Business Failure Prediction Models

Using various statistical techniques, business failure prediction models attempt to estimate the bankruptcy probability of a firm using a set of covariates such as financial ratios, market-related variables, or the type of industry. One powerful method is survival analysis. A non-parametric survival analysis technique is used to estimate survival and hazard functions. A semi-parametric Cox proportional hazard (PH) model is used to evaluate the effect of covariates on bankruptcy probability and duration of time before bankruptcy. In addition, discriminant and logit analyses are applied to predict the success or failure of a business.

Click to see our SAS Paper our recent Working Paper or a Powerpoint Presentation of our work

Multiple Factor Risk Models

In increasingly complex global markets, firms are faced with multiple sources of financial risk in their operations. Whether it is credit risk or market risk, it is crucial to identify and effectively manage the underlying risk. Often market risk arises in portfolios consisting of highly correlated market indices or interest rates. One powerful statistical method used in managing market risk is a principal components or factor analysis.

Monte Carlo Simulation

Monte Carlo methods can be applied to ROI analysis (to calculate the probability of a particular IRR or expected NPV) and Value-at-Risk (to simulate the distribution of potential losses for a portfolio or the default probability). In addition, the MC method is a powerful tool in the pricing of options.

Click Here to see our Powerpoint presentation on option pricing under stochastic interest rate changes.