A higher bond yield can be caused by a longer clock, a weaker borrower, a less liquid market—or all three at once. Calling every extra percentage point a “risk premium” hides which risk the market is actually pricing.
This is why two concepts must be kept separate: term premium and credit spread. Term premium belongs to the compensation investors require for uncertainty about interest rates across time. Credit spread belongs to the additional yield on a credit-risky claim relative to a chosen reference, reflecting expected loss, risk compensation, liquidity and other instrument-specific effects.
The two can move in the same direction. They can move in opposite directions. They can be estimated using different models. They should not be collapsed into one number merely because both appear inside a long-term bond yield.
This article is part of Batch 034 of the eduKateSG Finance Authority 400. The Yield Curve owns the cross-maturity structure. Duration owns price sensitivity. This page owns the yield-decomposition boundary between time compensation and borrower-specific credit compensation.
Term premium prices uncertainty about the path of rates. Credit spread prices the extra risk of the particular credit claim relative to its benchmark.
Educational boundary: neither term premium nor credit spread has one universally agreed decomposition. Term premium is model-estimated rather than directly observed; corporate spreads also contain liquidity, tax, technical and option effects in addition to expected default loss.
Start with the benchmark
Suppose a five-year corporate bond yields 5.8% and a comparable five-year government benchmark yields 3.5%. The observed spread is 2.3 percentage points, or 230 basis points.
CORPORATE YIELD − REFERENCE YIELD = OBSERVED CREDIT SPREAD.
That arithmetic is simple. Interpretation is not.
The 3.5% government benchmark already contains the market’s expectations of future short-term government yields and a term premium. The additional 2.3% corporate spread can contain expected default loss, compensation for uncertainty about that loss, liquidity differences, tax effects, market technicals and instrument-specific features.
The corporate yield therefore stacks several layers.
A useful conceptual stack
For teaching purposes, a long-term corporate yield can be written conceptually as:
CORPORATE YIELD ≈ EXPECTED FUTURE SHORT RATES + TERM PREMIUM + CREDIT / LIQUIDITY / INSTRUMENT SPREAD.
This is not an accounting identity with perfectly observable components. It is a diagnostic map.
The first two components can be used to describe the government or risk-free-like term structure. The last component asks why this particular borrower and bond need to pay more than the reference curve.
What is term premium?
The Federal Reserve Bank of New York’s ACM framework describes Treasury yields as having two conceptual components: expectations of the future path of short-term Treasury yields and a Treasury term premium. The New York Fed defines the term premium as compensation investors require for bearing the risk that interest rates over the bond’s life do not evolve as expected. See New York Fed Treasury Term Premia.
Crucially, the New York Fed also states that term premium is not directly observable. It must be estimated from models using financial and macroeconomic data.
This means a ten-year government yield cannot be split with certainty into an exact expectations number and an exact term-premium number. Different reasonable models can produce different estimates.
Why a term premium exists
Imagine lending for ten years at a fixed rate. You do not know today what one-year rates will be in Years 6, 7, 8 or 9. Inflation can surprise. Growth can surprise. Monetary policy can surprise. Bond supply and demand can surprise.
If investors dislike carrying that uncertainty, they can demand compensation for locking in the longer claim. That compensation is one interpretation of a positive term premium.
The premium can also be low or negative in some model estimates and market environments. Strong demand for long-duration safe assets, hedging value and other factors can make investors willing to accept lower long yields than a simple future-short-rate expectation would imply.
Term premium is not the same as maturity itself
Time is a dimension. Term premium is compensation for risk associated with holding that time under uncertainty.
A ten-year bond does not mechanically have a fixed “ten-year premium.” The premium can change while the legal maturity stays the same.
This is why the Yield Curve can move even without a corresponding change in expected policy rates.
What is a credit spread?
A credit spread is the yield difference between a credit-risky bond and an appropriate reference yield, usually chosen to match maturity or duration as closely as practical.
Investor.gov explains that corporate bonds carry credit risk because the issuer may fail to make promised payments. Corporate yields therefore often exceed government benchmarks to compensate investors for additional risk. See Investor.gov’s corporate-bond bulletin.
The spread is observable once the two yields and comparison method are chosen. The reasons inside the spread are not perfectly observable.
Expected default loss is only one part of credit spread
A simple expected-loss model might say:
EXPECTED CREDIT LOSS ≈ PROBABILITY OF DEFAULT × LOSS GIVEN DEFAULT.
If a one-year bond has a 2% default probability and 50% expected loss given default, the simplified expected loss is 1% of exposure.
But a 150-basis-point credit spread would not automatically mean the remaining 50 basis points are “profit.” Investors may require compensation for uncertainty, illiquidity, systematic risk, model error and other features.
Spread is a market price. Expected loss is a probability-weighted model of cash-flow impairment. They are related and not identical.
Recovery changes the economics of default risk
Two companies can have similar default probabilities and different expected losses because their recovery structures differ.
A senior secured bond with strong collateral may recover far more in default than subordinated unsecured debt from the same issuer. That can justify different spreads even if both claims share the same company-level probability of distress.
The prior Bond Default article owns the recovery waterfall.
Liquidity can widen a credit spread without the issuer getting weaker
Suppose two bonds have identical expected cash flows, but Bond A trades actively every day and Bond B rarely trades. An investor who may need to sell quickly can demand a higher yield for Bond B.
That extra yield is not expected default loss. It is compensation for liquidity risk and market friction.
During periods of stress, liquidity can disappear at the same time credit concerns rise, making observed corporate spreads widen far more than changes in expected default alone would suggest.
Spread depends on the benchmark you choose
A five-year corporate bond can be quoted against a government bond, a swap curve, a fitted zero curve or another market benchmark.
If the benchmark changes, the numerical spread changes.
Suppose a corporate yield is 5.5%. The five-year government yield is 3.0%, producing a 250-basis-point spread. A five-year swap rate might be 3.4%, producing a 210-basis-point spread. Both calculations can be correct under their stated benchmarks.
A spread should therefore always be quoted with its reference.
Nominal spread, zero-volatility spread and option-adjusted spread
Fixed-income analysis uses several spread measures.
- Nominal spread compares a bond’s yield with a benchmark yield at a chosen maturity.
- Zero-volatility spread adds a constant spread across a benchmark spot curve so the discounted cash flows match the bond’s price.
- Option-adjusted spread attempts to remove the value of embedded options from the spread under a model.
The measures answer different questions. A callable bond’s nominal spread can look attractive partly because the investor has sold an option to the issuer.
Do not compare spread numbers unless their construction is compatible.
Term premium can rise while credit spread falls
Imagine strong economic growth improves corporate cash-flow expectations. Credit spreads tighten because default risk appears lower. At the same time, long government yields rise because investors demand more term premium for inflation and rate uncertainty.
The corporate bond yield can still rise even though the company becomes safer.
For example:
| Before | After | |
|---|---|---|
| Government benchmark yield | 3.0% | 4.0% |
| Corporate spread | 2.0% | 1.2% |
| Corporate yield | 5.0% | 5.2% |
The corporate yield rose by 0.2 percentage points even though the credit spread improved by 0.8. The government-rate component rose more.
Credit spread can rise while term premium falls
Now imagine a recession scare. Investors expect future policy rates to fall and demand safe government bonds, pushing long government yields down. At the same time, corporate earnings expectations deteriorate and credit spreads widen.
A corporate bond can fall in price even while the government bond market rallies.
This is common enough conceptually that “rates fell, so bonds went up” should never be applied blindly to credit-risky bonds.
Worked decomposition: one yield, three moving layers
Consider a fictional ten-year corporate yield of 6.5%.
| Conceptual component | Illustrative contribution |
|---|---|
| Expected average future short-rate path | 3.0% |
| Estimated term premium | 0.8% |
| Observed corporate spread above government curve | 2.7% |
| Total yield | 6.5% |
Now suppose six months later the expected short-rate component falls to 2.6%, estimated term premium rises to 1.1% and corporate spread widens to 3.4%. Total yield becomes 7.1%.
The issuer’s borrowing cost rose even though expected future short rates fell.
The observed outcome can only be understood by separating the layers.
Term premium is model risk as well as market risk
Because term premium is estimated rather than observed, an analyst faces model risk. Different models can attribute the same ten-year yield move differently between future-rate expectations and term premium.
The New York Fed’s ACM estimates are useful because the method is explicit and data are published. They are not official forecasts of the Federal Reserve or a directly observed market price of term premium.
A Wintour v1.0 reading therefore labels term premium as an estimate, not a fact masquerading as a measurement.
Credit spread is observable but its internal causes are also estimated
If a corporate bond yields 6% and the reference government yield is 4%, the two-percentage-point spread is observable under the stated comparison.
But saying “1.2% is default risk, 0.5% is liquidity and 0.3% is risk aversion” requires a model or inference.
The spread itself is data. Its decomposition is analysis.
Credit spreads can widen before actual defaults rise
Markets price expectations. If investors expect recession, weaker earnings, tighter refinancing and lower recoveries, spreads can widen before companies actually miss payments.
This forward-looking behaviour is useful and imperfect. Spreads can overshoot during panic or compress excessively during optimism.
The price of risk is therefore partly a forecast and partly a reflection of current willingness to bear uncertainty.
Spread tightening does not prove the company improved
A corporate spread can tighten because the issuer’s balance sheet improved.
It can also tighten because investors collectively become more willing to take credit risk, because liquidity improves or because demand from funds increases.
Likewise, a spread can widen because the company deteriorated or because the entire market suddenly demands more compensation for corporate risk.
Issuer-specific and system-wide components must be separated where possible.
The yield curve and credit curve are different structures
A government yield curve maps time for a benchmark issuer. A corporate credit curve maps the additional spread required for one corporate borrower across maturities.
A company’s two-year spread can be 100 basis points while its ten-year spread is 180. Another distressed company can show an inverted credit curve where short bonds trade at very wide spreads because near-term default or refinancing risk is concentrated.
The shape of a credit curve therefore contains information about when the market expects credit stress to matter.
Maturity can increase term risk without increasing credit spread
Two government bonds from the same strong issuer can have almost no corporate-style credit spread difference yet very different yields because of maturity and term premium.
Conversely, two bonds with the same maturity can have very different yields because their credit spreads differ.
This is the cleanest way to remember the ownership boundary:
TERM PREMIUM CHANGES WITH THE PRICE OF TIME RISK.
CREDIT SPREAD CHANGES WITH THE PRICE OF THE CREDIT CLAIM RELATIVE TO THE BENCHMARK.
Duration converts each spread into price sensitivity
A bond can have duration to the benchmark curve and sensitivity to credit-spread changes.
Suppose a corporate bond has government-rate duration of 6 and spread duration of 5. A 0.50-percentage-point rise in the benchmark curve and a 1.00-percentage-point widening in credit spread can create roughly −3% and −5% first-order price effects respectively, before convexity and other interactions.
The Duration article owns the sensitivity method. This example shows why the yield decomposition should occur before the hedge or risk estimate.
A higher spread can coexist with a lower all-in yield
Suppose a corporate spread widens from 1.5% to 2.0% while the government benchmark falls from 4.5% to 3.0%.
The all-in yield falls from 6.0% to 5.0% even though credit conditions worsened relative to government bonds.
An issuer might therefore refinance at a lower absolute rate while the market is simultaneously expressing a more negative view of its credit quality.
Absolute yield and relative spread answer different questions.
A lower spread can coexist with a higher all-in yield
The reverse can happen when the government curve rises more than the credit spread falls.
This matters for corporate treasurers. A company can improve its balance sheet and market credit standing yet still face higher borrowing costs because the whole interest-rate environment moved against it.
Credit management cannot control the benchmark curve. It can influence the spread layer through leverage, liquidity, operating quality, collateral and market communication.
Term premium is system-wide; credit spread can be issuer-specific
A rise in estimated Treasury term premium can affect long-term financing conditions across many borrowers simultaneously because the benchmark curve itself moves.
An issuer-specific scandal or earnings collapse can widen one company’s spread while leaving the government curve largely unchanged.
A recession shock can do both: benchmark yields can fall through rate expectations while credit spreads widen system-wide.
Separating common and issuer-specific shocks helps identify what can be repaired by the company and what belongs to the wider financial environment.
Failure-first reading: what would make the decomposition misleading?
- The benchmark maturity does not match the corporate bond well.
- The bond contains an embedded call or conversion option.
- Liquidity is unusually poor.
- The government benchmark itself contains special collateral value.
- Term-premium estimates are highly model-dependent.
- The issuer is distressed and yield-to-maturity becomes a poor summary of expected recovery.
- Tax or regulatory treatment materially differs between the two securities.
The spread calculation may remain arithmetically correct while the economic interpretation becomes incomplete.
The credit-spread versus term-premium diagnostic
- What is the all-in yield?
- What reference curve is being used?
- How closely are maturity and duration matched?
- What is the observed spread?
- What part of the benchmark yield reflects expected future short rates?
- What term-premium estimate is being used?
- How model-dependent is that estimate?
- What expected default loss is plausible?
- What recovery assumptions matter?
- What liquidity premium may sit inside the credit spread?
- Are embedded options affecting the spread?
- Did the benchmark move, the spread move, or both?
- Is the move issuer-specific or system-wide?
- What duration converts each component into price risk?
Observable mastery test
A company’s ten-year bond yield rises from 5% to 5.4%. A commentator says, “Credit risk clearly worsened.” You discover that the ten-year government benchmark rose from 3% to 4%, while the corporate spread narrowed from 2% to 1.4%.
The correct conclusion is that the company’s all-in yield rose, but its spread relative to the benchmark tightened. The observed move is more consistent with a rise in the government-rate component overwhelming an improvement in the relative credit spread.
You understand the decomposition when “yield up” stops being a synonym for “credit worse.”
The World Return: how yield components reach the real borrower
Future-rate expectations / term premium → government benchmark curve → corporate credit and liquidity spread → all-in borrowing yield → refinancing and investment hurdle → corporate action → future cash flow and credit quality → new spread.
The decomposition matters because different problems require different responses. A company cannot directly control the global term premium. It can reduce leverage, extend maturities, improve liquidity or strengthen collateral to influence its credit spread.
Finance becomes useful when the source of the higher borrowing cost is identified before management tries to repair the wrong layer.
Research anchors
The Federal Reserve Bank of New York Treasury Term Premia page explains the decomposition of Treasury yields into expected future short rates and estimated term premium, and explicitly notes that term premium is not directly observable. Investor.gov’s corporate-bond bulletin provides the credit-risk and bond-yield foundation. The decomposition tables are original teaching examples.
Continue through Batch 034
Read The Yield Curve and Duration. Continue to Yield-Curve Inversion | What the Shape Can Signal—and What It Cannot Prove. Return to How Finance Works for the complete Finance map.