Credit Spread

Decomposing a credit spread is the single most useful exercise a candidate can perform. The first component is expected credit loss: the annualised probability of default multiplied by loss given default, where loss given default equals one minus the recovery rate. A bond with a 2% annual default probability and a 40% recovery rate carries roughly 120 basis points of expected loss compensation.

The remainder is a risk premium. Investors demand payment for the uncertainty around that expected loss, for the fact that defaults cluster in recessions precisely when investors can least afford them, and for the liquidity cost of exiting a position in a stressed market. Taxation and regulatory treatment can also drive a wedge. Empirically the residual premium is large, often exceeding the expected-loss component for investment-grade issuers, which is why realised default losses have historically fallen well short of spreads earned.

Several spread measures coexist because each strips out a different distortion. The G-spread is the bond's yield to maturity minus the interpolated yield on a government bond of matching maturity; the I-spread uses the swap curve instead. Both compare single yields and therefore embed cash-flow timing differences.

The Z-spread corrects this by adding a constant increment to every point on the spot curve until discounted cash flows equal market price. For callable, putable or convertible bonds the Z-spread is contaminated by option value, so the option-adjusted spread removes it using an interest-rate model: OAS equals Z-spread minus option cost for a callable bond, and Z-spread plus option value for a putable bond. Separately, the CDS basis compares the credit default swap premium with the cash bond spread; persistent negative basis typically signals funding or balance-sheet constraints rather than a pure credit view.

Credit portfolios are managed in spread-duration space. Spread duration measures the percentage price change of a bond for a one-percentage-point change in its credit spread, holding the benchmark curve constant. The approximate return impact equals the negative of spread duration multiplied by the change in spread, with a convexity adjustment for large moves. A bond with spread duration of seven that widens by 50 basis points loses roughly 3.5% of price.

Excess return over a duration-matched government position therefore equals spread carry earned over the holding period minus the spread-duration-weighted widening, minus realised default losses net of recovery. This decomposition explains why short-dated high-yield paper can outperform long investment-grade credit in a widening episode despite carrying far higher default risk: the shorter spread duration dominates.

Spreads are countercyclical. They compress through expansions as issuers deleverage and investors reach for yield, then gap wider at the turn, typically ahead of realised default rates, because markets price forward-looking loss expectations. Government yields often fall simultaneously in a flight to quality, so total yields on high-grade credit can decline even as spreads widen, an interaction candidates frequently misread.

Curve shape matters too. Investment-grade spread curves normally slope upward, while distressed issuers show inverted spread curves because near-term survival is the binding question. At Level I you should be able to compute expected loss from default probability and recovery, convert a spread change into a price or return impact using spread duration, rank G-spread, I-spread, Z-spread and OAS by what each isolates, and explain why the option-adjusted measure is the only fair comparison across bonds with different embedded options.