1) The Merton model simulates asset price paths over time to model default risk, with the loss given default calculated as the difference between debt face value and asset value at maturity.
2) The model can be used to derive a bond's credit spread by treating the bond as a risk-free bond minus a put option on the underlying asset's value.
3) While convenient, the Merton model has shortcomings as it predicts credit spreads that are often too low compared to real-world data and that decline over time contrary to market observations.
Download as TXT, PDF, TXT or read online on Scribd
0 ratings0% found this document useful (0 votes)
35 views
Lecture 6 Script
1) The Merton model simulates asset price paths over time to model default risk, with the loss given default calculated as the difference between debt face value and asset value at maturity.
2) The model can be used to derive a bond's credit spread by treating the bond as a risk-free bond minus a put option on the underlying asset's value.
3) While convenient, the Merton model has shortcomings as it predicts credit spreads that are often too low compared to real-world data and that decline over time contrary to market observations.
Download as TXT, PDF, TXT or read online on Scribd
You are on page 1/ 3
PROFESSOR: A more physical sense of what's
going on in the Merton model is obtained
by considering what happens along a single sample path for asset value over time. In this example here, the face value of debt is the solid red line, and the asset value winds up below that value at the debt maturity date. The loss given default on that path is just the difference between the face value and the asset value on that maturity date. You can think of the model is equivalent to running many risk neutral Monte Carlo paths of lognormal model for asset prices and calculating the present value of the expected loss given default by averaging across those many paths. With the predicted bond price from the Merton model, it's straightforward to derive the corresponding credit spread. We start with the bond price D0 and find the yield to maturity on the bond on a continuous basis from this formula here. Substituting in the expression for D0 that's based on the price of a risk free bond minus the price of a put option, we can rearrange and solve for the credit spread, which is given here. As you'd expect, the spread gets wider the more valuable the put option, which of course equals the value of the expected losses.
How the credits spread moves over time
depends on the model parameters, and it's especially sensitive to asset volatility and the leverage ratio, D divided by A. These graphs show the relationship for two different levels of assumed asset volatility and for a range of leverage ratios. The graphs reveal some counterfactual predictions of the model. One is that the credit spread is close to 0 when the bond is nearing maturity. The technical reason for that is because the price of the asset is assumed to follow a continuous path, and there's essentially no chance over a short horizon of a price move that is adverse enough for the possibility of loss to become significant. A second problem is that the predicted magnitudes of the credit spreads are far lower than those observed in the data. As I mentioned earlier, one reason for that is that the empirical credit spread reflects more than just the direct cost of default losses. Finally, the pattern of sharply declining credit spreads over time for some parameterizations is also contrary to what is observed in the market. The reason in the model for the declining spreads is that assets on average grow and the face value of the debt is fixed, which has the effect of reducing the likelihood of default over time as the leverage ratio tends to shrink. Despite these shortcomings, it's convenient to use this model to develop an understanding of the relative pricing of some more complicated debt structures, and we'll turn to that next. We'll also discuss some modifications to the basic model that address some of these shortcomings and that produce more reliable price estimates. Corporations often issue a combination of senior and junior debt. Doing that may lower the total cost of debt financing by creating securities that appeal to clienteles with different tolerances for risk and return. In the event of default, the senior debt is fully repaid before the junior debt holders receive any payments. Both groups of debt holders must be repaid before any value goes to the equity holders. To take a simple example, imagine that a firm has two issues of zero-coupon bonds outstanding, both maturing at the same time Cap(t). Fs is the face value of the senior debt, and Fj is the face value of the junior debt. This table describes the payoffs to the different claimants as a function of the asset value at maturity, V sub t. If the asset value is less than the face value of the senior debt, the entire value goes to pay the senior debt holders. If the asset value is enough to pay the senior debt holders but less than the sum of the face value of the junior and the senior debt, then the junior debt holders get the remaining value after the senior debt holders are paid in full. Finally, if the asset value is greater than the sum of the face value of the junior and senior debt, there isn't a default. The bondholders are paid in full, and any residual asset value goes to the equity holders. That description of the payoffs can be translated into expressions that are in terms of the value of call options on the underlying assets of the firm. The senior debt holders have a position that's equivalent to owning the assets of the firm and writing a call option.
That call option has a strike price
equal to the face value of their debt. Then the junior debt holders are long that same call option, and they're short another call option with the strike price equal to the sum of the face value of the senior and the junior debt. Finally, as we know, the equity holders are long the call option written by the junior debt holders, which has a strike price of the sum of the face values of the debt. For pricing more complicated structured claims, the stochastic model for asset prices can be implemented on a risk neutral binomial tree or with Monte Carlo that can then be used for finding and pricing the payoffs to different claimants.