The Merton model is a structural credit-risk model that treats a company’s equity as a call option on its assets. In the basic model, default occurs at a specified debt horizon if firm asset value is below the promised debt payment.
Key Takeaways
- Equity holders receive the value left after debt is paid, which creates a call-option payoff.
- The model requires firm asset value and asset volatility, but neither is directly observable.
- Under the basic risk-neutral formulation, (N(-d_2)) is a model-implied default probability at the horizon.
- Risk-neutral default probability is not automatically an observed or “real-world” probability of default.
- The model is most informative as a structured sensitivity and ranking framework, not as a stand-alone credit decision.
Merton Model Setup
The basic one-period model assumes:
- firm assets have current value (V);
- asset value follows a lognormal diffusion with volatility (\sigma_V);
- the firm has one zero-coupon debt obligation with face value (D);
- debt matures at time (T);
- the continuously compounded risk-free rate is (r);
- markets are frictionless under the pricing assumptions;
- default can occur at maturity when (V_T < D).
At maturity, equity is:
$$
E_T = \max(V_T - D, 0)
$$
This is the payoff of a call option on firm assets with strike (D).
Under the model:
$$
E = V N(d_1) - D e^{-rT} N(d_2)
$$
where:
$$
d_1 =
\frac{\ln(V/D) + \left(r + \frac{1}{2}\sigma_V^2\right)T}
{\sigma_V\sqrt{T}}
$$
$$
d_2 = d_1 - \sigma_V\sqrt{T}
$$
(N(\cdot)) is the standard normal cumulative distribution function. In this simplified risk-neutral setup, the horizon default probability is:
$$
Q(V_T < D) = N(-d_2)
$$
The model-implied debt value is:
$$
B = V - E
$$
Inferring Asset Value and Volatility
Equity value (E) and equity volatility (\sigma_E) can be observed or estimated for a traded company, but (V) and (\sigma_V) cannot. A common implementation solves the equity-value equation together with:
$$
\sigma_E E = N(d_1)\sigma_V V
$$
The estimates are sensitive to the equity-volatility window, liability mapping, debt horizon, payout assumptions, and numerical method.
Worked Example
Assume:
- firm asset value (V = $120) million;
- debt due in one year (D = $100) million;
- asset volatility (\sigma_V = 25%);
- risk-free rate (r = 4%);
- horizon (T = 1).
The model inputs produce approximately:
$$
d_1 \approx 1.014
\qquad
d_2 \approx 0.764
$$
The model-implied risk-neutral default probability is:
$$
N(-0.764) \approx 22.2\%
$$
Estimated equity value is:
$$
E
\approx
\$120\text{m}\,N(1.014)
-
\$100\text{m}\,e^{-0.04}N(0.764)
\approx
\$26.7\text{m}
$$
The implied debt value is approximately:
$$
B = \$120\text{m} - \$26.7\text{m} = \$93.3\text{m}
$$
This example applies the model mechanically. The 22.2% figure is not an audited forecast of one-year default. It is conditional on the assumed asset value, volatility, debt boundary, market model, and risk-neutral measure.
What Changes the Model Result?
Holding other inputs constant:
- higher asset value generally reduces modeled default risk;
- higher debt increases the default boundary;
- higher asset volatility increases downside probability and equity option value;
- a longer horizon creates more time for asset value to move;
- different liability mapping can materially change the result;
- payouts, dilution, new borrowing, and asset sales can invalidate a static setup.
Sensitivity analysis should vary correlated inputs rather than moving one parameter in isolation when the business scenario affects several at once.
| Approach | Default mechanism | Typical evidence |
|---|
| Merton model | Asset value below debt at a stated horizon | Equity value, equity volatility, liabilities, rates |
| Structural Credit Model | Asset value crosses a model-defined boundary | Market and capital-structure inputs |
| Jarrow-Turnbull Model | Default arrives through a reduced-form process | Prices, term structures, intensity, recovery |
| Corporate Failure Prediction | Statistical relationship between indicators and a defined failure event | Accounting, behavior, market, and qualitative data |
Practical Uses
The model can support:
- distance-to-default monitoring;
- comparison of market-implied credit changes across public firms;
- theoretical valuation of risky debt;
- analysis of how leverage and volatility affect credit quality;
- benchmarking of Credit Spread and rating signals;
- stress testing and challenger-model analysis.
It should be combined with cash flow, liquidity, covenants, debt maturity, collateral, guarantees, industry, and management evidence.
Limitations
- Default occurs only at the horizon in the basic model.
- The firm is reduced to one asset process and one debt payment.
- Asset value and asset volatility are estimated.
- Lognormal diffusion omits jumps and some distress dynamics.
- Liability structure, interim coupons, covenants, and liquidity shortages are simplified.
- Equity prices and volatility can contain market noise and liquidity effects.
- The model is difficult to apply to private firms and complex financial institutions.
- Risk-neutral output contains market pricing effects and may differ materially from historical default frequency.
Common Mistakes
- Calling (N(-d_2)) a real-world PD without conversion or validation.
- Substituting book assets directly for modeled asset value.
- Using total liabilities as the debt boundary without documenting the choice.
- Omitting off-balance-sheet commitments, guarantees, or debt seniority.
- Treating a lower stock price as proof of imminent default.
- Using one volatility estimate without sensitivity analysis.
- Applying the model after a major capital-structure event without refreshing inputs.
Model Governance
A controlled implementation should document model purpose, theory, input lineage, estimation method, limitations, user controls, benchmark comparisons, sensitivity, validation, monitoring, overrides, and change history. Current U.S. interagency model-risk guidance emphasizes use-specific understanding, validation, monitoring, and response when performance deteriorates.
Primary and Official References
- Structural Model of Credit Risk: The broader model family in which default follows from firm value relative to obligations.
- Jarrow-Turnbull Model: A reduced-form alternative that models default timing without the same asset-value mechanism.
- Probability of Default (PD): A likelihood estimate whose interpretation depends on horizon, default definition, and probability measure.
- Credit Spread: A market measure that can be compared with model-implied risky-debt pricing, subject to non-credit effects.
- Corporate Failure Prediction: Statistical and judgmental approaches using accounting, behavioral, market, and qualitative warning indicators.
Educational Use
This article is educational and does not provide individualized investment, valuation, lending, accounting, capital, or regulatory advice. Model output depends on restrictive assumptions and requires validation for its intended use.