Structural vs. reduced-form credit models — what's the real difference and which does the CFA exam emphasize?

Question

AcadiFi · Accepted Answer

The exam is predominantly conceptual for credit models, but you need to understand the intuition behind both approaches and their strengths and limitations. Let me compare them side by side.

## Structural Models (Merton Model)

**Core idea:** The company's equity is a European call option on its assets, with the face value of debt as the strike price.

**How it works:**
- Company has assets worth V and zero-coupon debt with face value K maturing at time T
- At maturity: if V > K, equity holders get (V - K) and debt holders get K
- If V < K, the firm defaults. Equity = 0, debt holders get V
- This is exactly the payoff of a call option on V with strike K

**Example with Stratton Mining:**
- Asset value today: $500M
- Volatility of assets: 30%
- Debt face value (due in 2 years): $350M
- Risk-free rate: 4.5%

Using Black-Scholes-Merton, you'd value the equity as a call option, and the value of risky debt = Asset value - Equity value. The credit spread is implied by comparing the risky debt yield to the risk-free rate.

**Strengths:** Economically intuitive, links equity and credit markets, endogenizes default.
**Weaknesses:** Requires observable asset value and volatility (neither is directly observable), assumes simple capital structure, poor at predicting short-term defaults.

## Reduced-Form Models

**Core idea:** Default is modeled as a random event governed by a statistical process (hazard rate), without explicitly modeling the firm's assets.

**How it works:**
- Define a **hazard rate** (lambda) — the instantaneous probability of default at any moment, conditional on having survived to that point
- The hazard rate can depend on observable variables: credit spreads, macro conditions, stock prices, etc.
- **Probability of survival** to time T = e^{-lambda x T} (if lambda is constant)
- **Expected loss** = Probability of default x Loss given default (LGD)

**Example:**
- Hazard rate for Whitfield Telecom: 2.5% per year
- Recovery rate: 40% (so LGD = 60%)
- Probability of surviving 3 years: e^{-0.025 x 3} = e^{-0.075} = 92.77%
- Probability of default within 3 years: 1 - 0.9277 = **7.23%**
- Expected loss (3 years) = 7.23% x 60% = **4.34%**

**Strengths:** Uses observable market data, can be calibrated to credit spreads, handles complex capital structures, better for short-term default prediction.
**Weaknesses:** Does not explain *why* default occurs (treats it as a statistical event), requires estimation of hazard rate.

## Side-by-Side Comparison

| Feature | Structural | Reduced-Form |
|---|---|---|
| Default trigger | Asset value < Debt | Random process |
| Key input | Asset value, volatility | Hazard rate, recovery |
| Observable? | No (must estimate) | Yes (from market data) |
| Short-term accuracy | Poor | Better |
| Economic intuition | Strong | Weaker |
| Capital structure | Simple only | Flexible |

## What the Exam Tests

Expect conceptual comparisons: "Which model is better for a firm with complex capital structure?" (reduced-form). "Which model endogenizes default?" (structural). You may also see calculations of survival probability using the hazard rate formula.

For full credit analysis coverage with practice vignettes, check out AcadiFi's CFA Level II Fixed Income curriculum.

Structural vs. reduced-form credit models — what's the real difference and which does the CFA exam emphasize?

Master Level II with our CFA Course

Related Questions

Practice Questions