Forecasting Corporate Failure: Understanding Statistical and Theoretical Approaches to Bankruptcy Prediction

The application of bankruptcy prediction models in litigation to evaluate whether a company was in default, or about to default, at a particular time can be helpful in a wide range of contexts when used appropriately, according to a recent AIRA Journal article authored by Managing Principal Andrew Wong and Konstantin Danilov. In "Forecasting Corporate Failure: Understanding Statistical and Theoretical Approaches to Bankruptcy Prediction" (Vol. 29, No. 1, 2015), the authors describe two categories of bankruptcy prediction models -- statistical and theoretical -- and examine their use in litigation. "Statistical models attempt to identify the most common symptoms exhibited by bankrupt companies, and then use this information to estimate the likelihood that a particular firm will go bankrupt in the future," the authors explain. "Conversely, theoretical models predict bankruptcy by attempting to identify and gauge the factors responsible for the causes of bankruptcy."

Noting that understanding "the strengths and weaknesses of these approaches is helpful when deciding which particular prediction model to apply in a bankruptcy setting," the authors describe the history of each model, explore their predictive performance and limitations, and examine several case examples. The authors conclude that, because of "differences in rationale, effectiveness, and applicability, each model is uniquely suited" to determine the likelihood that a company will go bankrupt. "As such, the applicability and validity of any approach will always depend on the context and the details of the particular case."