What have you actually learned to read?

You started this course able to see that a corporate bond yields more than a government bond. You finish able to say what that extra yield is paying for, decompose it, and read it as a signal about the market as a whole. Here is the whole toolkit in one place.

The spread and its parts

Spread = corporate yield − matched benchmark yield, quoted in basis points, and made of three things:

Spread ≈ expected loss + risk premium + liquidity premium

Expected loss is usually the minority of the spread. The rest is compensation for uncertainty and illiquidity — earned in most years, handed back in the concentrated bad ones.

The arithmetic worth memorising

• Expected loss = PD × LGD
• LGD = 1 − recovery rate
• Implied PD ≈ spread ÷ LGD — a risk-neutral price statement, never a forecast
• Price change ≈ − spread duration × change in spread

Four lines. Everything quantitative in this course is one of them.

Ratings, and the institutions around them

• The scale is ordinal — an opinion about relative creditworthiness, not a probability, not a recommendation.
• Issuer ratings cover the borrower as a whole and issue ratings are notched for seniority and recovery — and scales differ on whether the issuer letter measures default probability alone or expected loss.
• Migration is more common than default: transition matrices are diagonal-dominant, downgrades cluster, and a downgrade costs money through mark-to-market long before any default.
• The investment-grade boundary is a convention with outsized force, because mandates, indices, regulation and collateral rules are written around it.
• The issuer-pays model carries documented structural conflicts — shopping, revenue concentration, downgrade reluctance, the regulatory licence — partly addressed by disclosure, supervision and separation of commercial and analytical functions. Use a rating as one input with a known incentive structure.

The queue, and what you get back

• The waterfall runs administrative and super-priority claims, secured, senior unsecured, subordinated, hybrids, equity — the absolute priority rule, which bends in negotiated restructurings.
• Structural subordination means a holdco bond ranks behind opco creditors, whatever its label says.
• Recoveries order by seniority, vary enormously within each category, and fall when defaults rise — PD and LGD are positively correlated, which is why bad years are doubly bad.
• Covenants are principally an LGD tool: they protect recovery and force early conversations. Read the definitions, not just the levels — headroom is a function of both.

Spreads as a market signal

• The credit cycle runs through cheap funding, competition on terms, quiet accumulation of weak credit, a shock, rising defaults with falling recoveries, retrenchment, and reset. The mechanism is documented; the timing is not.
• Spreads move together because the price of risk is market-wide, fundamentals correlate, dealer capacity shrinks, and forced selling happens. Diversification removes single-name risk, not the common factor.
• CDS prices the same three components as a bond spread, encoded in a traded contract, and gives a continuous, spread-quoted read on credit — including for sovereigns.

The disciplines that keep you honest

• Compare like with like — same spread measure, same benchmark curve, same maturity, same seniority, same entity in the group.
• State your recovery assumption whenever you quote an implied default probability. Without it, the number is half-finished.
• Never read a spread as a forecast. It is a price, containing risk premium and liquidity alongside expected loss.
• Remember the correlation. PD and LGD deteriorate together; models that hold recovery fixed understate exactly the scenarios that matter.

Every number in this course was an observation about price and structure, never a recommendation about an issuer, a rating category or a credit instrument, and never a prediction about defaults or spread direction. Holding that line is not modesty — it is what makes the analysis worth anything.

Before you sit it

Each of these is a minute at your desk. Any one that is not names the lesson to reopen first.

• Turn a 2% default probability and a 40% recovery into basis points of spread — How do you turn a default probability into basis points?
• Name the one notch on the rating scale that matters more than all the others, and say why — Why is one notch on the scale worth more than all the others?
• Put four claimants in the order a liquidation pays them — Who gets paid first when a company runs out of money?
• Say why spreads on unrelated issuers widen on the same morning — Why do spreads blow out all at once?

Try it now

Before the checkpoint quiz, prove the toolkit on one credit:

1. Pick one credit and state its spread over the matched benchmark, in basis points. Either the high-quality corporate 5-year for August 2026 minus the Treasury 5-year averaged over the same month (the first two tables), or one government's credit default swap spread over Switzerland (the third).

2. Split it: derive expected loss from a stated PD and recovery, and name the remainder as risk-and-liquidity premium.
3. Write one comparative, non-predictive sentence about how that credit is priced relative to a peer — and name the one structural difference that most affects the comparison. If you can do that, you have earned this checkpoint.