Fixed Income Essentials — CFA Level I

The exam rarely asks “what is a bond” — it tests whether you can map a described feature onto the right bond indenture element (the legal contract between issuer and holder). Read the vignette for the three legal pillars: covenants (affirmative = the issuer must do, e.g. insure assets or pay on time; negative = must not, e.g. exceed a leverage ratio or take on more debt), the source of repayment, and collateral/seniority. A favorite item asks who absorbs losses first — the answer hinges on the priority waterfall: secured before unsecured, senior before subordinated, with equity holders last.

The classic trap is conflating the bond (the instrument and its full contract) with its narrower features. A bond is not its coupon (the periodic cash payment) and not its maturity (the principal-repayment date); both are merely terms inside the indenture. Students also wrongly assume par value equals price — par is the fixed redemption amount, while market price is the present value of future cash flows and can trade at a premium or discount. Hook: an indenture is the bond’s rulebook; coupon and maturity are just two clauses in it.

The exam loves to make you separate the fixed coupon rate (set at issuance on the face value) from the market yield that moves daily — a floating-rate note resets its coupon to a reference rate plus a quoted (fixed) spread each period, so its coupon does change while a plain-vanilla bond’s never does. The classic vignette gives a coupon plus a price (discount, par, or premium) and asks for the relationship to yield: because price and yield move inversely, a discount price always implies yield-to-maturity above the coupon. Watch the day-count trap too — US Treasury accrued interest uses actual/actual, but US corporate and municipal bonds use 30/360.

A frequent mistake is conflating coupon with current yield (annual coupon ÷ price) or with YTM; only at par do coupon rate, current yield, and YTM coincide (at a discount, current yield sits between the two). Don’t mistake a zero-coupon bond’s lack of payments for no return — its implied interest accretes as the price pulls to par. Hook: discount, yield’s up; premium, yield’s down.

Money-market instruments mature within one year; the note-versus-bond label is a loose market convention (US Treasury notes run 2–10 years, Treasury bonds 20–30 years). Perpetual bonds (consols) have no maturity at all. Term to maturity declines mechanically over a bond’s life — a 10-year bond becomes a 9-year bond a year after issuance.

The exam loves to make you pick the right duration for the right bond: when an item mentions a callable or putable bond, the only valid measure is effective duration, because the embedded option lets cash flows change with yield — choosing Macaulay or modified there is the planted trap. A classic computation gives you Macaulay and asks for modified duration: divide by (1 + periodic yield), so modified is always slightly smaller than Macaulay (under discrete compounding). Another favorite: a zero-coupon bond’s Macaulay duration equals its maturity, while any coupon bond’s duration is shorter.

Don’t confuse duration with maturity — maturity is a fixed calendar date, while duration weights all cash flows, so a higher coupon and a higher yield both lower duration. The other classic error is treating the linear duration estimate as exact; for large yield jumps it overstates the price drop (and understates the gain), which is precisely why convexity corrects it. Memory hook: duration is the slope, convexity is the bend — slope alone always misses the curve.

The exam rarely asks you to compute convexity; it tests ranking and direction. A classic item gives two bonds with identical (or nearly identical) duration and asks which is preferable — the higher-convexity bond wins, because for equal-sized yield moves it gains more when yields fall and loses less when they rise. The tell that convexity matters is a large yield move or the phrase “more accurate estimate,” signaling you must add the ½ × convexity × (ΔY)² term, which is positive for option-free bonds and therefore improves the duration-only estimate in both directions.

The trap is treating convexity as a substitute for duration: it is a second-order correction, so apply the duration (first-order) term first, then adjust. Duration is the slope; convexity is the curvature. Also distinguish effective convexity — which reprices the bond at shifted (higher and lower) yields and is the required measure for bonds with embedded options — from yield-based (approximate) convexity, and remember a callable bond can turn negative-convex when low yields make the call likely.

The exam’s favorite tell is which spread to use for a bond with an embedded option: OAS is the correct relative-value measure for callable/putable bonds because it strips out the option’s value, whereas the Z-spread does not. A classic item gives a callable bond’s Z-spread and option cost and asks for OAS — remember OAS = Z-spread − option cost for a callable (the issuer’s call has value, so OAS < Z-spread), while for a putable bond the option benefits the investor, so OAS > Z-spread. The trap is treating Z-spread and OAS as interchangeable: on an option-free bond they are equal, but only then.

Don’t reduce a spread to default risk alone — a credit spread compensates for credit and liquidity (and tax effects), so a wide spread isn’t purely default probability. Keep spread distinct from the treasury benchmark yield itself: the spread is the difference, whether measured over a single government bond (G-spread) or the swap curve (I-spread). Note the Z-spread is a constant add-on to every spot rate, not a single-point spread.

Exam items almost always force you to separate default probability from default loss: a question gives recovery data and asks for loss given default = 1 − recovery rate, or hands you PD, LGD, and EAD and wants expected loss — the trap is forgetting that higher seniority lowers LGD, not the probability of default (the curriculum treats PD as the same for an issuer and its issues, so seniority only shifts recovery, which is why ratings get “notched”). A second classic pattern asks which event constitutes default: missing a coupon or principal payment qualifies, and so can breaching a covenant — but a covenant breach is typically a technical default that can be cured or waived before it becomes a payment default.

Don’t conflate the pieces with their neighbors. Spread is the market price of this risk (compensation for expected loss plus a risk premium), so widening spreads signal rising default expectations — but spread also embeds liquidity and tax components, not just credit. The indenture defines what counts as default and the remedies; weaker (covenant-lite) protections raise loss severity, not default frequency. Memory hook: PD is whether, LGD is how bad — seniority only moves the second.

Exam items lean on one identity: callable bond value = value of an otherwise-identical straight (option-free) bond − value of the embedded call option. The classic vignette gives you the straight-bond price and the call-option value and asks for the callable price, or flips it to back out the embedded option — recognize that the option is subtracted because it is held by the issuer. A second favorite: when interest-rate volatility rises, the call option gets more valuable, so the callable bond’s price falls (a putable bond’s price would rise, since the put adds value). The “tell” that an answer hinges on call risk is any mention of rising volatility or the call moving toward the money.

The trap is confusing call risk with convexity mechanics. Callable bonds show negative convexity only at low yields, where the call is near/in-the-money; at high yields the call is far out-of-the-money and they behave like straight bonds with normal positive convexity. Don’t equate “callable” with “putable” — the put protects the holder and adds value for them. Memory hook: whoever owns the option gains, so the issuer’s call is subtracted.

Items rarely ask “what is an indenture”; they hand you a scenario and make you classify a clause, and the tell is a verb. If the issuer is doing something — “shall file audited statements,” “must keep a current ratio above 1.5,” “maintain insurance,” “pay taxes” — it’s affirmative, even though that last pair looks like a rule. If the clause forbids an action — “may not incur additional debt,” “no asset sales,” “restricted payments” — it’s negative. The hinge: negative covenants protect bondholders by constraining the issuer, so tighter covenants lower required yield while covenant-lite structures (a leveraged-loan/high-yield phenomenon) widen spreads. The curveball is that covenants aren’t free — they cost the issuer flexibility, traded for a lower coupon.

Don’t conflate the indenture with default: the indenture defines what counts as an event of default and the trustee’s remedies (acceleration, enforcement), but default is the breach itself. Likewise a call provision isn’t a separate document — the redemption schedule and call price are written into the indenture.

The Treasury yield curve is the benchmark for pricing all other USD fixed-income securities. Its shape — normal (upward-sloping), flat, or inverted — carries information about expected growth, inflation, and monetary policy.

The exam’s favorite move is separating the par, spot, and forward curves built from Treasuries: a par rate is the YTM of a coupon bond priced at par, spot rates discount single cash flows (so a bond prices correctly as a portfolio of zeros), and forward rates are implied future single-period rates. Bootstrapping solves for each later spot rate sequentially from the shorter spot rates already found — never one flat YTM applied to every cash flow. Distinguish on-the-run (most recently auctioned, most liquid, the benchmark) from off-the-run issues. A corporate bond’s yield equals the matched-maturity Treasury yield plus a spread, with the G-spread quoting that gap directly over the government bond. Don’t call Treasuries truly risk-free — they carry interest-rate/duration and inflation risk; only their default risk is near-zero. Memory hook: “on-the-run runs the benchmark.”