Credit Spread Decomposition: Default vs Risk Premium

A 10-year BBB corporate bond yields 2 percentage points more than a Treasury of the same maturity. A reasonable first guess is that this 2 points is the market's estimate of expected losses from default — but historical default rates on BBB bonds are far too low to justify anywhere near that much extra yield. Something else is being paid for. Figuring out what is one of the most useful exercises on a credit desk, because it tells you whether a spread is cheap, fair, or a trap.

Think of buying earthquake insurance in a region with a small, well-known chance of a quake. If insurers only charged you the statistically expected payout, they'd go bankrupt the one year quakes cluster, because payouts arrive at the worst possible time — when everyone needs cash and assets are falling together. So insurers charge more than the bare expected loss, for bearing risk that shows up exactly when it hurts. A credit spread works the same way: part of it pays for defaults that are expected to happen, and a much larger part pays lenders for being exposed to a risk that clusters in bad times.

Building the decomposition

Start from what a lender needs to break even in expectation. If a bond has probability of default $PD$ over the period and, conditional on default, the lender recovers a fraction $R$ of face value (so loses $1 - R$, the loss given default), the actuarially fair extra yield — the number that makes expected payoff equal to the Treasury payoff — is:

$$s_{\text{fair}} \approx PD \times (1 - R)$$

In words: the fair spread is the probability of default multiplied by how much you'd lose if it happened. This is the piece a pure actuary would charge. But the observed market spread $s_{\text{mkt}}$ is bigger:

$$s_{\text{mkt}} = \underbrace{PD \times (1-R)}_{\text{expected loss}} + \underbrace{\pi}_{\text{risk premium}} + \underbrace{\ell}_{\text{liquidity premium}}$$

In words: the spread you actually see priced into the bond equals the expected loss from default, plus $\pi$, a premium for bearing default risk that tends to realize in recessions (the earthquake-insurance logic), plus $\ell$, extra compensation for the fact that corporate bonds are harder to sell quickly than Treasuries, especially in a selloff.

Worked example: decomposing a BBB spread

A 10-year BBB corporate bond trades at a 180 basis point spread over Treasuries. Historical data says BBB issuers default at about 0.20 percent per year on average, and recovery in default averages 40 percent of face value (so loss given default is 60 percent).

1. Expected loss piece. $PD \times (1-R) = 0.0020 \times 0.60 = 0.0012$, or 12 basis points a year.
2. Remaining spread. $180 - 12 = 168$ basis points must be risk premium and liquidity premium combined.
3. Split the remainder. Empirical studies (Elton, Gruber, Agrawal & Mann; later work by Berndt et al.) typically attribute 20-30 percent of the residual to illiquidity and the rest to a genuine risk premium for default-risk exposure. Taking 25 percent as illiquidity: $168 \times 0.25 = 42$ basis points of liquidity premium, leaving $168 - 42 = 126$ basis points of pure default-risk premium.

So of the 180 basis points an investor earns, roughly 12 compensate for expected losses, 126 compensate for bearing the risk of those losses (the part that pays off precisely because defaults cluster in recessions, when everything else in the portfolio is also losing), and 42 compensate for the bond being harder to trade than a Treasury.

Worked example: why the premium exists, in dollars

Consider two portfolios of $100 million each, one holding Treasuries yielding 4.00 percent, one holding a diversified pool of BBB bonds yielding 5.80 percent (the same 180bp spread). Over 10 years, actuarially fair pricing (annual expected loss of 12bp) would predict the BBB pool loses about $100 million times 0.12 percent times 10 years, roughly $1.2 million cumulative to defaults, versus earning $100 million times 1.8 percent times 10 years, about $18 million of extra spread income over the decade. The BBB holder collects roughly 15 times more in extra yield than they lose to actual defaults, on average. The catch: that surplus is not free — it is compensation for the years it doesn't average out, like 2008 or 2020, when defaults spike and BBB bonds get marked down hard exactly when the holder least wants to be selling anything. The long-run average return looks generous because the downside is rare but severe, which is the textbook signature of a risk premium.

180 168 12 0

What this means in practice

Credit desks use this decomposition to ask "is this spread cheap relative to fundamentals?" A bond trading at a spread close to its actuarially fair expected loss, with little risk premium left over, is expensive — you are barely paid for the risk you are carrying. A spread that has blown out far beyond any plausible increase in default probability, as happens in liquidity crunches, is where distressed and relative-value credit funds hunt, betting the market has overpaid for illiquidity and panic rather than genuine credit deterioration. The decomposition is also the basis for credit default swap basis trades: if the CDS-implied default probability and the cash bond spread imply different risk premia for the same issuer, there may be a relative-value trade in the gap.

Key terms

• Credit spread — extra yield a risky bond pays over a risk-free benchmark of the same maturity.
• Probability of default (PD) — likelihood of missing scheduled payments over a horizon.
• Recovery rate / loss given default — fraction of face value recovered in default, and its complement, the fraction lost.
• Risk premium — compensation for bearing risk that tends to realize in bad economic states, over and above expected loss.
• Liquidity premium — extra yield for holding an asset that is harder to sell quickly than the risk-free benchmark.