commonly used to identify companies that may be committing fraud or, more politely, “managing earnings.”
Each of these three approaches—accounting-based, marketbased and fraud-based—is at least moderately effective on a
stand-alone basis in predicting bankruptcies. Since the models are not assumed to be highly correlated, each model has
the potential to add incremental value to the other models.
To validate bankruptcy models, an Accuracy Ratio test is typically performed. This widely accepted validation test is
designed to measure the predictive power of a model and to
compare different models. An Accuracy Ratio of 100% indicates a perfect model, and an Accuracy Ratio of 0% indicates
a model with no predictive power: the higher the percentage,
the closer the approximation to a perfect model.
Though not designed specifically as a bankruptcy predictor,
the stand-alone Accuracy Ratio for the AGR Model is 45%.
The Accuracy Ratio for the Ohlson Model is 77% and for the
Merton DD model it is 86%, which are consistent with
numerous academic studies.
To determine whether a better approach would improve
bankruptcy prediction, the three models above were combined in a single model, using regression techniques to determine which factors in combination best predict bankruptcies.
The resulting model was found to have an Accuracy Ratio of
91%, indicating a highly-predictive model.
Beyond adjusting current models, credit officers can institute
other strategies to gauge risk such as:
•
Increase the frequency of model updates. Some models
only utilize year-end data, which cannot react quickly
enough in today’s environment;
•
Add in market-based information to get a timely perspective on how the marketplace is assessing risk. Significant
stock declines and excess volatility indicate heightened
bankruptcy risk;
•
Incorporate forensic accounting or other non-traditional
bankruptcy risk approaches. An overreliance on fundamental research can create a false sense of security. Certain
key forensic accounting and corporate governance
measures have been found to identify companies where
fundamental data cannot be relied on.
Whether an actual bankruptcy filing occurs or not, companies
with high bankruptcy risk are more likely to suffer losses,
defaults, restructuring, asset sales, downsizing, equity dilution
and other events damaging to creditors. And analysis of past
Whether an actual bankruptcy filing
occurs or not, companies with high
bankruptcy risk are more likely to suffer
losses, defaults, restructuring, asset
sales, downsizing, equity dilution and
other events damaging to creditors.
bankruptcy cycles indicates that bankruptcy filings lag an economic rebound by up to a year, which may mean further
increases in bankruptcies in the quarters ahead, as companies
weakened by the recession struggle to survive a still-challenging environment.
As the economy recovers from a deep, painful recession, many
companies continue to be vulnerable to bankruptcy risk and
financial distress. New approaches to identifying and monitoring this key risk should be considered, particularly in light
of the changing nature of bankruptcy risk. ●
Jack Zwingli is CEO of Audit Integrity, an independent research firm
that measures and evaluates the integrity and transparency of
corporate accounting and governance practices. He may be reached at
[email protected] or 310-444-8820.
Times Are Difficult Enough Without Having
to Deal With Increased Levels of Fraud
LA L