Skip to content
Financial Context Wiki

Yield Curve: Par, Zero, Forward, Shape, and Valuation Mapping

Define the exact U.S. Treasury curve object, normalize quote conventions, distinguish par, zero, and forward rates, interpret curve dynamics without causal overreach, and map the right curve to securities and valuation.

Updated

For educational purposes only; not investment advice. Investing may result in loss.

Direct answer

A yield curve maps a specified yield or discount measure to maturity for a defined issuer, instrument set, currency, date, time, side of market, and quote convention. There is no single universal U.S. Treasury curve. The official nominal Constant Maturity Treasury, or CMT, series is a par yield curve; a zero-coupon curve maps maturity to spot yields or discount factors; a forward curve is derived from discount factors; the real Treasury curve is based on TIPS; and fitted research or term-premium curves are separate model products.

Treasury currently derives its official nominal par curve from indicative bid-side prices, not transactions, on the most recently auctioned nominal bills, notes, and bonds observed near 3:30 p.m. Eastern each trading day. It converts inputs to yields, bootstraps instantaneous forward rates at input maturities, and applies monotone-convex interpolation in forward-rate space. Fixed-maturity CMT points are theoretical par yields read from that curve and may not equal the yield on any security available to buy. Treasury can change inputs or methodology, does not publish the complete derivation program, and floors derived nominal CMT rates at zero even when some input securities have negative yields.

Par, zero, and forward rates answer different questions. A par yield is the coupon rate that prices a hypothetical coupon security at par under a stated convention. A zero yield discounts one cash flow at one maturity. A forward rate is the no-arbitrage rate implied between future dates by today’s discount curve; it is not a pure forecast. A term spread such as spread10y−2y = y10y − y2y describes two selected points, not the entire curve, and its sign does not identify one economic cause.

Seven-step yield-curve workflow

  1. Define the object and decision. State issuer, currency, nominal or real claim, instrument universe, par or zero or forward object, bid or ask or midpoint, date, time, timezone, frequency, source, vintage, and intended use. A macro signal, an index reference, a hedge, and a valuation curve can require different objects.
  2. Normalize instruments and quote conventions. Distinguish Treasury bills from coupon notes and bonds, TIPS, and floating-rate notes. Convert bank discount, investment yield, bond-equivalent yield, effective annual yield, clean price, accrued interest, and cash price consistently; fix settlement, coupon frequency, day count, and compounding before comparing rates.
  3. Construct discount factors. Map each security to dated cash flows, bootstrap discount factors or zero rates from observed prices, and document treatment of bills, coupon dates, reopenings, taxes, liquidity, special repo value, and excluded securities. Choose interpolation and extrapolation deliberately; the published CMT points alone are not Treasury’s complete internal discount curve.
  4. Derive and validate curve representations. From one discount function, derive zero, par, and forward curves under matching conventions. Reprice every input security, inspect residuals, discount-factor monotonicity, implausible forward oscillations, boundary behavior, and sensitivity to knots. Keep the official Treasury CMT curve distinct from Federal Reserve staff research curves and market-vendor curves.
  5. Describe level, slope, curvature, and movement. Record several maturity points, named term spreads, a curvature measure, and key-rate changes. Classify bull steepening, bull flattening, bear steepening, or bear flattening only after reporting which yields moved and by how much; one overall label can hide opposing moves in different segments.
  6. Separate arithmetic from economic attribution. Decompose a fitted zero yield, within one model and vintage, into expected future short rates and an estimated term premium rather than applying that identity mechanically to a par CMT. For nominal-versus-real comparisons, separate real rates, expected inflation, inflation-risk premium, liquidity, TIPS indexation lag and floor, and model effects. Treat recession probabilities and causal stories as model-dependent evidence, not deterministic timing rules.
  7. Map the curve to the claim and stress it. Discount a Treasury security’s dated cash flows with matching zero rates; add credit, liquidity, option, collateral, and servicing components for other claims. Map company debt to currency, maturity, fixed or floating terms, refinancing date, and issuer spread; map equities to claim-matched cash flows and required returns. Stress parallel, steepening, flattening, curvature, and key-rate shocks rather than moving every discount rate mechanically with one CMT point.

Worked examples

  • Treasury bill quote conversion. A 91-day bill has face value $100 and a bank discount rate of 5.0000%. Under the stated 360-day discount convention, P = $100 × [1 − 5.0000% × 91 ÷ 360] = $98.736111. The simple investment yield on price and a 365-day basis is investment yield = ($100 − $98.736111) ÷ $98.736111 × 365 ÷ 91 = 5.134337%. If the same holding-period return could be reinvested for a year, the effective annual rate would be EAR = ($100 ÷ $98.736111)^(365 ÷ 91) − 1 = 5.234134%. These are three conventions for one assumed cash flow, not three market outcomes; CMT values are reported on a bond-equivalent basis rather than as the bank discount rate or EAR.
  • Bootstrap par to zero and derive a forward. Assume a simplified semiannual curve with a six-month par bond-equivalent yield of 4.0000%. Because there is one maturity cash flow, DF₀.₅ = 1 ÷ (1 + 4.0000% ÷ 2) = 0.980392157. A one-year par yield of 4.5000% implies two coupon payments of $2.25 per $100 face, so DF₁ = [$100 − $2.25 × 0.980392157] ÷ $102.25 = 0.956421688 and z₁ = 2 × [0.956421688^(−1 ÷ 2) − 1] = 4.505639%. The six-month rate beginning six months from now is therefore f₀.₅,₀.₅ = 2 × (0.980392157 ÷ 0.956421688 − 1) = 5.012531%. The check $2.25 × 0.980392157 + $102.25 × 0.956421688 = $100.000000 reprices the one-year par security. Substituting par yields directly into a forward formula would mix curve objects.
  • The same uninversion from opposite shocks. Initially the two-year yield is 5.00% and the ten-year yield is 4.20%, so spread10y−2y = 4.20% − 5.00% = −80 bp. In a bull steepening, the two-year falls 120 bp to 3.80% and the ten-year falls 30 bp to 3.90%, producing 3.90% − 3.80% = +10 bp; the slope changes by +90 bp. In a bear steepening, the two-year stays at 5.00% while the ten-year rises 90 bp to 5.10%, also producing 5.10% − 5.00% = +10 bp and the same +90 bp slope change. The final spread sign is identical, but the rate level, duration shock, inflation narrative, financing pressure, and likely cash-flow context are not.
  • Discount each cash flow with the matching zero rate. A two-year Treasury note has $100 face value, a 4.0000% annual coupon, and semiannual payments. Assume six-, twelve-, eighteen-, and twenty-four-month zero bond-equivalent yields of 3.0000%, 3.5000%, 4.5000%, and 5.5000%. Curve-consistent value is P = $2 ÷ (1 + 3.0000% ÷ 2)^1 + $2 ÷ (1 + 3.5000% ÷ 2)^2 + $2 ÷ (1 + 4.5000% ÷ 2)^3 + $102 ÷ (1 + 5.5000% ÷ 2)^4 = $97.283998. Using the two-year rate for every cash flow instead gives Pflat = $2 ÷ (1 + 5.5000% ÷ 2)^1 + $2 ÷ (1 + 5.5000% ÷ 2)^2 + $2 ÷ (1 + 5.5000% ÷ 2)^3 + $102 ÷ (1 + 5.5000% ÷ 2)^4 = $97.195429, an error of $97.283998 − $97.195429 = $0.088569. The error generally grows with curve shape, maturity, and interim cash flows.

Risks and validation controls

  • Name issuer, currency, nominal or real claim, curve type, maturity points, side of market, source, date, time, timezone, frequency, and vintage.
  • Treat a CMT point as an interpolated theoretical par yield, not automatically the yield or executable price of a security.
  • Distinguish Treasury’s official monotone-convex par curve from H.15 averages, Federal Reserve staff research curves, vendor curves, and fitted term-premium models.
  • Convert bill bank discount rates, investment yields, bond-equivalent yields, and effective annual yields before comparison.
  • Reconcile clean price, accrued interest, cash price, settlement date, coupon schedule, day count, and compounding.
  • Do not substitute par yields for zero yields or use either as a forward rate without a consistent discount-function derivation.
  • Document bootstrap order, instrument inclusion, knot dates, interpolation, extrapolation, rounding, and treatment of missing or anomalous observations.
  • Account for Treasury’s zero floor on derived nominal CMT rates when studying low-rate or negative-yield episodes.
  • Freeze data vintage and methodology; historical values or fitted parameters can differ after revisions or methodology changes.
  • Name both endpoints and subtraction order for every term spread; 10-year-minus-2-year and 10-year-minus-3-month are different series.
  • Track absolute levels, several slopes, curvature, and key-rate changes; one spread cannot represent the full curve.
  • Use bull or bear and steepening or flattening labels only after calculating the underlying yield changes.
  • Treat an implied forward as a no-arbitrage break-even conditional on the curve, not as an unbiased forecast or promised future rate.
  • Treat expected-short-rate and term-premium components as model estimates with parameter, sample, vintage, and specification uncertainty.
  • Match nominal and real curve types before computing breakeven inflation and preserve inflation-risk, liquidity, indexation-lag, floor, and technical wedges.
  • Predefine recession spread, sampling, horizon, recession definition, threshold, vintage, and decision rule before testing predictive performance.
  • For banks, reconcile earning-asset yields, funding volumes and costs, deposit beta, repricing gaps, hedges, liquidity, and credit losses rather than using one Treasury spread.
  • For corporate debt, map each maturity, currency, fixed or floating rate, refinancing date, covenant, collateral, credit spread, and embedded option.
  • For equities, separate cash-flow revisions from discount-rate revisions and use a claim-, currency-, inflation-, and horizon-matched required return.
  • For futures, swaps, mortgages, ETFs, and options, incorporate contract delivery, collateral, basis, roll, convexity, prepayment, tracking, and volatility features rather than treating CMT as the product return.

Common misconceptions

  • “The U.S. Treasury yield curve is one unambiguous dataset.” Official par CMT, real par, bill, zero, forward, fitted research, and term-premium curves are different objects.
  • “A published CMT rate is a Treasury security I can buy.” It can be an interpolated theoretical par yield derived from indicative bid quotations.
  • “Par, zero, and forward yields are interchangeable at the same maturity.” Coupon structure and discounting make them different even under one internally consistent curve.
  • “An implied forward or inverted spread is a deterministic forecast.” Forward rates contain risk premia, and historical recession relationships are conditional model evidence rather than guarantees or clocks.
  • “Steepening is always bullish or always helps banks.” It can arise from falling short rates or rising long rates, while funding structure, hedges, credit, liquidity, and cash-flow expectations can dominate.

Authoritative sources

Navigation

Search the wiki...