Why Dollar Duration Matters
A duration stated as a number of years describes proportional price sensitivity, but a risk manager often needs to know the corresponding dollar amount. A 1% price change has a very different dollar effect on a $100,000 position than on a $100 million portfolio. Dollar duration and DV01 connect a duration measure to a position’s market value and provide a convenient way to estimate the dollar effect of small yield changes.
The calculation is direct, but its units and sign are easy to mishandle. In particular, a basis point is not a 1% change: one basis point is 0.0001 when a yield is expressed as a decimal. The calculations below use a deterministic, local sensitivity model. They do not assign a probability distribution to yield changes.
Here, “duration” in the calculation means a duration defined on a modified or equivalent price-sensitivity basis—not an unadjusted measure that has not been converted to price sensitivity. If a problem supplies only an unqualified duration, establish which definition is intended before applying the formula. The calculations assume the stated duration is appropriate for the yield movement being considered.
From Market Value to a Dollar Estimate
The relationship between dollar duration and a small yield change follows from the local slope of market value with respect to yield. If market value is \(M(y)\), modified duration at the current yield is defined by the negative proportional slope, \(D_{\text{mod}}=-(1/M)(dM/dy)\). Multiplying both sides by \(M\) gives the dollar sensitivity. A first-order approximation then estimates the change for a small yield movement.
This is a first-order, or tangent-line, approximation: it uses the slope at the current yield and treats that slope as constant over the scenario. It is most appropriate for a relatively small yield change, with cash flows and the relevant yield relationship behaving smoothly. It is not an exact repricing formula. Larger changes, embedded options, or changes in the shape of the yield curve can make the estimate less accurate; those effects require methods beyond this calculation.
Use positive market value for a long position and negative market value for a short position when calculating a signed exposure.
Multiply market value by modified duration. The result is sensitivity to a yield change of 1.00, not to a one-basis-point change.
Multiply dollar duration by 0.0001, because one basis point is one hundredth of one percent.
Convert the yield movement to a decimal and apply \(\Delta M \approx -M D_{\text{mod}}\Delta y\). A yield increase generally reduces the value of a long fixed-income position; a yield decrease generally increases it.
Worked Example: One Long Position and an 18-Basis-Point Increase
A hypothetical insurer holds a fixed-cash-flow bond portfolio with market value \(M=\$2{,}400{,}000\) and modified duration \(D_{\text{mod}}=5.8\). The current yield is 4.00%. Estimate the dollar change if the relevant yield increases by 18 basis points. Assume a small parallel shift, no change in cash flows, and use the first-order duration approximation. The scenario is deterministic; no mortality assumption or yield probability distribution is used. Method: calculate dollar duration, convert it to DV01, and then scale by the yield change.
Dollar duration. Multiply market value by modified duration:
Thus, the approximate loss for a unit increase in yield is $13,920,000. This is a sensitivity amount per 1.00 change in decimal yield, not the loss for a one-basis-point move.
DV01. One basis point is 0.0001 in decimal form:
Scenario change. Convert 18 basis points to a decimal yield change: \(18 \times 0.0001=0.0018\). Since the yield rises, the signed change for this long position is negative:
The estimated market value after the move is \(\$2{,}400{,}000-\$25{,}056=\$2{,}374{,}944\). As a check, 18 times the $1,392 DV01 is $25,056. The result is an estimated loss of $25,056, not an exact repricing.
Common mistake avoided: treating 18 basis points as 0.18. That would produce a loss of \(\$13{,}920{,}000 \times 0.18=\$2{,}505{,}600\), which is 100 times the correct first-order estimate.
Aggregating Dollar Duration Across Positions
For a portfolio exposed to the same parallel yield movement, signed dollar durations can be added. Long positions contribute positive dollar duration; short positions contribute negative dollar duration. The resulting net exposure determines the first-order portfolio change. Adding only the absolute values instead measures gross sensitivity, not the net change under a common yield movement.
Worked Example: Netting Long and Short Bond Exposures
A hypothetical portfolio has three positions, all exposed to a parallel shift in the same yield: a $1,200,000 long position with modified duration 3.2; an $800,000 long position with modified duration 7.5; and a $500,000 short position with modified duration 4.0. The current yield is 3.60%. Estimate the portfolio change for a 12-basis-point increase. Assume the supplied durations are modified durations, with no cash-flow changes or yield probability distribution. Method: calculate each signed dollar duration, net them, and apply the decimal yield change.
First long position: \(\$1{,}200{,}000 \times 3.2=\$3{,}840{,}000\).
Second long position: \(\$800{,}000 \times 7.5=\$6{,}000{,}000\).
Short position: its signed market value is negative, so its dollar duration is \(-\$500{,}000 \times 4.0=-\$2{,}000{,}000\).
Net dollar duration:
The net DV01 is \(\$7{,}840{,}000 \times 0.0001=\$784\) per basis point. A 12-basis-point increase is \(\Delta y=12 \times 0.0001=0.0012\), so:
The portfolio is estimated to lose $9,408. The DV01 check gives the same result: \(12 \times \$784=\$9{,}408\). Common mistake avoided: adding the absolute DV01s, \(\$384+\$600+\$200=\$1{,}184\), and using that as net exposure. The short position offsets some of the long sensitivity, so the net DV01 is $784, not $1,184.
Using DV01 to Estimate a Hedge
DV01 also makes sensitivities comparable across positions. In a simplified hedge calculation, a position’s DV01 can be offset by taking an opposite position with a suitable DV01. This is a sensitivity match, not a guarantee that the positions will offset under every market movement: the calculation assumes the relevant yield changes affect both positions in the stated way and ignores differences in their behavior beyond the local approximation.
Worked Example: Estimating a Simple DV01 Hedge
A hypothetical long bond position has market value $18,000,000 and modified duration 6.1 at a current yield of 4.25%. A hypothetical hedge instrument has market value $1,000,000 per unit and modified duration 8.0, so its DV01 per unit must be calculated. Estimate the number of short hedge units that offsets the bond position’s DV01. Then estimate the residual change for a 15-basis-point yield increase if only a whole number of hedge units can be held. Assume both positions respond to the same parallel yield shift, with unchanged cash flows; the calculation is deterministic and uses a first-order approximation. Method: compare the positive DV01 of the long position with the DV01 per hedge unit.
Long position dollar duration and DV01:
Hedge-unit dollar duration and DV01:
A short hedge unit has negative signed DV01. The ideal number of units is the long DV01 divided by the DV01 per unit:
If only whole units are allowed, compare 13 and 14. Thirteen short units leave net DV01 \(=\$10{,}980-13(\$800)=\$580\). Fourteen short units leave net DV01 \(=\$10{,}980-14(\$800)=-\$220\). Fourteen units have the smaller absolute residual, so choose 14 under this criterion.
For a 15-basis-point increase, the residual signed change is:
The positive result means an estimated $3,300 gain: after the hedge, the portfolio is slightly short in net sensitivity, so rising yields increase its value under this model. Check: the unhedged long position loses \(15 \times \$10{,}980=\$164{,}700\), while 14 short units gain \(15 \times 14 \times \$800=\$168{,}000\). The net is a $3,300 gain. Common mistake avoided: treating the short hedge as a positive exposure or rounding the ideal hedge quantity without checking which whole number leaves the smaller absolute residual.
Interpretation and Common Traps
A positive DV01 is conventionally reported as a loss magnitude for a one-basis-point yield increase in a long fixed-income position. The signed value change has the opposite direction: the formula includes a minus sign. Keeping those two statements separate avoids confusion when reporting a scenario result. For example, a long position with DV01 of $500 has an estimated $500 loss for a one-basis-point increase and an estimated $500 gain for a one-basis-point decrease, under the same local linear approximation.
| Item | Meaning | Unit or sign check |
|---|---|---|
| Dollar duration | Market value multiplied by modified duration | Dollars per 1.00 decimal yield change |
| DV01 | Dollar duration multiplied by 0.0001 | Positive loss magnitude per basis point for a long position |
| Yield change | Scenario change in yield | Convert basis points to decimal form before multiplying |
| Signed value change | Negative dollar duration multiplied by decimal yield change | Long fixed-income position generally loses when yield rises |
A frequent unit error is to confuse a 1-basis-point change with a 1% change. One percent is 0.01, or 100 basis points; one basis point is 0.0001. Another is to calculate a correct DV01 but then multiply it by a yield change already expressed as a decimal. Use either dollar duration times decimal change, or DV01 times the number of basis points—never mix the two forms.
Finally, duration-based dollar estimates are not exact price changes. They use the sensitivity at the current market yield and approximate the price-yield relationship locally. For a sufficiently large movement, for an asset with embedded options, or for a nonparallel change in yields, this simple estimate may be materially different from a full repricing. Dollar duration and DV01 are useful for exposure measurement and rapid scenario estimates, but their assumptions should accompany the reported result.
Check Your Understanding
For each problem, use the stated duration as modified duration and the first-order sensitivity approximation.
- A hypothetical long bond position is worth $750,000 and has modified duration 2.4. Calculate its dollar duration and DV01. (Answer: Dollar duration = $1,800,000; DV01 = $180 per basis point.)
- A $4,500,000 long position has modified duration 4.6. Estimate its dollar change if yield rises by 7 basis points. (Answer: \(-\$14{,}490\), an estimated loss of $14,490.)
- A portfolio holds $2,000,000 long at modified duration 5.0 and $600,000 short at modified duration 3.5. Calculate net DV01 and estimate the change for a 9-basis-point rise in yield. (Answer: Net DV01 = $790 per basis point; estimated change = \(-\$7{,}110\).)
- A hypothetical position is reported as market value $12,500 thousand, with modified duration 6.2. Estimate the gain or loss for an 11-basis-point yield decrease. (Answer: Dollar duration = $77,500,000; DV01 = $7,750; estimated gain = $85,250.)
- A long position worth $3,750,000 has modified duration 4.8. A hedge unit has market value $1,000,000 and modified duration 2.75. Find the whole number of short hedge units that leaves the smaller absolute residual DV01, then estimate the residual change for an 8-basis-point yield increase. (Answer: Long DV01 = $1,800; hedge DV01 = $275 per unit; use 7 short units, leaving net DV01 of \(-\$125\); estimated change = \(+\$1{,}000\).)