Skip to content

Duration Times Spread (DTS)

Multiply spread duration by the current credit spread to compare bond price sensitivity to the same proportional spread shock.

Duration Times Spread (DTS) multiplies a bond’s spread duration by its current credit spread. It lets a portfolio manager compare the price sensitivity of a shorter-duration bond with a wide spread and a longer-duration bond with a tight spread when both spreads move by the same percentage.

Which shock are you measuring?

Let DD be spread duration and ss the starting annual spread in decimal form. Hold the benchmark curve fixed and approximate the price effect of a small spread change Δs\Delta s using the starting duration. For a positive spread,

ΔPP≈−D Δs=−Ds⏟DTSΔss.\frac{\Delta P}{P} \approx -D\,\Delta s = -\underbrace{Ds}_{\mathrm{DTS}}\frac{\Delta s}{s}.
The two expressions describe the same first-order price move. Spread duration is the sensitivity to an absolute spread change, such as +20 bp. DTS is the sensitivity to a proportional change, such as +10% of the starting spread. Ben Dor and co-authors found that wider-spread US corporate bonds in their historical sample tended to experience larger spread moves in basis points. This made DTS a useful proxy for spread-driven return volatility in that sample. Ben Dor et al. (2007)

Two bonds, two spread shocks

Take two fictional bonds with option-adjusted spreads (OAS) calculated on the same benchmark curve, at the same time and under the same model conventions. Each spread duration measures sensitivity to its own OAS, with the benchmark and other model inputs held fixed.

BondSpread durationStarting OASDTS (bp-years)
A4 years150 bp600
B3 years200 bp600
One basis point is 0.00010.0001 in decimal form. Duration in years times a decimal annual spread gives a dimensionless sensitivity. The display value of 600 bp-years is 600/10,000=0.06600/10{,}000=0.06 on that scale.
A 10% relative widening takes A from 150 to 165 bp and B from 200 to 220 bp. The longer duration on A offsets its smaller 15 bp move. At first order, both prices fall by 0.06×0.10=0.6%0.06 \times 0.10=0.6\%. Now apply the same +20 bp to both. B’s price still falls by about 0.6%, while A’s falls by about 0.8% because each basis point has a larger effect on A’s price.
Fictional Bond A has 4 years of spread duration at 150 bp and Bond B has 3 years at 200 bp. With 10% relative widening their spreads rise by 15 and 20 bp and both prices fall by 0.6%. With a 20 bp widening for both, A loses 0.8% and B loses 0.6% at first order.
Fictional Bond A has 4 years of spread duration at 150 bp and Bond B has 3 years at 200 bp. With 10% relative widening their spreads rise by 15 and 20 bp and both prices fall by 0.6%. With a 20 bp widening for both, A loses 0.8% and B loses 0.6% at first order.
Both panels use the same price-loss scale. The estimates hold the benchmark curve fixed and omit spread convexity.
Exercise 4 in Credit-Sensitive Bonds works through an absolute spread shock in dollar terms (full lesson for members).

What a portfolio DTS tells you

For a cash bond portfolio, multiply each bond’s DTS by its starting market-value weight and add the contributions. That sum gives the portfolio’s first-order exposure to one common proportional spread shock. Estimating volatility also takes a model for the size and co-movement of relative spread changes. A jump-to-default loss depends on exposure and recovery. Ben Dor et al. (2007) and Houweling (2024)
Near zero, a small basis-point move can look enormous as a percentage of the starting spread. The ratio is undefined at zero, and multiplying a negative spread by 1.101.10 makes it more negative. An absolute basis-point shock gives a clearer scenario in those cases. For a large move, reprice the bond under the new spread to capture changes in sensitivity. The original study excluded bonds with nonpositive spreads. Ben Dor et al. (2007), data appendix

Sources

Last updated
  • credit spreads
  • spread duration
  • fixed income
  • credit risk
  • scenario analysis
← Back to the glossaryBack to top ↑