|
Fixed income instruments are often described as trading at
a spread over some benchmark
yield. For example, a 10-year
callable
corporate bond might have a
yield to maturity (YTM) of
6.7%. If the on-the-run 10-year
Treasury note's YTM is 5.5%, the
bond would be described as trading at a spread of 1.2%, or 120
basis points, over the Treasury. Such
spreads can be attributed to a number of factors, including
credit quality,
liquidity and embedded options.
If a bond has embedded options, its Option-adjusted spread
(OAS) is the spread at which it presumably would be trading over a
benchmark if it had no embedded optionality. More precisely, it is the
instrument's current spread over the benchmark minus that component of the
spread that is attributable to the cost of the embedded options:
 |
[1] |
or, rearranging:
 |
[2]
[3] |
OAS can be calculated with respect to various benchmarks:
Treasuries, swap rates, a short-term
"risk-free" rate, etc. Most often, the benchmark is Treasuries. To avoid
dependency on a particular benchmark, option-adjusted yield
may be quoted instead of OAS:
 |
[4] |
Prior to the 1970s, investors made only rudimentary
efforts to adjust their analysis
of fixed income instruments to recognize the effects of embedded options.
There were several reasons for this. Back then, the bond market was less
diverse than it is today. Instruments like
mortgage-backed securities (MBS)
didn't exist. I oversimplify only slightly if I describe the US market as
offering two types of bonds: callable
corporates and non-callable
Treasuries. Call features differed little from one bond to the next, so
investors could reasonably compare corporates based on their
yield to first call or
yield to worst. Corporates
offered yields in excess of Treasuries, and some of the excess yield could
presumably be attributed to embedded call features, but there was no
particular need to put a number on this. No one was shorting corporates
against Treasuries as a volatility play! Another issue was the fact that
analytics for assessing option values didn't exist. The
Black-Scholes model for pricing
options had not yet been published. Computer technology was cumbersome and
expensive. Finally, interest rates tended to be stable prior to 1970, so
embedded call options weren't worth much to begin with.
All this started to change in the 1970s. New forms of
fixed income instruments were brought to market. Interest rates became
increasingly volatile. A robust theory of
option pricing emerged, and the
processing power needed to implement the new theory became easier to use
and less costly.
Option-adjusted spreads were first widely employed in the mortgage-backed
securities market in the late 1980s. Investors were offered instruments
with extraordinary current yields—500 or 600 basis points over Treasuries. To analyze these,
they needed to
somehow subtract out the yield component that was attributable to the embedded
options. They wanted to know what the yield over Treasuries would be if
the exact same instruments did not have embedded options.
The value of option-adjusted spread analysis is that it enables
investors to separate out optionality and judge the degree to which an
instrument's yield compensates them for
credit risk,
liquidity risk or other such factors.
Suppose an investor is comparing two similar bonds. Both have comparable maturities, credit
qualities and liquidity, but they have different embedded options. The investor might
purchase whichever bond has the higher option-adjusted spread—that bond
would offer higher compensation for the risks being taken.
This is how OAS is used in theory. Practice is not so
simple. OAS is more a philosophy that can be implemented in different ways
than it is a well defined metric of yield. Models abound. Proprietary
models used by bond dealers tend to be sophisticated. Those that are
available to investors can be crude. Routinely, an investor will survey a
number of dealers on the option-adjusted spread those dealers calculate
for a particular MBS and be troubled by the broad range of replies.
Definitions and modeling assumptions vary.
|
|
 |
|
asset-liability
management
Techniques for protecting a firm's solvency in the context of accrual accounting.
callable bond A bond which allows the issuer to repurchase the bond for a specified
price on certain dates prior to the bond's maturity.
compounding
Simple, compound and continuously compound interest.
convertible security
A hybrid security that can or must be exchanged for some other security, usually
the issuer's common stock.
duration and convexity
Risk metrics employed in fixed income markets.
fixed income
term structure Refers collectively to a spot curve, forward curve,
discount curve, yield curve or any other curve that describes the time value of
money at a particulate point in time.
interest
rate parity An arbitrage condition that must hold between the spot interest
rates of different currencies.
interest rate spreads
Spreads between interest rates.
mortgage
backed security A security interest in
mortgage collateral.
option pricing theory
The
body of financial theory used by financial engineers to value options and other
derivative instruments.
return Any of a number of metrics for the
change in an asset's or portfolio's accumulated value
scenario analysis
Formalized "what if" analysis typically performed as a part of asset-liability management
or corporate risk management.
yield
Any of several metrics of the income or return to be earned from an investment. |
|
|
|
 |
|
Hayre (2001)
discusses OAS for mortgage-backed securities in some depth.
|
|
|
|
 |
 |
Ads by Contingency Analysis
|
|
|
 |
|
|
|
|
|
|