Skip to main content
News_

Predicting corporate bankruptcy

28 September 2017
Existing models just aren't cutting it
Professor Stewart Jones is using new and more sophisticated machine-learning technologies that can translate big data into highly accurate predictions.

Research highlights with Professor Stewart Jones

Professor Stewart Jones says the new technologies are set to benefit banks, companies and investors, as their outputs can assist with making the right choices around lending money, managing business operations and making investments.

This new research will help a variety of stakeholders better understand how and why companies fail. "We can learn from those failings and make business stronger in the future," Professor Jones says.

I think the US market is at some points going to falter. It’s hard to judge how serious the correction will be. Hopefully it won’t be a GFC 2, but many of the issues that created the first GFC are re-emerging.
Professor Stewart Jones

As examples, Professor Jones points to the underregulated over-the-counter derivatives market which has again risen in value to more than $US500 trillion globally and the rapidly expanding trillion-dollar student loan market in the US.

He says the hedge funds and investment banks are now "cutting and dicing up these loans into collateralised loan obligations".

"They’re being rated by the credit rating agencies and it seems to be a replay of what happened during the global financial crisis," he says. "With default rates on student loans at around 11 per cent and with a much higher number in a deferral or repayment program it’s looking like an overvalued market."

Financial modelling in the field of credit risk and bankruptcy forecasting, developed by Professor Jones, is also pointing toward a downturn.

"Since the GFC, there’s been a lot more concern about the measurement of risk, the pricing of risk using more sophisticated models for prediction," Professor Jones says. "I'm working in more advanced technologies, such as machine learning technologies, which are known to be a lot more accurate in the prediction of risk events."

Published work

Jones S, Johnstone D and Wilson R 2017 'Predicting Corporate Bankruptcy: An Evaluation of Alternative Statistical Models', Journal of Business Finance and Accounting, vol.44:1-2, pp. 3-34

Jones S, 2017 'Corporate Bankruptcy Prediction: A High Dimensional Analysis', Review of Accounting Studies, vol.22:3, pp. 1366-422