Option-Adjusted Spread (OAS): Bond Optionality, Models, and Relative Value

What is Option-Adjusted Spread?

Option-Adjusted Spread (OAS) is a model-based spread measure that attempts to separate a bond's spread over a benchmark interest-rate curve from the value of embedded options.

The concept is especially useful for securities whose expected cash flows can change when interest rates change, including:

• callable bonds;
• putable bonds; and
• many mortgage-backed and asset-backed securities.

A simple Credit Spread or Z-spread can treat contractual cash flows as if they were fixed.

But a call, put, or prepayment option can alter when those cash flows are actually received.

OAS therefore asks a more demanding question:

After modeling the embedded option across possible interest-rate paths, what constant spread over the benchmark curve makes the modeled value equal the observed market price?

That makes OAS powerful, but it also makes OAS dependent on the model.

OAS is not simply YTM minus Treasury yield

A simple nominal spread might be calculated as:

1Bond YTM - Treasury yield

That can be useful for a rough comparison, but it uses one bond yield and one benchmark yield.

OAS instead belongs to a curve-based valuation framework.

A model typically projects or represents many possible future interest-rate paths, values the bond's cash flows under those paths, accounts for option exercise or prepayment behavior, and finds the spread that reconciles the modeled present value with the market price.

Conceptually:

1Benchmark rate model
2+
3embedded-option behavior
4+
5constant spread
6-> modeled cash flows and discounting
7-> observed market price

The constant spread is the OAS under that model and assumption set.

Why embedded options distort raw spread comparisons

Suppose two bonds have the same issuer and similar maturity.

Bond A is noncallable.

Bond B can be called by the issuer if rates fall.

Bond B gives the issuer something valuable: the ability to refinance the debt and end the investor's high-coupon cash flows early.

An investor generally wants compensation for granting that option.

Part of Bond B's observed yield spread can therefore reflect the value of the issuer's call option rather than pure credit compensation.

Comparing the raw spread of Bond B with Bond A can make the callable bond look unusually generous.

OAS attempts to remove the modeled option effect so the residual spread can be compared more meaningfully.

Z-spread versus OAS

The Z-spread is the constant spread added to a benchmark spot-rate curve that makes the present value of a bond's specified cash flows equal its price.

For an option-free bond with fixed cash flows, Z-spread can be a useful curve-aware spread measure.

For an option-sensitive bond, the difference is important:

1Z-spread
2-> discounts a specified cash-flow path
3-> does not by itself model how embedded options change cash flows
4
5OAS
6-> models option-sensitive cash flows across rate scenarios
7-> adjusts the spread for modeled option value

For a callable bond, the Z-spread is often greater than OAS because part of the raw spread compensates the investor for the call option granted to the issuer.

The exact relationship depends on the bond, model, and conventions.

A conceptual callable-bond example

Assume a callable corporate bond has:

1Observed Z-spread: 180 bps
2Modeled OAS:        125 bps

A simplified interpretation is that the model attributes part of the raw spread to the embedded call option.

The 55 bp difference should not be casually labeled an exact "option cost" without understanding the model, but it illustrates the economic separation OAS is trying to achieve.

If another similar bond has no embedded option and a 120 bp OAS, the two securities may be much closer in model-adjusted spread than their raw headline yields initially suggested.

OAS helps make unlike structures more comparable.

OAS depends on interest-rate volatility

The value of an option depends partly on volatility.

For a callable bond, greater interest-rate volatility increases the potential value of the issuer's call option because there is a wider range of future rate outcomes in which calling can become advantageous.

CFA Institute's fixed-income framework notes that, for a callable bond, higher assumed interest-rate volatility can lower the calculated OAS, all else equal.

That relationship illustrates an important investor boundary:

1OAS is not a model-free observed fact.

Change the volatility assumption and the modeled option value can change. The OAS required to reconcile the model with the same market price can therefore change too.

The interest-rate model matters

OAS calculations can depend on how future rates are modeled.

A model may differ in its assumptions about:

• the current benchmark curve;
• interest-rate volatility;
• mean reversion;
• rate-path dynamics;
• calibration instruments;
• correlations; and
• numerical implementation.

Two reputable systems can produce different OAS values for the same security if their models and inputs differ.

That does not automatically mean one calculation is wrong.

It means an OAS comparison is strongest when securities are evaluated under a consistent model and assumption set.

Mortgage-backed OAS adds prepayment assumptions

Mortgage-backed securities create an additional layer of complexity.

Homeowners can often prepay their mortgages, especially through refinancing.

When rates fall, refinancing incentives can rise. Principal may return sooner than expected.

When rates rise, refinancing can slow. Cash flows may extend.

That creates negative-convexity behavior and makes the cash-flow path dependent on both rates and borrower behavior.

An MBS OAS model therefore needs assumptions about prepayments in addition to interest-rate paths.

Potential drivers include:

• mortgage rate incentives;
• loan seasoning;
• burnout;
• housing turnover;
• borrower characteristics;
• loan balance; and
• policy or market conditions.

An MBS OAS is only as meaningful as the modeling framework that produces those expected cash flows.

OAS and negative convexity are connected

Convexity describes the curvature of the price-yield relationship.

For an option-free bond, convexity is generally positive.

A callable bond can become negatively convex when falling yields increase the probability of a call and cap price appreciation.

An MBS can behave similarly when falling rates accelerate prepayments.

Because the option changes both value and rate sensitivity, OAS is often analyzed alongside:

• effective duration;
• effective convexity; and
• scenario-based price changes.

A single OAS number does not describe all of those exposures.

OAS is a spread measure, not a duration measure

It is easy to confuse a rich fixed-income data screen because many related metrics appear together.

They answer different questions:

1Yield to Worst
2-> conservative qualifying contractual yield scenario
3
4OAS
5-> model-adjusted spread after embedded option value
6
7Effective Duration
8-> modeled price sensitivity to benchmark-curve changes
9
10Effective Convexity
11-> modeled curvature of that sensitivity

A bond can have an attractive OAS and still have substantial duration or convexity risk.

Likewise, a low-duration security can still have weak credit quality or poor liquidity.

Wider OAS does not automatically mean cheap

A wider OAS means the model requires a larger residual spread over the benchmark curve to reconcile with price after accounting for the modeled option.

That can signal attractive compensation, but it can also reflect genuine risk.

A wide OAS can accompany:

• deteriorating credit quality;
• poor liquidity;
• structural subordination;
• model uncertainty;
• unfavorable collateral behavior;
• event risk; or
• market stress.

The right question is not:

1Which security has the highest OAS?

It is:

1Is the OAS adequate for the credit, liquidity, structural, and model risks being assumed?

Relative value requires context.

OAS is not default probability

Even after adjusting for option value, OAS should not be interpreted directly as expected credit loss.

The residual spread can still include compensation for:

• expected default loss;
• uncertainty around that loss;
• liquidity;
• risk aversion;
• market technicals;
• taxes; and
• model imperfections.

For a corporate bond, credit fundamentals remain essential.

For securitized products, collateral quality, structure, prepayments, and loss allocation can matter as much as the headline spread.

OAS is a valuation lens, not a complete credit model.

OAS and YTW can disagree for good reasons

Yield to Worst evaluates specified contractual redemption scenarios and takes the lowest qualifying yield under the convention.

OAS instead values the option across modeled rate paths.

A callable bond can therefore look different under the two measures.

YTW may be driven by a particular call date. OAS may indicate that the option's modeled value across many paths leaves a different amount of spread compensation than the raw yield suggests.

Neither result automatically invalidates the other.

They answer different questions:

1YTW: What is the lowest qualifying yield under the redemption scenarios?
2
3OAS: What spread remains after modeling the embedded option across rate paths?

Using both can reveal more than using either alone.

OAS comparisons require consistent models

A common analytical mistake is to compare OAS values taken from different providers without checking their methodologies.

Differences in:

• benchmark curves;
• volatility surfaces;
• prepayment models;
• option exercise assumptions;
• settlement conventions; and
• security data

can produce different results.

For cross-sectional screening, consistency can matter more than false precision.

Comparing 125 bps from one model with 140 bps from a materially different model may tell you less than comparing two bonds within the same system.

OAS can change even if price barely moves

Because OAS is derived from several inputs, it can change when the model environment changes.

Suppose the observed bond price is nearly unchanged, but assumed rate volatility rises.

For a callable bond, the modeled issuer option can become more valuable. The OAS needed to reconcile the unchanged market price can fall.

That means an OAS move does not always represent a pure change in market credit sentiment.

Investors should check whether the movement came from:

• bond price;
• benchmark curve;
• volatility;
• expected cash flows; or
• another model input.

OAS is not live authority inside this encyclopedia

This page explains the concept and analytical boundaries.

A current OAS for a real security would require current bond pricing, benchmark curves, security terms, volatility assumptions, and an option-aware valuation model.

For mortgage securities, it can also require a current prepayment model and detailed collateral information.

The encyclopedia does not invent those inputs or present static examples as current market facts.

A future Grizzly Bulls fixed-income research surface would need separately reviewed live-data and methodology authority before publishing current OAS values.

How investors should use OAS

A disciplined OAS workflow asks:

1. Does the bond contain an economically meaningful embedded option?
2. Which benchmark curve and interest-rate model are being used?
3. What volatility assumption drives the option value?
4. For securitized products, what cash-flow or prepayment model is used?
5. Is the comparison set calculated under the same methodology?
6. How does OAS compare with raw spread and YTW?
7. What do effective duration and convexity say about rate sensitivity?
8. Is the remaining spread adequate for credit, liquidity, and model risk?

OAS is valuable because it tries to compare spread economics after recognizing that options alter bond cash flows.

It becomes dangerous when the model output is treated as an observable law of nature.

For broader valuation context, investors can continue into the Cyclically Adjusted Risk Premium and Macroeconomic Conditions Index. These are separate market research frameworks and do not supply OAS inputs or live bond valuations to this article.

Sources and further reading

• CFA Institute, 2026, Valuation and Analysis of Bonds with Embedded Options: https://www.cfainstitute.org/insights/professional-learning/refresher-readings/2026/valuation-analysis-bonds-embedded-options
• CFA Institute, 2026, The Arbitrage-Free Valuation Framework: https://www.cfainstitute.org/insights/professional-learning/refresher-readings/2026/arbitrage-free-valuation-framework
• CFA Institute, 2026, Curve-Based and Empirical Fixed-Income Risk Measures: https://www.cfainstitute.org/insights/professional-learning/refresher-readings/2026/curve-based-and-empirical-fixed-income-risk-measures
• CFA Institute, 2026, Yield and Yield Spread Measures for Fixed-Rate Bonds: https://www.cfainstitute.org/insights/professional-learning/refresher-readings/2026/yield-and-yield-spread-measures-for-fixed-rate-bonds

Explore more topics in the Financial Research Encyclopedia.