TL;DR. Duration is the single number that tells you how much a bond’s price will move when interest rates change. A bond with a duration of 10 falls roughly 10% in price if yields rise 1%, and roughly 16% if its duration is 16. That math is why the 30-year Treasury has battered investors while T-bills have barely moved, and why “safe” long-dated fixed income behaves nothing like cash when the Fed shifts.
The mechanic: bond prices and yields move opposite each other
A bond is a contract to receive a fixed stream of coupons plus the return of face value at maturity. When you buy one in the secondary market, you cannot change those cash flows — but you can set the price. If interest rates on similar new bonds rise, the existing bond’s fixed coupons become less attractive, and its price must fall until its yield-to-maturity matches the new market level. The opposite happens when rates fall: the existing coupon looks rich, so the price rises.
That inverse relationship is the bedrock fact of fixed income. The interesting question isn’t whether prices move — it’s how much they move per unit of rate change. That is what duration answers.
Duration: one number that measures the pain (or the gain)
FINRA puts it cleanly: “Bond duration is a measure of the degree to which a bond investment is likely to change in value if interest rates were to rise or fall” (FINRA Investor Insights). The rule of thumb the same page gives:
For every 1 percentage-point change in interest rates, a bond will rise or fall in the opposite direction by an amount equal to its duration number.
- Duration 2 → roughly ±2% per 100 basis points of rate move.
- Duration 10 → roughly ±10% per 100 basis points.
- Duration 16 → roughly ±16% per 100 basis points.
That last one is not a typo. A long-duration bond can lose roughly a sixth of its market value in a single year if its yield climbs by one percentage point — even though no coupon was missed and no default occurred. It is a mark-to-market reality of holding fixed income with a long stream of distant cash flows.
A worked example with real numbers
Suppose you own a 10-year Treasury note with a 4% coupon, priced today at par ($1,000) to yield 4%. Its modified duration is roughly 8.2 years.
Now the 10-year yield rises from 4% to 5% — a 100 basis-point move. The duration rule estimates:
Estimated price change ≈ −8.2 × 1.00% = −8.2%
The bond’s price falls from $1,000 to roughly $918. You still collect the $40 coupon each year, but if you sell now you have locked in an 8.2% capital loss. If instead rates fell by 100 bps to 3%, the same bond would gain roughly 8.2%, landing near $1,082. Duration is symmetric in its first-order estimate — and it is also imperfect, which is where convexity comes in.
Where duration comes from: Macaulay and modified
The version most cited in finance textbooks is Macaulay duration — the weighted-average time until you receive the bond’s cash flows, with the present value of each cash flow as the weight. It comes out in years.
Modified duration is Macaulay duration divided by (1 + yield per period). It converts the weighted-average time into a price-sensitivity number you can multiply directly by a yield change to estimate the percent price move. Modified duration is what FINRA’s rule of thumb implicitly uses.
Three things drive duration higher:
- Longer maturity. A 30-year bond has a much longer Macaulay duration than a 2-year bond, all else equal.
- Lower coupon. A zero-coupon bond has duration equal to its maturity — every dollar comes back at the end. Higher coupons pull duration shorter because you recover principal sooner via cash payments.
- Lower yield. As yields fall, the present value of distant cash flows rises faster than near-term ones, lengthening duration.
Convexity: the second-order correction
Duration assumes the price-yield relationship is a straight line. It isn’t — it’s a curve. The actual price-yield curve bows upward, which means:
- When yields fall, the bond’s price rises more than duration predicts.
- When yields rise, the bond’s price falls less than duration predicts.
This curvature is captured by convexity, a second-derivative measure. The better full-formula estimate is:
Price change ≈ −ModDur × Δy + ½ × Convexity × (Δy)2
For small yield moves (≤25 bps), duration alone is fine. For larger moves — like the +100 bps used in the example above — convexity matters. Convexity is “good” for a bondholder: the actual loss is somewhat less than duration estimates, and the actual gain somewhat more. Long-duration zero-coupon bonds have the highest convexity, which is one reason institutional managers sometimes pay up for them.
Duration across today’s Treasury curve
Modified durations vary sharply across the curve. Using U.S. Treasury yields as of June 9, 2026 (from the Federal Reserve’s H.15 release):
| U.S. Treasury maturity | Yield (Jun 9, 2026) | Approx. modified duration | Est. price impact of +100 bps |
|---|---|---|---|
| 3-month T-bill | 3.79% | 0.25 | −0.3% |
| 2-year note | 4.13% | 1.9 | −1.9% |
| 5-year note | 4.26% | 4.5 | −4.5% |
| 10-year note | 4.53% | 8.4 | −8.4% |
| 20-year bond | 5.02% | 12.7 | −12.7% |
| 30-year bond | 5.01% | 15.8 | −15.8% |
That is the answer to the question on every income investor’s mind right now: a 30-year Treasury can lose roughly $158 per $1,000 face if yields rise 1 percentage point. A 3-month T-bill loses about $3. Same asset class, completely different risk profile.
Common mistakes
1. Treating duration as a static number
As yields change and the bond ages, duration changes too. The 10-year bond you bought last year is now a 9-year bond with a shorter duration. Re-check periodically — especially for bond funds, where the published duration is a snapshot.
2. Confusing duration with maturity
A 30-year zero-coupon Treasury has a duration of 30. A 30-year Treasury with a 5% coupon has a duration closer to 16. They share a maturity but live in completely different risk worlds.
3. Ignoring convexity on large moves
The duration approximation can overstate losses by a meaningful amount when yields move 100+ bps. For institutional risk management and any large position, convexity-adjusted numbers matter.
4. Assuming a bond fund’s duration is its average maturity
It isn’t. Fund duration is the weighted-average duration of holdings, which is shorter and more sensitive to the coupon distribution than the maturity ladder suggests. Always look at the published effective duration on the fact sheet.
5. Forgetting that duration cuts both ways
Long duration is painful in a rising-rate environment and a tailwind in a falling-rate one. The question isn’t “is duration good or bad” — it is “do you want this exposure given where rates are going next.”
Related concepts and what to learn next
- Yield curve and term premium. Duration tells you sensitivity to one yield; the curve tells you the full term structure. (See our Yield Curve Explained and Term Premium Explained pieces.)
- Credit spreads and spread duration. For corporate bonds, both rates and spreads matter. Spread duration measures sensitivity to credit spread changes, separate from rate duration.
- Effective duration. For bonds with embedded options (callables, mortgage-backed securities), modified duration breaks down because cash flows can change. Effective duration handles it.
- Key-rate duration. A single duration number masks where on the curve a portfolio is sensitive. Key-rate duration decomposes that exposure into 2-year, 5-year, 10-year, and 30-year buckets.
Sources
- FINRA. “Brush Up on Bonds: Interest Rate Changes and Duration.” finra.org
- Federal Reserve. “H.15 Selected Interest Rates,” release for June 9, 2026. federalreserve.gov
- U.S. Securities and Exchange Commission. Investor education on bonds and interest rate risk. investor.gov
- U.S. Department of the Treasury. Daily Treasury Par Yield Curve Rates. home.treasury.gov
Disclosure: This article was produced with AI assistance and reviewed before publication. It is for informational purposes only and is not investment advice.