How to Calculate Spread Duration: Formula, Example, and Convexity

To calculate spread duration, divide the bond’s percentage price change by the change in its credit spread (expressed as a decimal), then flip the sign: Spread Duration = −(ΔP / P) ÷ Δs. The result tells you roughly how much the bond’s price will move for every one-percentage-point shift in its credit spread. A spread duration of 5.0 means the price drops about 5% if the spread widens by 100 basis points, and rises about 5% if it tightens by the same amount.

The arithmetic is simple. Getting the inputs right is where most errors happen.

The Three Inputs You Need

You need the bond’s current market price, the credit spread at the start of your observation window, and the credit spread at the end.

Use the dirty price, not the clean price. The dirty price is the full settlement value including accrued interest since the last coupon payment; the clean price strips that out. Most trading screens quote the clean price because it makes day-to-day comparisons tidier, but spread duration asks about the actual cash value of the position. Using the clean price throws off the percentage-change figure, and the error is largest mid-coupon-cycle when accrued interest is a meaningful share of the total.

The credit spread is the gap between the bond’s yield to maturity and the yield on a comparable-maturity Treasury. “Comparable” usually means the nearest on-the-run Treasury, the most recently issued security at that maturity point. Benchmark maturities are 2, 5, 10, and 30 years. For a bond that matures off the benchmark grid, say a 7-year corporate, you can either compare against the nearest benchmark or interpolate between the 5-year and 10-year yields for a synthetic 7-year rate. Interpolation is more precise.

For real-time corporate bond prices, FINRA’s TRACE system is the most reliable public source. It publishes the time, price, yield, and volume of over-the-counter bond trades reported by broker-dealers, with over 80% of transactions available within five minutes of execution. You can search by CUSIP or ticker through FINRA’s Fixed Income Security Lookup at no cost. Treasury yields used as benchmarks also appear through TRACE, though on a next-day basis.1FINRA.org. What Is TRACE and How Can It Help Me

The Formula

Spread Duration = −(ΔP / P) ÷ Δs

ΔP is the dollar change in the bond’s price, P is the initial price, and Δs is the change in the credit spread expressed as a decimal (25 basis points becomes 0.0025). The negative sign is a convention: bond prices and spreads move in opposite directions, so when spreads widen (positive Δs) and the price falls (negative ΔP), the two negatives cancel and spread duration reads as a positive number for normal bonds.

The formula assumes the risk-free rate holds steady during your observation window. In reality, Treasury yields and credit spreads often move together, which tangles interest-rate risk and credit risk inside the same price change. Spread duration works best over short windows where you can reasonably attribute the price movement to the spread alone, or where you have already hedged the interest-rate piece.

A Worked Example

Suppose you hold a corporate bond priced at $1,025 (dirty price per $1,000 face value). Its credit spread is 150 basis points over the 10-year Treasury. Over the next week the spread widens to 175 basis points while the Treasury yield barely moves. The bond’s dirty price falls to $1,008.

  • Convert the spread change to a decimal. 175 minus 150 is 25 basis points. Divide by 10,000 to get 0.0025.
  • Calculate the percentage price change. The bond dropped $17 from $1,025, so −17 ÷ 1,025 = −0.01659, or roughly −1.66%.
  • Divide and flip the sign. Spread Duration = −(−0.01659) ÷ 0.0025 = 6.64.

A spread duration of 6.64 means the price moves about 6.64% for every 100-basis-point shift in the credit spread. That is moderately high sensitivity, typical of an intermediate-maturity investment-grade bond.

One detail that trips people up: the spread change and the price change must cover exactly the same window. If you use Friday’s closing price alongside Monday’s spread quote, the mismatch contaminates the result. Obvious in theory, common in practice when bond data lives across platforms that update on different clocks.

Callable Bonds Use a Different Formula

Everything above assumes fixed coupons paid on schedule until maturity. Callable bonds break that assumption. If the issuer can redeem the bond early, the expected cash flows shift whenever yields move, and the standard calculation no longer captures price sensitivity correctly.

For callable bonds, use effective spread duration, sometimes called option-adjusted spread duration:

Effective Spread Duration = (P₋ − P₊) ÷ (2 × P₀ × Δs)

P₋ is the modeled price after a small spread decrease, P₊ is the modeled price after a spread increase of the same size, P₀ is the current price, and Δs is the size of the shift. Rather than observing an actual price change, you use an option-adjusted model to reprice the bond under both scenarios, accounting for the probability that the issuer calls the bond at each possible call date.

When a callable bond trades at a premium and rates are falling, its price gains slow down because the call option is more likely to be exercised. That produces negative convexity, and the effective spread duration ends up shorter than the modified duration you would calculate for an identical non-callable bond. Ignoring the call provision overstates sensitivity and leads to oversized hedges.

Adding Convexity for Larger Spread Moves

Spread duration is a first-order approximation: it draws a straight line through a relationship that is actually curved. For small changes of 10 to 25 basis points, the line is close enough. For larger moves, convexity matters.

The improved estimate adds a convexity term:

%ΔP ≈ (−Spread Duration × Δs) + (½ × Convexity × Δs²)

Convexity is always positive for option-free bonds, so the duration-only estimate overstates losses when spreads widen and understates gains when spreads tighten. In a crisis where spreads blow out 200 or 300 basis points, duration alone can miss by a meaningful amount. During the 2008 financial crisis and the March 2020 liquidity shock, bonds that duration said “should have” dropped 10% often moved 7% or 14% depending on which way convexity cut.

For normal conditions and investment-grade bonds, the duration-only number is enough. For stress-testing a severe scenario, add the convexity term.

Calculating Portfolio Spread Duration

Portfolio spread duration is a market-value-weighted average of the individual bonds. Multiply each bond’s spread duration by its share of total portfolio market value, then sum the results.

Treasuries carry a spread duration of zero because they have no credit spread. If a portfolio is 40% Treasuries and 60% corporate bonds with an average corporate spread duration of 5.0, the portfolio-level figure is 0.60 × 5.0 = 3.0, not 5.0. This is why some managers report spread duration both for the full portfolio and for the credit-sensitive portion alone.

A simplified three-bond example:

  • Bond A: $500,000 market value, spread duration of 3.2
  • Bond B: $300,000 market value, spread duration of 6.1
  • Treasury position: $200,000 market value, spread duration of 0

Total portfolio value is $1,000,000. Portfolio spread duration is (500,000 / 1,000,000 × 3.2) + (300,000 / 1,000,000 × 6.1) + (200,000 / 1,000,000 × 0) = 1.60 + 1.83 + 0 = 3.43. A 50-basis-point spread widening across the board would cost this portfolio roughly 1.72%, or $17,200.