The credit default swap remains one of the most consequential innovations in financial engineering—a bilateral contract that fundamentally altered how institutions transfer and express views on credit risk. Yet beneath its deceptively simple payoff structure lies a valuation framework demanding rigorous treatment of default probability, recovery assumptions, and discount rate selection. Mispricing any of these components doesn't just produce theoretical errors; it creates exploitable dislocations that sophisticated participants actively hunt.

Since the 2008 crisis reshaped CDS market infrastructure—pushing standardized contracts toward central clearing and fixed-coupon conventions—the analytical landscape has grown more nuanced, not less. The interplay between single-name protection, index products, and tranched structures generates a rich set of relative value relationships that reward precise quantitative reasoning. Understanding these dynamics requires moving beyond textbook no-arbitrage arguments into the institutional frictions and liquidity regimes that govern real-world pricing.

This treatment develops the CDS valuation framework from first principles, examines the persistent and time-varying basis between CDS spreads and bond yields, and extends the analysis to index products and their tranched derivatives. The goal throughout is not merely theoretical elegance but practical applicability—connecting hazard rate calibration and recovery modeling to the trading decisions and risk management challenges facing quantitative analysts and portfolio managers in institutional settings.

CDS valuation decomposes into two present value calculations: the premium leg, representing the stream of periodic spread payments from protection buyer to seller, and the protection leg, representing the contingent payment triggered by a credit event. The fair spread equates these two legs at inception. Under a reduced-form framework, the survival probability to time t is expressed as exp(−∫₀ᵗ λ(s) ds), where λ(s) is the hazard rate—the instantaneous conditional default intensity. For a constant hazard rate λ, survival probability decays exponentially, and the premium leg simplifies to a summation of discounted survival-weighted coupon payments across payment dates.

The protection leg requires modeling both the timing and magnitude of the loss given default. Under standard ISDA conventions, recovery is applied to par—typically assumed at 40% for senior unsecured obligations—making the loss given default 60 cents on the dollar. The protection leg's present value integrates the product of the discount factor, default probability density, and loss given default across the contract's tenor. When calibrating to market-observed CDS spreads, we extract the implied hazard rate term structure by bootstrapping across multiple tenors, analogous to how zero-coupon yields are stripped from bond prices.

The shift to fixed-coupon conventions (100 basis points for investment grade, 500 for high yield) introduced upfront payments at trade inception—the difference between the present value of the running coupon and the fair spread. This seemingly administrative change has meaningful consequences for mark-to-market risk and margin dynamics. A contract trading at a spread significantly different from the fixed coupon embeds substantial upfront value, altering the counterparty exposure profile and the sensitivity to recovery rate assumptions.

Recovery rate sensitivity is often underappreciated. The hazard rate implied by a given spread depends critically on the assumed recovery: higher recovery assumptions require higher default intensities to justify the same observed spread. This creates a circularity that practitioners resolve either through market-implied recovery extraction (using recovery swaps or senior-subordinated CDS pairs) or through stress-testing across a recovery range. For distressed credits, where recovery uncertainty dominates, the standard 40% assumption can generate severely misleading risk metrics.

The risky duration (RPV01)—the present value of a one-basis-point annuity conditional on survival—serves as the fundamental sensitivity measure, linking spread changes to profit and loss. For a five-year investment-grade name with a risky duration of approximately 4.5 years, a one-basis-point spread widening produces a loss of roughly $4,500 per $10 million notional. This linearity holds well for small spread moves but breaks down for distressed credits where convexity effects—the nonlinear relationship between spread and price—become material.

Takeaway

The fair CDS spread is not a single number but an output of interacting assumptions about default intensity, recovery, and discounting. Changing any one input reshapes the entire calibration, making sensitivity analysis across the assumption space more informative than any point estimate.

In a frictionless world, the CDS spread on a reference entity should equal the credit spread on that entity's par-priced floating-rate bond over the risk-free rate. This equivalence—the CDS-bond basis—is zero under idealized no-arbitrage conditions. In practice, the basis is persistently nonzero and time-varying, driven by a constellation of institutional frictions, funding costs, and supply-demand imbalances that create genuine opportunities for relative value strategies.

A negative basis (CDS spread below bond spread) signals that protection is cheap relative to the cash bond market. This typically arises when bond spreads widen due to forced selling, illiquidity, or financing constraints while CDS spreads remain anchored by dealer hedging flows. The classic negative basis trade—buying the bond and buying CDS protection—captures the basis as carry, provided the trader can finance the bond position and sustain the trade through mark-to-market volatility. The trade's apparent simplicity masks significant execution risks: funding costs, counterparty exposure, and the possibility of basis widening before convergence.

A positive basis (CDS spread above bond spread) can emerge from several sources: the cheapest-to-deliver option embedded in CDS contracts, supply-demand dynamics in the protection market (e.g., structured credit desks buying protection to hedge synthetic CDO positions), or differences in the risk-free rate benchmark used for bond spread measurement versus CDS discounting. During periods of elevated correlation and systemic stress, positive basis regimes can persist as hedging demand overwhelms arbitrage capital.

The 2008 crisis provided a dramatic case study. The CDS-bond basis on investment-grade names collapsed to deeply negative levels—exceeding −200 basis points in some cases—as bond markets seized and leveraged basis traders faced margin calls. The dislocation revealed that basis convergence trades are fundamentally carry trades with embedded liquidity risk: they earn steady income in normal environments but suffer catastrophic drawdowns when funding markets freeze. Post-crisis regulatory changes, including higher capital charges for counterparty credit risk and mandatory clearing, have structurally altered basis dynamics by changing the cost of maintaining both legs of the trade.

For practitioners, monitoring the basis requires decomposing it into its fundamental and technical components. Fundamental drivers include the bond's coupon structure, accrued interest treatment in credit events, and the CDS contract's restructuring clause (CR, MR, XR variants). Technical drivers encompass repo specialness, dealer balance sheet constraints, and seasonal patterns in index rolls. A robust basis trading framework models these components separately, establishing fair-value estimates for the basis and trading around deviations—treating the basis itself as an asset with its own mean-reversion properties and risk factors.

Takeaway

The CDS-bond basis is not arbitrage mispricing waiting to be collected—it is compensation for bearing liquidity, funding, and convergence risk. Profitable basis trading demands modeling these friction costs as carefully as the credit risk itself.

CDS indices—CDX in North America, iTraxx in Europe—aggregate single-name CDS contracts into standardized, equally-weighted portfolios that trade as a single instrument. The index spread approximates the average spread of its constituents, but does not equal it exactly. The index-intrinsic spread, computed by repricing each constituent CDS at current market levels and averaging, typically diverges from the traded index spread due to the index's own supply-demand dynamics, the convenience of trading a liquid on-the-run product, and the skew created by distressed names whose upfront pricing conventions differ from the index's running-spread convention.

This index-versus-intrinsics basis (or skew) is a critical relative value signal. When the traded index is cheap relative to intrinsics, dealers with the infrastructure to delta-hedge single names can buy the index and sell the components, earning the basis as the index converges toward intrinsic value at maturity. The trade requires managing the idiosyncratic risk of each constituent and the correlation exposure embedded in any portfolio versus single-name position. Index roll dates—when the market transitions from one series to the next, with updated constituent lists reflecting credit rating migrations—create additional technical dislocations.

The true analytical complexity emerges with index tranches, which slice the index's loss distribution into subordinated layers. The equity tranche (typically 0-3% for CDX IG) absorbs the first losses and offers leveraged exposure to spread and default risk. Senior tranches (e.g., 15-30%) are exposed only to catastrophic, highly correlated default scenarios. Tranche pricing depends fundamentally on the correlation structure among constituent defaults—a dimension entirely absent from single-name or index-level analysis.

The standard market pricing convention uses the Gaussian copula model with a single implied correlation parameter per tranche—the compound correlation. However, this framework produces the well-documented correlation smile: equity and senior tranches imply different correlations, revealing the model's inability to capture the true multivariate default distribution with a single parameter. Base correlation—which prices each tranche as the difference between two first-loss positions—provides a monotonically increasing curve that practitioners use for interpolation and relative value assessment, despite its own theoretical limitations.

For portfolio managers, tranches offer a mechanism to express views on correlation, tail risk, and idiosyncratic versus systematic credit exposure with precision unavailable in index or single-name space. Selling equity tranche protection and buying mezzanine protection isolates correlation exposure: the position profits if realized correlation exceeds the level implied by tranche pricing. These correlation trades require continuous delta-hedging against index spread movements and careful monitoring of jump-to-default risk—the discrete, unhedgeable loss from a sudden credit event in a concentrated position. The analytical infrastructure needed to manage a tranche book extends well beyond CDS pricing into Monte Carlo simulation, copula calibration, and scenario-based stress testing.

Takeaway

CDS indices simplify credit exposure but tranches reintroduce profound complexity through correlation dependence. The lesson: aggregation creates new risk dimensions that cannot be understood by analyzing the components in isolation.

Credit default swaps sit at a unique intersection of mathematical elegance and institutional messiness. The valuation framework—hazard rates, survival probabilities, risky durations—provides a rigorous foundation, but the real analytical challenge lies in the gaps between theory and market reality: recovery uncertainty, basis frictions, and the correlation assumptions embedded in structured products.

The progression from single-name CDS to index tranches illustrates a broader principle in quantitative finance: each layer of product complexity introduces risk dimensions invisible at the prior level. Spread risk gives way to basis risk, which gives way to correlation risk. Managing this hierarchy demands not just mathematical sophistication but a deep understanding of the institutional plumbing that generates persistent mispricings.

For practitioners, the imperative is clear: model the frictions, not just the payoffs. The most profitable opportunities in CDS markets arise precisely where elegant theory collides with the messy realities of funding, liquidity, and counterparty constraints.