What is a credit spread actually made of?

A spread of 200bp looks like a single number. It is not. It is a sum of at least three distinct payments bundled into one quote, and confusing them is the most common mistake in credit analysis.

The three components

Spread ≈ expected loss + risk premium + liquidity premium

• Expected loss. The part that, on average, gets consumed by actual defaults. If a pool of similar bonds loses 1.2% a year to defaults over the long run, then 120bp of the spread is not income at all — it is a reserve you are being handed to cover losses that will arrive. The next lesson computes this precisely.
• Risk premium. Compensation for uncertainty about that loss, not for the loss itself. Defaults do not arrive smoothly at 1.2% a year; they arrive in clusters, in exactly the years when everything else you own is also falling. Investors demand extra payment for a risk that shows up at the worst possible moment.
• Liquidity premium. Compensation for the fact that a corporate bond is harder to sell than a government bond. Most corporate bonds trade over the counter, thinly, at a wider bid-offer. You are being paid for the possibility that when you want out, exiting is expensive.

A small residual also exists — tax treatment in some markets, and structural quirks in how a bond is documented — but the three above carry the story.

A worked decomposition

Take an illustrative bond quoted at 300bp:

• Expected loss: 120bp (worked out from default and recovery assumptions in the next lesson)
• Liquidity premium: 60bp
• Risk premium: 300 − 120 − 60 = 120bp

Read that carefully. Only 40% of the spread is compensation for losses that are actually expected to happen. The majority is payment for uncertainty and for illiquidity.

The pattern this reveals

This is a well-documented feature of credit markets, sometimes called the credit spread puzzle: for higher-quality bonds especially, spreads have historically been several times larger than the realised losses over the same periods. In a benign decade, an investor who simply collected the spread would have kept most of it.

It is tempting to read that as a standing bargain. It is not, and the reason is structural: the excess over expected loss is precisely the part that is paid for bearing the bad state. It is compensation you earn in most years and give back, sometimes violently, in the years when defaults cluster and liquidity evaporates together. Unit 4 shows exactly that mechanism at work.

Why the split matters practically

Two bonds can quote the same 300bp for completely different reasons:

• Bond A: expected loss 240bp, risk and liquidity premium 60bp — a genuinely weak credit priced tightly.
• Bond B: expected loss 40bp, risk and liquidity premium 260bp — a solid credit that is small, rarely traded, and hard to exit.

Same headline number, entirely different exposure. One is a fundamental credit question; the other is a market-structure question. Reading the spread as one lump hides that completely.

In the data

Published estimates split a borrower's price of risk into pieces rather than quoting one total. The table is Aswath Damodaran's annual set for Brazil: a default spread of 0.021275 beside a country risk premium of 0.03241.

The premium is deliberately larger than the default spread, because the two answer different questions: what lenders to Brazil's government are charged, and what extra return its shares must offer. Every value is a fraction, not basis points: 0.021275 means 213bp.

Try it now

Take one spread apart:

1. The two tables below are the same month of high-quality corporate 5-year yield and the Treasury 5-year. Average the three Treasury readings, subtract the result from the corporate yield, and write the spread in basis points.

2. The table below is a lookup: the default spread a borrower in each rating bucket is typically charged. Six buckets are shown, as fractions, so 0.005993 means 60bp. The corporate curve is built from bonds rated roughly AAA to A; set your spread beside the Aa2 and A1 values. Same neighbourhood, and nothing in either number says how much of it is expected loss.

3. Now the expected loss, with the assumption stated: an illustrative single-A credit with a 0.1% annual default probability and a 60% loss given default expects 0.001 × 0.60 = 6bp a year (the next lesson derives the formula). Subtract that from your spread. The remainder is the combined risk-and-liquidity premium, and noticing how large a share of the spread it is, is the whole point of this lesson.