A single bond has a coupon rate, a current yield and a yield to maturity, and they are three different numbers. The gap between the last of them and a government bond is the market's own estimate of what could go wrong.
13 min read · Level 2
The most consumed piece of information about a bond is the rating and the least consumed is the spread, and the ranking should be the other way around. A rating is a third party's ordinal opinion about default likelihood, published slowly, revised late, and paid for by the borrower. A spread is the aggregated judgement of everybody currently willing to trade the instrument, updated continuously, free to read, and it embeds the rating along with everything the rating omits. More importantly, the spread converts into something a rating never will: an explicit break-even probability that you can accept or reject with reasons. When you turn 238 basis points into a one-in-four chance of default over seven years, you have stopped consuming an opinion and started forming one, and you can be shown to be wrong - which is the property that makes a view worth holding. None of this means the rating is worthless; the historical link between rating and default frequency is real and the mandate effects around the investment grade line are a genuine market force. It means that a rating is a starting point that most holders treat as a conclusion, and that the free number sitting next to it is doing far more work.
Start by pinning down what the bond actually yields, because the word is used for three different things and a seller will quote whichever is highest.
Bond C: a corporate issue, 5% coupon, seven years remaining to maturity, quoted at 94.50 per 100 of par - so 945 for a 1,000 unit.
Coupon rate 5.0000% 50 / 1,000 par. A fact about the contract. It never changes. Current yield 5.2910% 50 / 945. What the annual income is against what you pay today. Yield to maturity 5.9847% The rate that discounts the schedule back to 945.
Only the last of the three is a return. Current yield ignores the fact that 945 becomes 1,000 at maturity, which is 55 of additional return the holder receives and current yield does not count. Coupon rate ignores the price you paid altogether.
Yield to maturity has no closed-form solution for a multi-period bond; it is found numerically, by trying rates until the discounted schedule equals the price. Here, discounting seven coupons of 50 and 1,000 of principal at 5.9847% gives 945.00. That is why the figure is not a round number, and why any yield to maturity you have not solved for is a figure you are taking on trust.
The ordering is diagnostic and worth memorising, because it tells you which side of par a bond is on before you have looked at anything else.
Discount bond (price below par): coupon < current yield < yield to maturity Bond C: 5.0000% < 5.2910% < 5.9847%
Premium bond (price above par): coupon > current yield > yield to maturity A 6% coupon, 5-year bond at 1,089.04: 6.0000% > 5.5094% > 4.0000%
The premium case is the one that catches people. That bond pays a 6% coupon into a 4% market and looks generous on the ticket, and its yield to maturity is 4.0000%, because the buyer paid 1,089.04 for something that will return 1,000. The extra coupon is largely the repayment of the premium, spread over five years.
Because once you know what the bond yields, the next question is what the extra is for - and the market has already answered it in a number you can read for free.
Suppose seven-year government bonds in the same currency yield 3.60%. Bond C yields 5.9847%. The difference,
5.9847% - 3.60% = 2.3847 percentage points = 238 basis points
is the credit spread. It is the market's compensation for everything the government bond does not carry: the chance the issuer does not pay, the chance you cannot sell without a concession, and any options in the contract that work against you.
You can convert a spread into an implied default rate, which is the single most useful thing to do with it, because it turns an abstract 238 basis points into a proposition you can accept or reject. Suppose that on default a holder recovers 40 cents in the dollar - so a default costs 60 cents. On a simple annual expected-loss basis, the spread breaks even when
annual default probability = spread / (1 - recovery) = 2.3847% / 0.60 = 3.9745%
Over the bond's seven remaining years, that is a cumulative probability of
1 - (1 - 0.039745)^7 = 1 - 0.752845 = 24.7155%
So buying Bond C at 945 is roughly a statement that you think this issuer is less than about one-in-four likely to default at some point in the next seven years, given a 40% recovery. Change the recovery assumption and the proposition changes with it: at 25% recovery the break-even annual rate falls to 3.1796% and the cumulative figure to 20.2432%.
Be honest about what this calculation is. It is a crude one-factor break-even. It ignores when in the seven years the default happens, ignores the coupons already collected before it, ignores the tax position and ignores the fact that a spread also pays for illiquidity rather than default alone. It is not a valuation. What it is good for is converting a yield into a question you can actually have an opinion about, which the yield on its own does not do.
Then there is the rating. A credit rating is an ordinal opinion, published by an agency the issuer usually pays, about the relative likelihood that an obligation is repaid. It is genuinely informative - the historical relationship between rating and default frequency is one of the more durable regularities in finance - and it is routinely asked to do work it cannot do.
What a rating tells you the agency's relative ranking of default likelihood What it does not tell you whether the price compensates you for that likelihood how far the price will move if rates change whether you could sell the bond next Tuesday what you would recover if the default happened anything at all about a callable bond's option
Above all it does not tell you whether the bond is a good purchase, because it says nothing about price - and price is the entire question. The investment grade boundary is worth naming for a reason that is about markets rather than businesses: many institutional mandates are written to hold only investment grade paper, so a downgrade across that line forces sales by holders who have no view at all. That is a real and predictable price effect, and it is a fact about the rules, not about the borrower.
Educational, not personalised financial advice. Nothing in the Investor Academy is a recommendation to buy, sell or hold anything — what is right for you depends on your own circumstances and the rules where you live.
The last thing to check on any specific issue is whether the schedule you are discounting is the schedule you will actually receive - because the issuer often holds an option to end it early.
Bond D: 6% coupon, eight years to maturity, callable by the issuer in three years at 102 (that is, the issuer may repay 1,020 per 1,000 of par at that date and walk away). It is quoted at 108.
Current yield 60 / 1,080 = 5.5556% Yield to maturity eight years, redeeming at 1,000 = 4.7735% Yield to call three years, redeeming at 1,020 = 3.7718%
Three yields again, and this time the gap is 100 basis points of return that may simply not happen. Which one is real depends on a decision the issuer makes, not you - and the issuer will call when calling is good for the issuer, which is when rates have fallen and it can refinance more cheaply. That is precisely the moment you would least like to have your money back, because you would be reinvesting it into the same lower rates.
The convention is to quote the lower of the two as the yield to worst, and the discipline is to underwrite on that figure. If 3.7718% is not enough for this credit, the bond is not attractive, whatever the 4.7735% on the sheet says. A call option held by your counterparty is not a neutral feature; it is a right you sold them, and the coupon is the price you were paid for it.
Then the covenants. These are the promises in the bond's own documentation that constrain what the issuer may do while your money is outstanding: limits on additional borrowing, on paying dividends, on selling assets, sometimes a requirement to maintain a stated interest cover. Covenants matter most in the case where they are hardest to think about - the deteriorating issuer - because they are what forces a conversation before the money is gone rather than after. A bond with a materially higher yield and materially weaker covenants than its peers is not offering you more return for the same thing.
Finally the ability to pay, which is the part the financial statements answer. Interest cover tells you whether this year's earnings comfortably carry this year's interest. Net debt against earnings tells you the scale of the obligation. Free cash flow tells you whether the interest is being paid out of operations or out of new borrowing. And the maturity schedule - what falls due, and when - tells you the date on which the issuer must persuade somebody to lend again, which is where most defaults actually happen. A company does not usually fail because it had a poor quarter; it fails because it had a poor quarter in the month it needed to refinance.
The first mistake is comparing bonds on current yield. It is the number most often displayed on a screen, and it is the only one of the three that is neither a contractual fact nor a return. Two bonds with identical current yields can have yields to maturity a percentage point apart if one is at a discount and the other at a premium.
The second is treating a high yield as an opportunity that others have missed. A spread is a price set by people who read the same accounts you do. The question is never why is this yield so high - the market has an answer - it is whether you disagree with the answer, and on what evidence.
The third is buying a rating instead of a bond. The rating is an input, freely available, already in the price, and silent on everything except default likelihood. If your reason for buying is that it is rated investment grade, you have not appraised anything.
The fourth is underwriting a callable bond on its yield to maturity. The option belongs to the issuer, and the issuer exercises it when it suits them, which is exactly when it does not suit you. Use the yield to worst.
The fifth is reading covenants as boilerplate. They are the only part of the document that changes what the issuer may do with your money, and the difference between a covenanted and an uncovenanted issue from the same borrower is a real difference in what you own.
Coupon rate, current yield and yield to maturity, from the price and the schedule. Their ordering tells you immediately whether the bond is at a discount or a premium, and the identity of the one on the seller's sheet tells you something too. If the figure presented is the current yield on a discount bond, it is understating the return; if it is the current yield on a premium bond, it is overstating it.
There is no formula. Try rates until the discounted schedule matches the price. Doing it once for a bond you are being offered is cheap insurance, and it settles a question that is otherwise unanswerable from the sheet: whether the number in front of you reconciles to the price in front of you at all.
Seniority, security, guarantees, and where this issue sits relative to the borrower's other debt. Two bonds from the same issuer can differ more from each other than either differs from a bond of a different company. The issue-level documentation, not the issuer-level summary, is where this lives.
Same currency, same remaining maturity, as close as you can get. The difference is what you are being paid for taking this issuer instead of the sovereign. A spread compared against the wrong maturity mixes credit risk with curve position and produces a figure that means nothing.
Divide the spread by one minus your recovery assumption for the annual figure, then compound it out over the remaining life. State the recovery assumption explicitly, because the answer moves a long way with it - 24.72% cumulative at a 40% recovery becomes 20.24% at 25%. Then answer the only question that matters: is the real probability higher or lower than that, and why do you think so?
Interest cover and net debt against earnings for the current position, free cash flow to check that interest is being paid out of operations, and then the maturity schedule. Most defaults happen at a refinancing date rather than at a bad result, so a borrower with comfortable cover and a large maturity falling due in a closed market is more fragile than the ratios suggest.
Calls, puts, sinking funds, coupon step-ups, make-whole provisions. Compute the yield under each scenario and use the lowest. On Bond D that means underwriting at 3.7718% rather than 4.7735%, because the issuer holds the call and will use it when it suits them. Then read the covenants and decide whether they actually bind or merely describe.
For any individual bond you are considering buying directly, and as a periodic re-examination of an issue you already hold - the same sequence run against newer accounts and a newer price is what catches a deteriorating credit before the rating agency publishes.
Avoid this appraisal where you cannot obtain the issue-level documentation, avoid it where the borrower's accounts are not intelligible to you, and avoid using it to justify a purchase you have already decided on - a break-even default rate you compute after choosing the bond will be met with whatever recovery assumption makes it look acceptable.
A pattern that recurs whenever a large issuer approaches the investment grade boundary is that the market moves ahead of the agency and then moves again when the agency acts. The first move is analysts and traders repricing the credit as the accounts deteriorate. The second is mechanical: funds whose mandates permit only investment grade holdings become forced sellers on the day of the downgrade, regardless of what they think of the bond, and the resulting supply moves the price further than the change in default probability alone would justify. Holders sometimes read the second move as new information about the business. Usually it is information about the rules governing other people's portfolios.
A downgrade is two events - a revised opinion and a forced sale - and only the first is about the borrower. Know which one you are watching before deciding the market has overreacted.
Callable bonds are frequently bought on their yield to maturity in a falling-rate environment, which is exactly the environment in which the call is most likely to be exercised. The holder receives a par-plus-a-little redemption, records a small gain, and then discovers that the replacement bond available with the returned cash yields materially less than the one just repaid. The disappointment is real and the outcome was contractual: the option was described in the documentation and paid for through the coupon on the day the bond was bought. What was missed was that the option is only ever exercised at the holder's expense, because an issuer refinancing more cheaply and a holder reinvesting more cheaply are the same event.
Underwrite a callable bond at its yield to worst. If that number is not enough, the higher one on the sheet is not a reason to buy.
Create a free account to read the rest and apply it to your own investing.