CDS duration measures how much the mark-to-market value of a credit default swap changes when the reference entity’s credit spread moves by a small amount. A contract with a duration of 5 would gain or lose roughly 5% of its notional value for every 1% shift in the credit spread. Unlike bond duration, which tracks sensitivity to interest rate changes, CDS duration is entirely about credit risk, and getting it right is what separates a hedged book from an accidental bet on a single name’s creditworthiness.
Why It Is Not Bond Duration
Modified duration on a bond tells you how the price moves when the risk-free interest rate changes. If Treasury yields rise 50 basis points, modified duration estimates the price drop. CDS duration ignores interest rates almost entirely. It captures how the contract responds to changes in the credit spread of the reference entity, whether that is a corporation, a sovereign, or a structured product.
The distinction matters because the two risk factors often move independently. A company’s spread can widen sharply while Treasury yields stay flat, or even fall. A portfolio manager holding both bonds and CDS needs separate duration measures for each factor, or the hedge ratios will be wrong.
Risky PV01: The Building Block
The measure underneath CDS duration is risky PV01, sometimes written RPV01. It is the present value of receiving one basis point per year over the remaining life of the CDS, discounted by both the risk-free rate and the probability that the reference entity survives to make each payment. That survival-weighted discounting is what makes it “risky” rather than a plain annuity.
If a five-year CDS has a risky PV01 of 4.3, a stream of one-basis-point annual payments, adjusted for the chance the reference entity defaults before each payment date, is worth 4.3 basis points of notional in today’s dollars. The higher the default probability, the lower the risky PV01, because there is a greater chance the premium stream gets cut short by a credit event.
Risky PV01 also ties directly to how the CDS premium is priced. The annual spread a protection buyer pays is set so that the present value of those spread payments, weighted by survival probability, equals the present value of the expected default payout. Risky PV01 is the multiplier that converts the spread into a dollar value for the premium leg.
DV01 and CS01: The Daily Number
Traders rarely talk about risky PV01 in isolation. The working risk metric is DV01, also called CS01 (Credit Spread 01). The two terms are interchangeable in the CDS market. DV01 answers a simple question: if the credit spread moves by one basis point, how many dollars does the position gain or lose?
The calculation is straightforward. Multiply the risky PV01 by the notional amount, then scale to one basis point. A $10 million notional CDS with a risky PV01 of 4.3 has a DV01 of roughly $4,300. If the spread widens by one basis point, the protection seller loses about $4,300 in mark-to-market value and the protection buyer gains the same amount.
The one-basis-point shock is the industry’s standard unit for comparing risk across contracts. A desk might be short protection on Company A with a DV01 of $15,000 and long protection on Company B with a DV01 of $12,000. The net DV01 of $3,000 tells the trader how exposed the book is to a broad one-basis-point widening across both names.
What Pushes DV01 Up or Down
Two factors dominate. The first is maturity. A ten-year CDS has roughly twice the DV01 of a five-year CDS on the same reference entity, because the premium stream is longer and the survival-weighted present value of each basis point is spread over more years. Longer-dated contracts carry significantly more spread risk per unit of notional.
The second is the credit quality of the reference entity. A name trading at very wide spreads, say 500 basis points or more, implies a high probability of default. That elevated default probability reduces risky PV01 because fewer premium payments are expected to actually occur. Counterintuitively, very distressed names have lower DV01 per unit of notional than investment-grade names of the same maturity. The spread is high, but the duration is short.
Sensitivity Along the Credit Curve
DV01 as described so far assumes the entire credit curve shifts in parallel, with every maturity point moving by the same basis point. Real spread movements are rarely that uniform. The five-year point on a curve might widen while the ten-year point stays flat, or the short end might invert relative to the long end.
Risk systems address this by breaking the credit curve into individual tenor points and calculating a separate DV01 at each one. This is analogous to key rate duration in the bond world. A portfolio that looks flat on aggregate DV01 can still carry substantial risk if it is long the short end and short the long end of the same issuer’s curve.
Where the Linear Estimate Breaks
DV01 is a linear approximation. It assumes the relationship between spread changes and value changes is a straight line. For small moves of a few basis points, the approximation is excellent. For larger moves, it starts to miss.
The actual price-spread relationship is curved. When spreads widen significantly, the protection buyer gains more than DV01 alone would predict. When spreads tighten sharply, the buyer loses less than the linear estimate suggests. This asymmetry is convexity, and it works in the protection buyer’s favor. A trader who is long protection on a name that suddenly blows out by 200 basis points will find more profit in the position than a simple “DV01 times 200” calculation would have indicated.
Convexity is a second-order effect. It only becomes material during significant credit events or broad market dislocations. For routine daily risk management, DV01 does the heavy lifting. But ignoring convexity when stress-testing a portfolio against large spread shocks leads to underestimating gains on long protection positions and overestimating gains on short protection positions.
How Traders Actually Use It
The most common use is hedging. A bank holding $50 million in corporate bonds can buy CDS protection sized so that the DV01 of the CDS position offsets the credit spread duration of the bond holdings. If the bond portfolio has a credit DV01 of $22,000, the hedge needs to deliver approximately $22,000 in DV01 on the other side. The trader picks the notional and maturity combination that hits that target.
Risk aggregation is the other core application. A credit desk running positions across dozens of reference entities needs a single number summarizing total spread exposure. Summing DV01 across all positions gives that figure. Risk limits are typically set in DV01 terms: a desk might have a limit of $500,000 net DV01, meaning the portfolio’s value cannot move by more than $500,000 for a one-basis-point parallel spread shift.
Duration also feeds profit-and-loss attribution. When a trader’s book makes or loses money overnight, the risk system decomposes the P&L into components: how much came from spread changes (DV01 times the actual spread move), how much from the passage of time, and how much from curve reshaping. Without an accurate DV01, none of that decomposition works, and the trader cannot tell whether today’s profit came from skill or from an unintended position the book happened to be carrying.
Contract Conventions That Shape the Number
CDS contracts trade with standardized coupons, typically 100 basis points for investment-grade names and 500 basis points for high-yield names. When the market spread differs from the fixed coupon, the difference is settled as an upfront payment at trade inception. That upfront exchange does not change the DV01 of the position, because DV01 measures ongoing sensitivity to spread movements rather than the one-time settlement.
Premium payments are made quarterly, and standard maturities roll on a fixed ISDA schedule tied to March 20, June 20, September 20, and December 20. A “five-year” CDS purchased in January will actually mature on the next June 20 five years out, giving it slightly more than five years of remaining life. These conventions are uniform across the market, which is what makes DV01 comparable between counterparties and venues.
As the contract approaches maturity, its risky PV01 declines naturally because fewer premium payments remain. A five-year CDS bought today will behave like a four-year CDS a year from now, with a correspondingly lower DV01. A trader who wants to maintain a constant level of spread exposure has to roll the position into a new on-the-run contract, which resets the duration back to the full term.