Expected credit loss is the product of two key variables: the probability of default (PD) and the loss given default (LGD). LGD equals one minus the recovery rate. These variables — their historical levels, the factors that drive them, and how they vary across credit cycles — are fundamental to credit analysis, portfolio management, and loan pricing.
Default rates measure the percentage of issuers (or dollar amount of outstanding debt) that fail to make scheduled payments in a given period. Investment-grade default rates have historically been extremely low — often below 0.1% annually — reflecting the low-risk profile of this borrower universe. Speculative-grade default rates are higher and more volatile: in benign credit environments they may run 1–3%, while during recessions or credit crises they can spike to 10–15% or higher. The default rate peak during the 2008–2009 financial crisis exceeded 12% for speculative-grade issuers.
Recovery rates measure how much lenders ultimately recover on defaulted obligations, expressed as a percentage of face value. Recovery rates vary significantly by seniority: first-lien senior secured loans have historically recovered 60–80 cents on the dollar, senior unsecured bonds 40–60 cents, and subordinated debt 20–40 cents. These averages mask wide variation based on industry, macro conditions, and specific deal characteristics.
The "negative correlation" between default rates and recovery rates is one of the most important — and often underappreciated — features of credit markets. When macroeconomic conditions deteriorate and default rates rise, recovery rates tend to fall simultaneously, because: (1) more credits are distressed at the same time, flooding the market with collateral; (2) going-concern values are depressed in recessions; and (3) buyers of distressed assets demand larger discounts when uncertainty is high. This correlation means credit losses in downturns are typically worse than models that treat PD and LGD as independent variables would predict.
Credit models must incorporate realistic assumptions about both variables and their co-movement. Stress testing — running the model under severe but plausible recession scenarios — is essential for understanding tail risk in credit portfolios. The 2008–2009 crisis served as a reminder that historical averages can be misleading when structural changes (such as cov-lite lending or high leverage multiples) shift the underlying risk profile of the market.