Merton Model for Default Probability

Introduction

In our previous lesson, we established the fundamental framework for quantifying credit risk through the Expected Credit Loss formula: $ECL = EAD \times PD \times LGD$ . We now turn our attention to the first, and often most challenging, component of this equation: the Probability of Default (PD).

This lesson introduces the pioneering Merton structural model, our first quantitative tool for estimating a company's PD. This is a "structural" model because it derives the risk of default from the firm's own financial structure—specifically, the relationship between its assets and its liabilities. We will explore the elegant core idea behind the model, learn how to calculate its key metric, the "Distance to Default," and apply the full formula to estimate the probability that a firm will be unable to meet its debt obligations.

The Core Idea: Equity as a Call Option

The genius of Robert Merton's 1974 model was to recognize a powerful analogy between a company's capital structure and an option contract. In this framework, the equity (or stock) of a company with debt is not just a simple share of ownership; it can be viewed as a European call option on the total assets of the firm.

Let's break down this analogy:

The Underlying Asset: The total value of the firm's assets, $V$ .
The Strike Price: The face value of the firm's debt, $D$ , which it must repay at a future maturity date, $T$ .
The Option Holder: The firm's shareholders (equity holders).

At the debt's maturity date $T$ , two scenarios can unfold:

Assets exceed Debt ( $V_{T} > D$ ): The shareholders will "exercise their option." They use the firm's assets to pay off the debt holders the amount $D$ , and they keep the remaining (residual) value, $V_{T} - D$ .
Assets are less than or equal to Debt ( $V_{T} \leq D$ ): The shareholders will let their "option" expire worthless. They receive nothing. The firm defaults, and the debtholders take possession of the firm's remaining assets, which are worth $V_{T}$ .

Therefore, the payoff to equity holders at time $T$ is $max (V_{T} - D, 0)$ , which is precisely the payoff of a European call option.

This insight is crucial because it allows us to use the powerful tools of option pricing theory, which you are familiar with from the Black-Scholes-Merton module, to analyze credit risk.

The following reading provides a clear explanation of this concept with simple numerical examples.

Estimating Default Probabilities - FRM Part 2 Study Notes

This text from AnalystPrep introduces the Merton model's core concept of viewing a firm's equity as a call option on its assets.

Please read the section "Using the Merton Model for Firm Valuation". Pay close attention to the payoff formulas for both equity and debt, and work through the two brief examples provided. This will solidify your understanding of the option analogy.

From Asset Value to Default Probability

Now, let's connect this back to default. In the Merton model, default happens if, and only if, the value of the firm's assets at maturity $T$ is less than the face value of its debt $D$ . Our goal is to calculate the probability of this event: $P (V_{T} \leq D)$ .

To do this, we make a familiar assumption: the firm's asset value, $V_{t}$ , follows a Geometric Brownian Motion (GBM). As you know from our work on stochastic processes, this implies that the asset value at a future time $T$ , $V_{T}$ , follows a lognormal distribution.

This is best understood visually.

This graph illustrates the Merton model. The firm's asset value follows a stochastic path. At time T, its value is a random variable with a lognormal probability distribution. Default occurs if the asset value falls into the shaded region, i.e., below the value of its liabilities (the default point).

The key question becomes: given where the asset value is today ( $V_{0}$ ), what is the probability it will end up in that shaded default region at time $T$ ? To answer this, we calculate a metric called the Distance to Default (DD).

The video below provides an excellent intuitive and graphical explanation of how the model works.

FRM: Expected default frequency (EDF, PD) with Merton Model

This video, "Expected default frequency (EDF, PD) with Merton Model" from Bionic Turtle, walks through the conceptual and graphical basis of the Merton model.

Please watch the following segments: Introduction to the structural approach A graphical explanation Introducing the D2 formula Focus on how the model combines the initial "cushion" (the difference between assets and liabilities) with the expected growth and volatility of the assets to determine the likelihood of default.

As the video explains, the Distance to Default measures how many standard deviations the expected asset value is from the default point. In the context of the Black-Scholes-Merton framework, this is given by the term $d_{2}$ .

Calculating the Probability of Default

The formula for the Distance to Default (DD) is:

DD = d_{2} = \frac{ln ( V _{0} / D ) + ( r - \frac{1}{2} σ _{V}^{2} ) T}{σ _{V} T}

Let's dissect this formula, which should look very familiar from the Black-Scholes model:

$V_{0}$ : The current market value of the firm's assets.
$D$ : The face value of the firm's zero-coupon debt (the default point).
$r$ : The risk-free interest rate.
$σ_{V}$ : The volatility of the firm's assets.
$T$ : The time to maturity of the debt.
Numerator: This represents the expected log-return of the assets over the debt level, under the risk-neutral measure. It's the "cushion" the firm has before defaulting.
Denominator: This is the total volatility (standard deviation) of the asset value over the time horizon $T$ .

By dividing the cushion by the volatility, we standardize the distance into units of standard deviations.

Once we have the Distance to Default, calculating the probability of default is straightforward. Since asset values are assumed to be lognormally distributed, the standardized distance $d_{2}$ can be related to the standard normal distribution. The probability of default is the probability that the asset value falls below the debt level $D$ . This corresponds to the area in the left tail of the distribution.

The probability of default (PD) is given by:

PD = N (- d_{2}) = N (- DD)

Where $N (\cdot)$ is the cumulative distribution function (CDF) for a standard normal distribution. Since $d_{2}$ measures how many standard deviations the mean is above the default point, $- d_{2}$ represents the position of the default point on a standard normal scale. $N (- d_{2})$ gives us the cumulative probability up to that point—the area of the tail representing default.

The following reading formally presents these formulas.

Estimating Default Probabilities - FRM Part 2 Study Notes

This section of the AnalystPrep article provides the precise formulas for Distance to Default and the resulting Probability of Default.

Please read the section titled "Distance to Default". This will confirm the formulas we've just discussed.

A Practical Calculation

Let's put this all together with a numerical example. The key challenge in applying the Merton model in the real world is that the firm's asset value $V_{0}$ and its volatility $σ_{V}$ are not directly observable from market data. We can only see the company's stock price (equity value $E_{0}$ ) and its volatility ( $σ_{E}$ ). In practice, analysts use a system of two simultaneous equations (the Black-Scholes formula for the call option and another relating the volatilities) to solve for the unobserved $V_{0}$ and $σ_{V}$ .

For exam purposes, however, you will typically be given the necessary inputs directly. The following video demonstrates a complete calculation in a spreadsheet, assuming the inputs are known.

FRM: Expected default frequency (EDF, PD) with Merton Model

Let's return to the Bionic Turtle video to see a step-by-step calculation.

Watch the Excel demonstration from the spreadsheet example. Follow how the presenter takes the inputs and applies the d_2 formula to calculate the Distance to Default, and then uses the NORMSDIST function in Excel (which is N(\cdot)) to find the final probability of default. Note that the video uses slightly different notation with an expected return \mu instead of the risk-free rate r. For risk-neutral pricing and consistency with the BSM framework, we use r. The mechanics are identical.

Here is a summary of the calculation steps from the video, which you can replicate:

Problem:
A firm has total assets valued at $1,000. It has debt with a face value of $600 maturing in 1 year. The risk-free rate is 5% and the asset volatility is 25% per annum. Calculate the probability of default.

Inputs:

$V_{0} = 1000$
$D = 600$
$r = 0.05$
$σ_{V} = 0.25$
$T = 1$

Step 1: Calculate the Distance to Default ( $d_{2}$ )

d_{2} = \frac{ln ( 1000/600 ) + ( 0.05 - \frac{1}{2} \cdot 0.2 5 ^{2} ) \cdot 1}{0.25 1}

d_{2} = \frac{ln ( 1.6667 ) + ( 0.05 - 0.03125 )}{0.25}

d_{2} = \frac{0.5108 + 0.01875}{0.25} = \frac{0.52955}{0.25} = 2.1182

Step 2: Calculate the Probability of Default (PD)

PD = N (- d_{2}) = N (- 2.1182)

Using a standard normal distribution table or a calculator function, we find:

PD \approx 0.0171 or 1.71%

This means, according to the Merton model, there is a 1.71% chance that the firm's asset value will be below its debt obligations in one year, causing a default.

Conclusion

In this lesson, we have delved into our first quantitative model for credit risk. By ingeniously reframing a firm's capital structure in terms of option theory, the Merton model provides a powerful and intuitive method for estimating the probability of default.

Key Takeaways:

The Merton model is a structural model that links default risk to a firm's asset and liability structure.
It views a firm's equity as a European call option on its assets, with the face value of debt acting as the strike price.
Default occurs if the asset value at maturity is less than the debt, which is analogous to the call option expiring out-of-the-money.
The Distance to Default (DD), given by the $d_{2}$ parameter from the Black-Scholes-Merton formula, measures the firm's financial cushion in units of standard deviation.
The Probability of Default (PD) is then calculated as $N (- DD)$ , where $N (\cdot)$ is the standard normal cumulative distribution function.

While the Merton model is elegant, its simple assumptions (e.g., single zero-coupon debt, no dividends, lognormal asset values) have limitations. Commercial models like Moody's KMV are based on Merton's framework but include practical modifications.

Having explored a structural model, our next lesson will introduce an alternative paradigm: reduced-form models. These models, also known as intensity-based models, treat default as an unpredictable, exogenous event, modeling its arrival with a statistical process rather than deriving it from the firm's balance sheet.