Macaulay vs. Modified vs. Effective Duration: Bond Measures Explained

A bond’s maturity is the contractual date when principal is scheduled to be repaid. Duration is a model-based description of cash-flow timing or price sensitivity. It is not simply another name for maturity, and it does not promise when an investor will recover the purchase price.

The label alone is incomplete. Before using a reported duration, identify whether it is Macaulay, modified, effective, money duration, or another measure. Then check the bond or portfolio, valuation date, yield or curve input, compounding convention, and price basis. CFA Institute separates yield-based measures, which assume specified cash flows and a change in the bond’s own yield, from curve-based measures that reprice against a benchmark curve and can accommodate uncertain cash flows. {source:cfaYieldBasedBondDuration2026} {source:cfaCurveBasedRiskMeasures2026}

Macaulay duration is the present-value-weighted average time until a bond’s scheduled cash flows are received. For cash flow CF_t at time t, price P, and yield y, each weight is PV(CF_t) divided by P. The weights add to one when the price is the present value of all modeled cash flows. Duration is the sum of each time multiplied by its weight:

PV(CF_t) = CF_t / (1 + y/m)^(m×t)

D_Mac = Σ[t × PV(CF_t)] / P

The weighting explains why Macaulay duration is usually shorter than maturity for a coupon bond: some value is returned before the final principal payment. A zero-coupon bond has only one scheduled cash flow, so its Macaulay duration equals its time to maturity under the stated cash-flow and discounting assumptions.

This is a present-value average, not a forecast of the investor’s personal holding period or a guarantee that reinvested coupons will earn the same yield. It can be useful for describing timing and for certain immunization frameworks, but those frameworks require additional assumptions about yield moves, reinvestment, and the liability being matched. FINRA also cautions that the word “duration” is not just a time-to-maturity label. {source:finraDurationExplainer}

Consider a hypothetical option-free bond with $1,000 face value, a 5% annual coupon, three years remaining, and a 4% annual yield to maturity. Assume annual payments, valuation just after a coupon date, and annual compounding. The scheduled cash flows are $50 in year one, $50 in year two, and $1,050 in year three.

Discounting those payments at 4% gives present values of about $48.08, $46.23, and $933.45. Their sum is a full price of about $1,027.75. Divide each present value by that price to get weights of approximately 4.68%, 4.50%, and 90.82%. The larger final weight reflects the principal repayment, while the first two coupons still pull the average time earlier.

Multiplying each time by its weight gives 1 × 4.68% + 2 × 4.50% + 3 × 90.82%, or about 2.8615 years. The result is less than the three-year maturity because part of the bond’s present value comes from earlier coupons. Small differences in the last decimal can arise from rounding; the weights should be calculated from unrounded present values. {source:cfaYieldBasedBondDuration2026}

Modified duration adjusts Macaulay duration for the stated yield-compounding convention. For a plain fixed-cash-flow bond with nominal annual yield y compounded m times per year, the familiar relation is D_mod = D_Mac / (1 + y/m). With annual payments in the example, m is 1, so modified duration is 2.8615 / 1.04, or about 2.7514 years.

For a small change in the bond’s own yield, the first-order percentage price estimate is ΔP/P ≈ −D_mod × Δy, where Δy is a decimal yield change. A rise of 1 basis point is 0.0001, not 0.01; a rise of 100 basis points is 0.01. With modified duration 2.7514, a 1% yield rise gives a local estimate of a 2.7514% price decline, assuming other inputs stay fixed. The negative sign represents the inverse local relationship between yield and price for this plain-bond setup.

Modified duration is a slope near the starting yield, not a full repricing for any-sized shock. A dollar estimate also needs the position’s market value: a common first-order DV01 magnitude is market value × modified duration × 0.0001 for a one-basis-point own-yield move. Sign conventions vary, and the result is not a fixed loss limit. CME distinguishes this percentage sensitivity from the dollar effect reported as DV01, BPV, or VBP. {source:cfaYieldBasedBondDuration2026} {source:cmeTreasuryRiskMeasurement}

Effective duration uses prices from a full valuation under rate scenarios. Let P₀ be the starting price, P₋ the price after a specified benchmark-curve decrease of Δr, and P₊ the price after the same-sized increase. A central-difference estimate is D_eff = (P₋ − P₊) / (2 × P₀ × Δr), with Δr expressed as a decimal. The result is commonly interpreted as a percentage-price sensitivity in years for the defined curve shock.

The method is useful when future cash flows may change as rates move, as with callable or putable bonds and mortgage-backed securities. Instead of holding projected cash flows fixed and changing only one yield input, the valuation model can update expected exercise or prepayment behavior and then reprice. The reported number therefore depends on the benchmark curve, shock size and shape, model, volatility assumptions, cash-flow assumptions, and valuation date. CFA Institute identifies effective duration as a curve-based measure for securities with uncertain cash flows; FINRA likewise distinguishes it from Macaulay and modified duration for bonds with redemption features. {source:cfaCurveBasedRiskMeasures2026} {source:finraDurationExplainer}

Effective duration is not automatically “more accurate” in every setting. It answers a different question: how the modeled full price changes under a specified benchmark-curve scenario. If the instrument has stable, known cash flows and the question concerns its own yield, modified duration may be the more direct measure. For a curve move, state whether the shift is parallel and which curve is used; for an option-sensitive bond, state the model that recalculates the cash flows.

Return to the $1,000-face example. At 4% yield its price is about $1,027.75 and modified duration is about 2.7514. The local estimate for a 100-basis-point rise is a decline of about 2.7514%, or $28.28, to roughly $999.47. The local estimate for a 100-basis-point fall is a gain of about 2.7514%, or $28.28, to roughly $1,056.03.

Now reprice the scheduled cash flows rather than using the straight-line estimate. At 5% yield, the bond is worth $1,000, a decline of about $27.75 or 2.70%. At 3%, it is worth about $1,056.57, a gain of about $28.82 or 2.80%. These exact cash-flow values are asymmetric around the starting price because the price-yield relationship is curved. The duration-only estimate is useful locally, while convexity helps explain the second-order difference for a larger move. The arithmetic is hypothetical and assumes unchanged cash flows, annual compounding, and the same valuation convention; it is not a market quote or forecast. See bond convexity and price sensitivity for the next-order adjustment.

The example’s D_Mac and D_mod are yield-based measures; it does not calculate effective duration because no benchmark curve or rate-bump valuation model was specified. To obtain effective duration, an analyst must revalue the bond under the defined benchmark-curve scenarios. If cash flows change under those scenarios, the model must apply the same exercise and prepayment assumptions consistently to the up and down valuations.

Use Macaulay duration when the question concerns the present-value timing of specified cash flows. Use modified duration for a first-order percentage-price response to a small change in the bond’s own yield under a stated compounding convention. Use effective duration when the requested scenario shifts a benchmark curve and the instrument may have rate-dependent cash flows. None tells you which way rates will move.

For a portfolio, aggregation usually depends on the market-value weights and the provider’s conventions. A single effective-duration number can summarize a defined broad curve shift but can hide twists and bends; key-rate duration maps sensitivity to selected curve points. When you need dollars per basis point, DV01 and BPV explain the currency measure. Credit-spread exposure, liquidity, inflation, currency, and optionality may require separate analysis; benchmark-rate duration does not measure all bond risk.

Bond features also matter. All else equal in the standard option-free fixed-rate setup, later cash flows tend to give more duration, while a higher coupon or yield tends to bring more present value toward earlier payments and reduce duration. As a bond ages, its remaining cash-flow profile changes, so its duration is not fixed. A call option can limit price gains when rates fall; a floating-rate note can reset its coupon periodically and may have less rate sensitivity than a fixed-rate bond of similar maturity, although credit spread, reset lag, floors, and liquidity remain. Treat these as mechanisms under stated assumptions, not universal rankings. {source:cfaYieldBasedBondDuration2026} {source:finraDurationExplainer}

Two sources can call a figure “duration” while measuring different things. Record the instrument or portfolio, valuation date, full or clean price basis, yield or benchmark curve, compounding frequency, bump size, parallel or nonparallel shock, option model, and whether cash flows are held fixed or recalculated. Match the position size and currency before comparing dollar measures. If any of those inputs differ, explain the difference before treating the numbers as comparable.

Also separate price sensitivity from expected total return. Duration does not include every source of realized return, such as coupon income, reinvestment, credit migration, spread changes, transaction costs, taxes, or a changing portfolio. It is a conditional sensitivity based on a defined starting point and model. Read a fund fact sheet or analytics screen for its methodology and as-of date, then consult the prospectus or security documentation for contractual features. Do not infer that a low duration means low overall risk or that a quoted duration forecasts a future loss. {source:finraDurationExplainer} {source:cmeTreasuryRiskMeasurement}