Default Prediction for Companies


CSE691 Computational Finance


Spring 2004


Introduction

Certain fundamental analysis models can be to predict the probability that a company will default on its debt or go bankrupt in a given period of time. This project seeks to analyze and test models that predict the default probabilities for each publicly traded company and derive measures of corporate health.The data used was that provided by corporate balance sheets and income statements reported on http://finance.yahoo.com. More comprehensive financial data is usually confidential propriety information and is not available freely.

I have studied and analyzed the following prediction models and applied them to our current database of 4000 stocks. I have also done a separate analysis for the S&P 500 stocks:

  • Z - Score Model
  • Vaseick Kealhofer Model(EDF Credit measure)
  • Black - Scholes Model


  • I have also tried to correlate the different strategies and come up with a unified view towards predicting default.

    This project was done under the guidance of Prof.Steven Skiena as part of the course on Computational Finance taken at the Computer Science Department,Stony Brook University in Spring 2004.

    Initial background work

    Click here to see the initial project proposal.
    Click here to see the first presentation.
    Click here to see the second presentation.

    The Z Score Model

    This is one of the venerable models for assessing the distress of industrial corporations. The unique characteristics of business failures are examined in order to specify and quantify the variables which are effective indicators and predictors of corporate distress. Specifically, a set of financial and economic ratios are analyzed in a corporate distress prediction context using a multiple discriminant statistical methodology.

    Click here for the theory and results.

    Expected Default Frequency Credit measure(EDF)

    There are three steps required to calculate an EDF measure : (1) Estimate the surrent market value and volatility of the firm's assets (2) Determine how far the firm is from default, i.e its distance - to - default (3) Scale the distance - to - default to a probability.

    Click here for the theory and results

    Black-Scholes Model

    The model is one of the most widely accepted of all financial models, right from its inception in 1973 and has its roots in Option Pricing Theory. The model assumes that the firm's equity is a perpetual option with the default point acting as an absorbing barrier for the firm's asset value. When the value hits the default point, the firm is assumed to default.

    Click here for the theory and results

    Common Analysis

    Pairwise Correlation analysis to see how well correlated the bankruptcy prediction is for all S&P 500 stocks by each pair of methods. Click here.

    Measuring the consistency of prediction across time. Click here.

    Break down of analysis by industrial sector. Click here.

    Break down of analysis by company size(market cap) and correlating the default probability with market capiltaization. Click here.

    Miscellaneous

    I have also created a separate database for the S&P 500 Balance Sheet information and Income Statement information from the existing database.

    Click here to create the database.

    References and Literature Survey

  • The Informational Content And Accuracy of Implied Asset Volatility as a Measure of Total Firm Risk.
  • Working Paper A Model For Bankruptcy Prediction
  • Assessing the probability of Bankruptcy
  • Predicting Financial Distress of Companies : The Z-Score and ZETA Models
  • Capital Structure and the prediction of Bankruptcy
  • Bayesian Analysis of Duration Models : An Application to Chapter 11 Bankruptcy
  • Correlated Default Risk
  • Debt Valuation with Endogenous Default and Chapter 11 Reorgnization
  • Choosing Bankruptcy Predictors Using Discriminant Analysis, Logit Analysis, and Genetic Algorithms
  • Predicting Corporate Financial Distress : A Time-Series CUSUM Methodology
  • Forecasting Bankruptcy More Accurately: A Simple Hazard Model
  • Multi - Period Corporate Failure Prediction With Stochastic Covariates
  • The Choice Among Traditional Chapter 11, Prepackaged Bankrupcty, and Out-of-Court Restructuring
  • The Link between Default and Recovery Rates: Theory,Empirical Evidence and Implications
  • Research Insights - The Barra Credit Series: Forecasting Default in the Face of Uncertainity
  • finance.yahoo.com
  • Options,Futures and Other Derivatives 5th Edition John C.Hull