WhitmanTrading

What Is Credit Risk?

Credit risk is the risk that a borrower fails to make payments owed, causing the lender a loss. Its payoff is fundamentally asymmetric: the best case is being repaid exactly as agreed, so the upside is capped while the downside runs to the full amount lent.

Credit risk is the oldest risk there is: you lent somebody money and they may not give it back. The shape of that bet is what makes lending a fundamentally different discipline from investing.

How it works

A price series where a borrower fails to pay.
Credit risk is the borrower not paying. Illustrative chart - not real market data.

A borrower promises a schedule of payments. Credit risk is the possibility that some or all of those payments do not arrive.

A steady series where full repayment is the best outcome.
The best case is being repaid as agreed. Illustrative chart - not real market data.

The best possible outcome is the contract performing. No borrower ever pays back more than agreed because their business went unexpectedly well.

A rising series where upside is capped.
So the upside is capped and the downside is not. Illustrative chart - not real market data.

The worst is losing the entire principal. Between those two sits a range of partial recoveries, and the distribution is heavily skewed toward the good case with a long thin tail.

A falling series contrasting the equity payoff.
Which is the opposite shape to owning equity. Illustrative chart - not real market data.

The arithmetic lenders use

A choppy series where expected loss is calculated.
Expected loss is probability times exposure times severity. Illustrative chart - not real market data.

Probability of default is how likely the borrower is to fail over a stated period, estimated from ratings, models or market prices.

A slow series where default risk accumulates over years.
And different again over a long horizon. Illustrative chart - not real market data.

Exposure at default is how much is outstanding when it happens, which for a revolving facility is often more than the current balance — borrowers draw down as they deteriorate.

A calm series where no defaults occur.
A quiet stretch hides what it measures. Illustrative chart - not real market data.

Loss given default is the share not recovered after security is realised. Multiply the three together and you have expected loss, which is what the interest rate must cover before any profit.

A worked example

A portfolio of loans with a 2% annual default probability and a 60% loss given default. Expected loss is 1.2% of the book each year.

Price the loans at 5% over funding costs and the expected margin is 3.8% before operating costs — a reasonable business if the estimates hold.

A falling series with a stop level marked.
A stop fills where the market is. Illustrative chart - not real market data.

Now double the default rate in a recession. Expected loss goes to 2.4%, and recoveries fall at the same time because collateral is worth less when everybody is selling it.

So both inputs deteriorate together. The loss rate rises by considerably more than the default rate, which is the correlation that makes credit downturns worse than a linear projection suggests.

Why clustering is the whole problem

Individual defaults look independent. One company failing tells you little about the next, and a diversified book should therefore be predictable.

Common factors bind them. Interest rates, the economic cycle, a sector shock — anything affecting many borrowers at once converts a diversified book into one correlated exposure.

Which is why credit losses arrive in waves. Years of below-average losses, then a period where several years of expected loss occurs at once.

And it is why capital, not pricing, is the real defence. Pricing covers the expected loss; capital covers the unexpected one, and the gap between them is where lenders fail.

The original data

On this site’s shared series 95% of bars sit below a prior peak, the maximum decline is 3.76%, and the longest below-peak stretch runs 73 bars, which finished +3.61%.

Credit returns do not look like that. They are flat and positive for long periods, then step down sharply, and there is no gradual recovery — a defaulted loan does not come back the way a price does.

A candlestick chart annotated with the cost of a round trip.
A round trip costs a share of a bar. Illustrative chart - not real market data.

And this site’s fee measurement shows what the margin must survive: 75 basis points costs 20.2% of a thirty-year balance. A lending margin of a few percent, less expected losses, less costs, is a thinner business than the headline rate suggests.

A price series with volume shown beneath.
Volume and price measure different things. Illustrative chart - not real market data.

How it is managed

Security. Taking a claim over assets reduces loss given default without changing the probability of default at all — the two are separate levers.

Covenants. Conditions requiring the borrower to maintain financial ratios give the lender a right to act before a default, which is worth more than the ratio itself.

Diversification. Spreading across borrowers, sectors and geographies works against idiosyncratic failure and not against the common factor.

And credit derivatives. Buying protection transfers the exposure to somebody else, which introduces counterparty credit risk in its place — a smaller risk, usually, and not no risk.

How it is priced in markets

A credit spread is the market’s estimate. The extra yield a borrower pays over a government benchmark is compensation for expected loss plus a premium for the uncertainty around it.

Ratings are the other estimate. Agencies assign letters reflecting a view on default probability, and they move slowly by design, which makes them stable and often late.

The two disagree frequently. Market spreads move daily on sentiment and liquidity as well as on credit, so a bond can trade at distressed levels while still carrying an investment grade rating.

And the disagreement is informative. A widening spread on an unchanged rating is the market saying something the agency has not yet said, and it has historically been the earlier of the two signals.

When it fails

The characteristic failure is a book that looks diversified and shares one driver. A lender holds thousands of loans across many borrowers, all secured on property in one region. Each loan was assessed individually and the book passes every concentration test written in terms of borrower names. Then property prices fall, and default probability and recovery rate deteriorate together across the entire portfolio. The diversification was real against individual borrower failure and absent against the variable that actually determined the outcome.

A candlestick series with a gap through a level.
A gap skips the level entirely. Illustrative chart - not real market data.

A second failure is pricing off a benign period. Loss estimates from years without a recession understate the cycle.

A third is assuming recovery rates are stable. They fall precisely when defaults rise.

A fourth is lending into a growing market late. The last loans written before a downturn carry the worst terms and the highest losses.

A declining series cut short at a decision point.
Five years, almost no losses. Lend more? Illustrative chart - not real market data.

And a fifth is treating expected loss as the risk. It is the average; the capital exists for the years that are not average.

Counterparty credit risk covers the version where the exposure moves with the market. Consumer credit risk covers lending to households. And credit spread covers what the market charges for all of this.

What I actually do

Lending has the worst payoff shape in finance. Everything goes right and you get your money back plus interest; something goes wrong and you can lose all of it. Every practice in credit — covenants, security, diversification, pricing — exists to compensate for that asymmetry rather than to remove it.

— Michael Whitman

This page is educational, not financial advice. Test every idea on your own charts before risking money.