Credit Default Risk Explained
Credit Default Risk Explained
Credit risk = probability of default × loss if default occurs. The first is estimated from credit ratings, financial metrics, and market prices. The second depends on seniority, collateral, and recovery processes.
Key takeaways
- Default probability is the likelihood a company will be unable to pay interest or principal on schedule. It varies by rating: AAA is under 0.1%; CCC is 10–30% over 5 years.
- Loss-given-default (LGD) is what creditors lose if the company defaults: 100% minus recovery rate. Senior secured bonds recover ~60%; subordinated bonds recover ~20%.
- Expected loss = default probability × LGD. A BB bond with 5% default probability and 60% LGD has an expected loss of 3%.
- Credit ratings (from S&P, Moody's, Fitch) are the market's consensus on default probability and serve as proxies for expected loss.
- Credit analysts use financial metrics (leverage, coverage ratios), stress scenarios, and peer comparison to assess and monitor default risk.
Probability of default
The probability that a company will default in a given time period (1 year, 5 years, 10 years) depends on its creditworthiness. A large, stable utility with low debt is unlikely to default; a highly leveraged, cyclical company in a distressed industry is more likely.
Credit rating agencies publish implicit default probabilities based on historical default statistics. For example, according to Moody's 20-year historical data:
- AAA: ~0.1% cumulative default probability over 20 years.
- AA: ~0.3%.
- A: ~0.8%.
- BBB: ~2–3%.
- BB: ~10–15%.
- B: ~25–35%.
- CCC: ~50–70%.
These are historical averages; actual default probability for a specific company varies. A strong A-rated tech company (low leverage, high margins) might have a 1-year default probability of 0.01%; a struggling A-rated cyclical manufacturer (high leverage, thin margins) might be 0.5%.
Default probabilities also vary by time horizon. A company's 1-year default probability might be 0.5%, but its 5-year default probability might be 5% (as conditions could deteriorate). Longer horizons capture more uncertainty.
Estimating probability of default: Methods
Method 1: Credit ratings and historical data. The simplest approach is to use the implied default probability from a bond's credit rating. This assumes the rating agency's historical default rates are predictive of the future. Rating agencies have track records: Moody's, S&P, and Fitch publish annual default studies. An A-rated bond today has roughly the same likelihood of defaulting within 5 years as an A-rated bond in the past.
Method 2: Structural models (Merton model). The Merton model, developed by economist Robert Merton, treats a company's equity as a call option on its assets. If assets fall below the debt level, equity is worthless and the company defaults. By observing the company's equity price, volatility, debt level, and other factors, the model calculates a probability of default. Merton models are used by credit derivatives traders and risk managers; they're less intuitive for buy-and-hold bondholders but valuable for real-time pricing.
Method 3: Market prices of credit default swaps (CDS). The CDS market prices the cost to insure a bond against default. A company with a 1% 1-year default probability might have a CDS spread of 50–100 bp (investor pays 50 bp annually to buy protection). The CDS spread is an observable, real-time market price of default risk. Many analysts use CDS spreads as a check on credit ratings and their own estimates.
Method 4: Fundamental analysis. Analysts examine the company's financial statements, business model, competitive position, and management. They calculate financial metrics (debt/EBITDA, interest coverage, cash flow/debt) and stress the company's finances under adverse scenarios (recession, price competition, supply-chain disruption). From this analysis, they estimate the probability the company will run out of cash or be unable to refinance debt. This is more labor-intensive but allows for idiosyncratic insights.
Loss-given-default (LGD)
If a company defaults, creditors don't necessarily lose 100% of their investment. They recover something from selling assets, from the reorganization/bankruptcy process, or from other creditors' concessions. Loss-given-default (LGD) is 100% minus the recovery rate.
Recovery rates depend on seniority:
- Senior secured bonds (backed by specific collateral like equipment or real estate): Recovery 50–70%. Example: A bank's secured term loan backed by collateral.
- Senior unsecured bonds (first claim on general assets): Recovery 40–60%. Example: A utility's unsecured bonds.
- Subordinated bonds (paid after senior creditors): Recovery 10–40%. Example: A bank's subordinated debt.
- Equity (residual claim): Recovery 0–20% (usually near zero). Example: Common stock after bankruptcy.
Recovery varies by company and situation. In a Chapter 11 bankruptcy, the time from default to emergence is typically 1–3 years; creditors recover based on what assets are worth at emergence, discounted for the wait. In a fire-sale liquidation (Chapter 7), recovery is usually low (20–30% for senior debt) because assets are sold quickly under pressure.
For example, during the 2008 financial crisis:
- Lehman Brothers creditors (senior unsecured): recovered ~20–30% over several years (vs. expectation of 50–60%), because the bankruptcy was complex and markets were frozen.
- General Motors (senior debt): recovered ~35%, because the company was reorganized and equity holders were wiped out, but senior debt was converted and traded in the restructured equity.
- Small commercial real estate companies: senior lenders often recovered 80–90%, because collateral was valuable and processes were fast; subordinated lenders recovered little.
Expected loss and risk pricing
Expected loss (EL) is the probability-weighted loss:
EL = Probability of Default × Loss-Given-Default
Example 1: A BB-rated industrial company with a 5-year default probability of 10% and an estimated recovery rate of 40% (LGD = 60%):
- EL = 10% × 60% = 6%
An investor in a 5-year bond should expect (on average) to lose 6% of principal over those 5 years, assuming the rating and recovery assumptions hold.
Example 2: A senior unsecured bond of an A-rated utility with a 5-year default probability of 0.5% and recovery of 50% (LGD = 50%):
- EL = 0.5% × 50% = 0.25%
The expected loss is much lower, reflecting the lower risk. Investors in A bonds demand lower yields to compensate for lower expected loss.
In an efficient market, the yield spread above Treasuries should roughly correspond to expected loss plus a premium for illiquidity and other factors. A bond with 6% expected loss should yield significantly more than a bond with 0.25% expected loss, all else equal.
Credit rating notches and "fallen angel" risk
Credit rating agencies assign ratings in notches (AAA, AA, A, BBB, BB, B, CCC) with modifiers (+/−). For example, a company might be rated "BBB+" (stable BBB) or "A−" (weaker A).
Within a notch, there is wide variation in default probability. A BBB+ company might have a 1% 5-year default probability; a BBB− company might have 5%. Yet both are rated BBB. Investors who understand this nuance can identify "fallen angels"—companies at risk of downgrade—before the rating agencies act.
For example, in the 2008 financial crisis, many financial companies were rated A or AA with stable outlooks in 2006 and 2007. By 2008, after the subprime collapse, they were downgraded to BB or lower. Investors who had analyzed leverage, off-balance-sheet risks, and mortgage exposure would have seen the downgrade coming; those who relied solely on ratings were caught off-guard.
Modeling default scenarios
Professional credit analysts build stress scenarios to estimate default probability under adverse conditions. For example, an analyst covering an airline might model:
Base case: Economic growth continues; fuel prices remain around $80/barrel; airline margins stay at 5–8%. Default probability: 0.5%.
Recession scenario: GDP contracts 2%; fuel prices spike to $120/barrel; margins compress to 1–2%; one major competitor exits the market. The airline's cash flow deteriorates, debt refinancing becomes difficult, and default probability rises to 8–10%.
Severe stress: Economic depression; fuel prices soar to $150+/barrel; margin collapse to negative (the airline loses money daily); key debt maturities loom in 2 years. Default probability: 30–40%.
By weighting these scenarios (e.g., 70% base case, 20% recession, 10% severe stress), the analyst can estimate the company's overall expected default probability: 0.7% × 0.005 + 0.2 × 0.10 + 0.1 × 0.35 = 0.875%.
This is more nuanced than simply accepting a credit rating or a backward-looking historical default rate. It accounts for the company's specific situation and the current macro environment.
Early warning signals
Credit analysts monitor warning signs of deteriorating default probability:
- Rising leverage (debt/EBITDA above historical levels or peer average).
- Declining coverage (EBITDA/interest expense falling; earnings becoming volatile).
- Weakening competitive position (market share loss, customer concentration, tech disruption).
- Management turnover (CFO, CEO, or board changes often precede troubles).
- Liquidity stress (delayed payables, increased reliance on credit facilities, frequent refinancing).
- Rating agency warnings (negative outlook assigned before a downgrade; CDS spreads widening ahead of rating action).
A company showing several of these signs warrants closer scrutiny. Bondholders who catch these early can sell before spreads blow out or exit bonds before defaults occur.
Process
Next
Default probability and loss-given-default are the two sides of credit risk. When defaults do occur, recovery is never 100%, but it can range widely. The next article examines recovery rates in detail: what senior secured bonds recover, what subordinated bonds recover, and real-world examples of how recoveries are realized in bankruptcy.