbing losses which provide credit
protection to the more senior tranches. The key features of whole loan CMOs are:

• The CMO structure is based on redistribution of collateral losses.

• There are Senior, Mezzanine and Subordinate tranches. Subordinate (or Equity) tranches take
first losses to protect more senior classes.
• Prepayments flow through to the senior tranches first while losses are taken by the Subordinate
tranches first.

Asset Backed Securities

Asset Backed Securities are similar to Whole Loan CMOs in structure; however, the underlying collateral
used to create ABS structures can take on a variety of forms. The most common types of ABS seen in
the market are backed by:

• Home Equity: These are loans taken out by home owners using the equity they have built up in
their home as collateral.
• Sub-Prime Mortgages: Low quality residential mortgages given to borrowers with bad credit.
• Manufactured Housing: Very low quality residential loans that finance mobile housing projects.
• Auto Loans: Loans to buy automobiles.
• Credit Cards: Represent the cash flows from credit card receivables. The receivables are
backed by a revolving pool of credit card debt that is managed by a Master Trust.

©2009 BlackRock, Inc. All Rights Reserved. 62

Chapter 9: Structured Products BlackRock Solutions

Asset Backed Securities have a number of credit features that are used to attract investors into buying
senior tranches. Investors in senior ABS tranches receive higher yields than they could get in the agency
CMO market and are protected, to a limited degree, from defaults.

Credit Features of an ABS structure:

• Subordination: Most ABS structures have equity tranches that absorb first losses and protect
the senior tranches.
• Over-collateralization (OC): This feature refers to having more notional in the underlying
collateral than in the ABS tranches. The extra collateral acts as buffer to absorb losses.
• Excess Spread: Refers to the excess coupon from the underlying assets over what is needed to
pay the ABS structure. Excess spread cash flows are held in reserve as extra insurance against
losses.
• Triggers: These are provisions such that if certain credit events, defined by delinquency or
defaults, take place on the collateral, all cash flow payments will be redirected to senior
tranches to maintain their level of credit protection.
• Available Funds Cap: Limits the floating interest payment due to a tranche if there is a lack of
funds from the collateral.

The bank structuring the ABS deal usually sells off the senior tranches to investors and ends up holding
on to the junior tranches. Excess Spread and OC cash flows are used to create tranches called NIM
bonds which are sold to investors with high risk tolerances in return for extremely high yields. Since the
main risk being redistributed in ABS structures is credit risk, prepayments are good for the investors.
Most deals have clean up calls embedded in them that allow the issuer to call back the deal and pay all
outstanding principal if the value of the structure exceeds par.

©2009 BlackRock, Inc. All Rights Reserved. 63

Chapter 10 – Value at Risk
Portfolio managers often look for metrics that will provide them with insight into the worst case losses
their portfolio might experience over time. One such measure is Value at Risk (VaR). Risk metrics like
duration tell an investor how sensitive their portfolio is to changes in specific market factors like interest
rates. However, given that market factors (interest rates, volatilities, spreads etc) are all correlated,
how does a portfolio manager get a sense of his PnL if all these factors move simultaneously? Given
historical market moves, what is the worst case loss a portfolio can be expected to experience?

Objectives of the chapter:

• Understand the definitions, assumptions and limitations of a Value at Risk (VaR) metric.
• Understand how to compute analytical and historical VaR for Fixed income securities
• Understand the working of a simple equity risk model.

Fixed Income Value at Risk

Value at Risk (VaR) gives an upper bound on the loss that one would expect a portfolio to experience
over a given time horizon within a specified confidence interval. The confidence interval, usually
represented as a percentage, is the probability associated with the expected loss boundary over a
specified time horizon.

While this definition may sound complicated, the concept is really quite simple. Before diving into the
details, let us begin with a simple example. Taking the example of a common credit index, let us assume
the following:

Portfolio Barclays Capital Credit Index

Market Value $2,000,000,000
Confidence Level 95%
Annual 95% VaR 10%
Annual 95% VaR in dollars $200,000,000

So the one year 95% VaR of Barclays Capital Credit Index is 10%. There are two ways we can interpret this
number:

• Over the next year, there is a 95% chance that this portfolio will lose less than 10% of its value.
• Or, over the next year, there is a 5% chance that this portfolio will lose more than 200 million
dollars.

We can also think about this graphically. Assume we know the probability distribution for the change in
market value of the portfolio over the next year. The x-axis on the graph below represents the change in
portfolio value while the y-axis represents the probability of that change happening over the next year. If
the shaded area is 5% of the total graph area and the value corresponding to the 5th percentile is 200M,
then we can say the annual 95% VaR of this portfolio is 200M. Given this probability distribution, there is
5% possibility that this index portfolio will lose more than 200M over the next year.

©2009 BlackRock, Inc. All Rights Reserved. 64

Chapter 10: Value at Risk BlackRock Solutions

Probability (pdf)

5%

-200M 0
Change in Market Value

Consider another case. Suppose the distribution of the change in portfolio value over a month is as shown
below. The shaded area is 1% and the value corresponding to the 1st percentile is 150M. This indicates
that there is a 1% probability that this portfolio will lose more than 150M over the next month. In this
case we would say that the 99% monthly VaR of this portfolio is 150M.

pdf(ΔV)

1%

-150M

From these two examples, we can see that there are three important elements to a Value at Risk
number: a confidence level (typically either 95% or 99%), a time period (a month or a year) and an
estimate of investment loss (expressed either in dollar or percentage terms).

There are two different types of VaR: Analytic VaR and Historical VaR.

Similar to our examples above, both methods come up with probability distributions for the change in the
portfolio’s market value in order to determine its VaR. However each method uses a different technique
to come up with the distribution and thus have different limitations. In both cases, the change in market
value used to come up with the distribution is computed parametrically.

©2009 BlackRock, Inc. All Rights Reserved. 65

Chapter 10: Value at Risk BlackRock Solutions

Parametric Model
A portfolio has exposure to various risk factors such as interest rates, volatility, spreads, mortgage basis
etc. Metrics such as duration, convexity, spread duration etc represent risk exposures that capture the
sensitivity of a portfolio’s value to changes in the corresponding risk factor.

For a given change in the level of the risk factors, using the corresponding risk exposures we can
approximate the change in a portfolios value ( V ) parametrically using Taylor expansion.

∂V ∂V ∂V
Δ Portfolio Value (V ) ≈ Δrates + Δspread + ΔVol + K
∂rates ∂spread ∂Vol

Analytic VaR
Analytic VaR is computed based on the assumption that all risk factors are normally distributed.
Therefore, combinations of those risk factors, including the portfolio parametric return, will also be
normally distributed. The standard deviation ( σ ) of the normal distribution is estimated from historical
data.

pdf (Δ V )

Normal Distribution of Portfolio Return

ΔV ≈ N ( μ = 0, σ 2 )

-4 -3 -2 -1 0 1 2 3 4

Let us take the example of a simplistic portfolio that has just one risk exposure: duration. In order to
compute analytic VaR on this portfolio:

• Assume normal distribution of risk factors. In our example we need to assume that the changes in
interest rates are normally distributed with mean 0 and a standard deviation derived from
historical rate changes.
• Get the risk exposures of the portfolio. In our example duration is the only relevant exposure.
• Get the distribution of portfolio returns i.e. we need to get the probability distribution for the
change in portfolio market value over our horizon. In order to do this we multiply the risk
exposure (duration) with the distribution of the risk factor (change in interest rates).

©2009 BlackRock, Inc. All Rights Reserved. 66

Chapter 10: Value at Risk BlackRock Solutions

• The normal distribution assumption gives us an easy way to get the distribution for the change in
portfolio value from the distribution of the risk factors. Once we obtain the distribution of
portfolio returns, we can compute an Analytical VaR number easily as described before.

Risk exposure * Normal distribution of Δ risk factor = Normal distribution of Δ Portfolio Value

Duration * Change in interest rates = Change in Market Value

Model Limitations
While interpreting this VaR metric a Portfolio Manager must understand the two main shortcomings of this
approach. Analytic VaR provides good estimates of first order risk within a portfolio but does not capture
non-linear exposures such as convexity (if we assume that Δ rates is normally distributed, then
( Δ rates)2 cannot be normally distributed). Secondly, the distributions of most market risk factors are
not normally distributed and often have fat tails which signify that the probability of extreme events
occurring is greater than that predicted by the normal distribution.

Historical VaR
Historical VaR does not make any assumptions about the distribution of the risk factors. The basic
premise of a Historical VaR model is to estimate a time series of parametric returns using historically
observed risk factor changes. The historical returns are sorted and, given a desired confidence level,
provide a Historical VaR measure that does a better job capturing the fat tails of a return distribution.

Let us take a look at a simple example of a portfolio that has exposure to interest rates, spreads and
volatility. We first pick 500 daily historical observations from the market; giving us 500 sets of interest
rate changes, spread changes and volatility changes.

⎛ Δr1 ⎞ ⎛ Δspread1 ⎞ ⎛ ΔVol1 ⎞

⎜ ⎟⎜ ⎟⎜ ⎟
⎜ Δr2 ⎟ ⎜ Δspread 2 ⎟ ⎜ ΔVol 2 ⎟
⎜ M ⎟, ⎜ M ⎟, ⎜ M ⎟
⎜ ⎟⎜ ⎟⎜ ⎟
⎜ Δr ⎟ ⎜ Δspread ⎟ ⎜ ΔVol ⎟
⎝ 500 ⎠ ⎝ 500 ⎠ ⎝ 500 ⎠

Because we don’t make any assumption about the distribution of risk factors, we can include second
order terms from the Taylor expansion when we calculate the change in value of the portfolio ΔV .

∂V ∂ 2V
ΔV ≈ Δrates + (Δrates )2 + ∂V Δspread + ∂V ΔVol + K
∂rates ∂rates 2
∂spread ∂Vol
Applying the risk exposures to the historical change in risk factors result in 500 historical portfolio return
observations.
⎛ ΔV1 ⎞
⎜ ⎟
⎜ ΔV2 ⎟
⎜ M ⎟
⎜ ⎟
⎜ ΔV ⎟
⎝ 500 ⎠

©2009 BlackRock, Inc. All Rights Reserved. 67

Chapter 10: Value at Risk BlackRock Solutions

These values can be plotted in a histogram to get the distribution of returns in a portfolio. We can then
pick the observation corresponding to our desired confidence level to report a Historical VaR. For
example, if we had 10 daily historical observations, then the 2nd lowest parametric portfolio return would
be our 20% daily historical VaR number.

Model Limitations
Using the historically-observed time series of risk factor changes allows us to capture any non-normality
associated with the return distribution and thus more accurately captures the true probability of extreme
events. One drawback of Historical VaR is that each historical return is equally weighted, so Historical
VaR does not react to recent market changes in volatility as quickly as Analytic VaR.

Equity Risk Model

The examples we have used so far have described how to compute VaR for fixed income securities with
exposure to interest rates, volatilities, spreads etc. While the definition of VaR for equities doesn’t
change, the process of estimating risk exposures and risk factors are different. There are several
different types of equity risk models in the market. Here we focus on the very widely used multi-
factor model.

As before the model is best understood by walking through a simple example. Let us assume that the
stock universe consists of three stocks.

Stock Market Capitalization Industry Daily Return 03/02/2009

A $1,000,000 Finance -4%
B $2,000,000 Finance -2%
C $3,000,000 Auto -8%

In order to compute VaR we first need to define the set of risk factors that can influence the return of
a stock. Unlike the fixed income market where the factors are related to economic variables like
interest rates, stocks returns are influenced by stock characteristics such as the size of the holding
company, the industry the company is in, the volatility of the stock, and the country the company is in.

©2009 BlackRock, Inc. All Rights Reserved. 68

Chapter 10: Value at Risk BlackRock Solutions

Factor Esitmation
In the fixed income market, both the security’s exposure (like duration) and the time series of the
corresponding risk factors (change in interest rates) are known and easily observable in the market. In
equity space, both the exposures and the factors need to be estimated from the universe of stocks.
Given the stock universe defined above, let us walk through a simple example of how we would
estimate a time series of risk factors and exposures. For simplicity, let us assume that there are only
two sources of the returns: size and industry.

For a given date and given company, the daily return of owning the stock can be represented as a
function of the size and industry of the company. Similar to the parametric model equation, return is
expressed as risk exposures multiplied by risk factors. The risk factor returns in our example are size
(fs), Finance industry (ff) and Auto industry (fa).

n
rstock = ∑ exp× f = exp s ⋅ f s + exp f ⋅ f f + exp a ⋅ f a
i =1

The next step is to define risk exposures to each risk factor. For the industry risk factor, the
corresponding risk exposure can either be 1 or 0. If the stock is in finance industry, the risk exposure of
the stock to finance industry factor is 1. Otherwise it is zero. The risk exposure corresponding to the
size factor is a little more complicated to estimate. Let us define the risk exposure to the size factor as
the normalized market capitalization of the stock. For our estimation universe of stocks, the mean
market capitalization is $2 Million and the standard deviation is $1 Million. On 3/2/09, the normalize
market cap for each stock can be computed as follows:

Market Cap − Mean

stdev
Therefore, the risk exposures of these three stocks to size factor are

Stock Risk Exposure to Size Factor

A ($1,000,000-$2,000,000)/ $1,000,000 = -1
B ($2,000,000-$2,000,000)/ $1,000,000 = 0
C ($3,000,000-$2,000,000)/ $1,000,000 = +1

Now that we have our exposures, we can calculate the corresponding risk factor returns on 3/2/09
using the stock returns from our estimation universe on 3/2/09. In this example, we have

rA = −1 ⋅ f s + 1 ⋅ f f + 0 ⋅ f a = −4%
rB = 0 ⋅ f s + 1 ⋅ f f + 0 ⋅ f a = −2%
rC = 1 ⋅ f s + 0 ⋅ f f + 1 ⋅ f a = −8%

Solving the above equations, the risk factor returns on 3/2/09 are solved to be

f s = 2%, f f = −2%, f a = −10%

This computation can then be repeated for historical dates in order to get a time series of risk factor
return values.

©2009 BlackRock, Inc. All Rights Reserved. 69

Chapter 10: Value at Risk BlackRock Solutions

VaR Computation
Similar to an analytic VaR computation for fixed income securities, a normal distribution of stock
returns can be estimated from today’s stock exposure (to size and industry factors) and the distribution
of factor returns. The risk factors are assumed to be normally distributed with mean zero and a
standard deviation ( σ ) estimated from the historical factor returns we computed before. The VaR of a
portfolio of stocks can be estimated from the portfolio return distribution which would be the sum of the
individual stock return distributions.

The example we have walked through highlights just one of many methods that can be used to
compute VaR for equities. As you would have noticed in our example, the computation of risk
exposures can vary depending on how we chose to define them. Changing the definition of risk
exposures will lead to different values for the corresponding risk factors. Therefore it is important to
understand all the underlying model assumptions when comparing VaR numbers from different models.

In our example, in order to estimate the factor returns for 3/2/09, we had three equations and three
unknowns. In reality we have a much larger number of stocks in the estimation universe and have more
return equations than unknown factors. Therefore we need to run a multi-factor regression to find the
optimal solution for factor returns. Each stock also has idiosyncratic return which cannot be explained
by the systematic risk factors such as size and industry.
n
r = ∑ exp× f + ridio
i =1

A good equity risk model usually uses a well-constructed equity index with thousands of stocks as its
estimation universe and provides portfolio managers with critical information for understanding and
quantifying the risk of their investment decisions.

©2009 BlackRock, Inc. All Rights Reserved. 70