Forecasting default with the Merton distance to default model

Sreedhar Bharath, Tyler Shumway

Research output: Contribution to journalArticle

460 Citations (Scopus)

Abstract

We examine the accuracy and contribution of the Merton distance to default (DD) model, which is based on Merton's (1974) bond pricing model. We compare the model to a quot;naïve" alternative, which uses the functional form suggested by the Merton model but does not solve the model for an implied probability of default. We find that the naïve predictor performs slightly better in hazard models and in out-of-sample forecasts than both the Merton DD model and a reduced-form model that uses the same inputs. Several other forecasting variables are also important predictors, and fitted values from an expanded hazard model outperform Merton DD default probabilities out of sample. Implied default probabilities from credit default swaps and corporate bond yield spreads are only weakly correlated with Merton DD probabilities after adjusting for agency ratings and bond characteristics. We conclude that while the Merton DD model does not produce a sufficient statistic for the probability of default, its functional form is useful for forecasting defaults.

Original languageEnglish (US)
Pages (from-to)1339-1369
Number of pages31
JournalReview of Financial Studies
Volume21
Issue number3
DOIs
StatePublished - May 2008
Externally publishedYes

Fingerprint

Default probability
Probability of default
Hazard models
Predictors
Functional form
Merton model
Bond yields
Out-of-sample forecasting
Sufficient statistics
Bond pricing
Credit default swaps
Reduced-form model
Rating agencies
Corporate bonds
Yield spread

ASJC Scopus subject areas

  • Finance
  • Accounting
  • Economics and Econometrics

Cite this

Forecasting default with the Merton distance to default model. / Bharath, Sreedhar; Shumway, Tyler.

In: Review of Financial Studies, Vol. 21, No. 3, 05.2008, p. 1339-1369.

Research output: Contribution to journalArticle

@article{4b7348dc44414e5ebfe57198b794bbc9,
title = "Forecasting default with the Merton distance to default model",
abstract = "We examine the accuracy and contribution of the Merton distance to default (DD) model, which is based on Merton's (1974) bond pricing model. We compare the model to a quot;na{\"i}ve{"} alternative, which uses the functional form suggested by the Merton model but does not solve the model for an implied probability of default. We find that the na{\"i}ve predictor performs slightly better in hazard models and in out-of-sample forecasts than both the Merton DD model and a reduced-form model that uses the same inputs. Several other forecasting variables are also important predictors, and fitted values from an expanded hazard model outperform Merton DD default probabilities out of sample. Implied default probabilities from credit default swaps and corporate bond yield spreads are only weakly correlated with Merton DD probabilities after adjusting for agency ratings and bond characteristics. We conclude that while the Merton DD model does not produce a sufficient statistic for the probability of default, its functional form is useful for forecasting defaults.",
author = "Sreedhar Bharath and Tyler Shumway",
year = "2008",
month = "5",
doi = "10.1093/rfs/hhn044",
language = "English (US)",
volume = "21",
pages = "1339--1369",
journal = "Review of Financial Studies",
issn = "0893-9454",
publisher = "Oxford University Press",
number = "3",

}

TY - JOUR

T1 - Forecasting default with the Merton distance to default model

AU - Bharath, Sreedhar

AU - Shumway, Tyler

PY - 2008/5

Y1 - 2008/5

N2 - We examine the accuracy and contribution of the Merton distance to default (DD) model, which is based on Merton's (1974) bond pricing model. We compare the model to a quot;naïve" alternative, which uses the functional form suggested by the Merton model but does not solve the model for an implied probability of default. We find that the naïve predictor performs slightly better in hazard models and in out-of-sample forecasts than both the Merton DD model and a reduced-form model that uses the same inputs. Several other forecasting variables are also important predictors, and fitted values from an expanded hazard model outperform Merton DD default probabilities out of sample. Implied default probabilities from credit default swaps and corporate bond yield spreads are only weakly correlated with Merton DD probabilities after adjusting for agency ratings and bond characteristics. We conclude that while the Merton DD model does not produce a sufficient statistic for the probability of default, its functional form is useful for forecasting defaults.

AB - We examine the accuracy and contribution of the Merton distance to default (DD) model, which is based on Merton's (1974) bond pricing model. We compare the model to a quot;naïve" alternative, which uses the functional form suggested by the Merton model but does not solve the model for an implied probability of default. We find that the naïve predictor performs slightly better in hazard models and in out-of-sample forecasts than both the Merton DD model and a reduced-form model that uses the same inputs. Several other forecasting variables are also important predictors, and fitted values from an expanded hazard model outperform Merton DD default probabilities out of sample. Implied default probabilities from credit default swaps and corporate bond yield spreads are only weakly correlated with Merton DD probabilities after adjusting for agency ratings and bond characteristics. We conclude that while the Merton DD model does not produce a sufficient statistic for the probability of default, its functional form is useful for forecasting defaults.

UR - http://www.scopus.com/inward/record.url?scp=39749084409&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=39749084409&partnerID=8YFLogxK

U2 - 10.1093/rfs/hhn044

DO - 10.1093/rfs/hhn044

M3 - Article

VL - 21

SP - 1339

EP - 1369

JO - Review of Financial Studies

JF - Review of Financial Studies

SN - 0893-9454

IS - 3

ER -