PRESS RELEASE 2 MARCH 2017
Study confirms: OK-Score Model systematically warns for Business Failures
European Rating House as external compliance officers for the years 2013 until 2016 has now published the 4th edition of “Assessing the Accuracy of the OK-Score Model”. The portfolio consisted end 2016 of 2,741 credit scores dating from the years 2000-2016, containing 70 cases of Business Failure. The OK-Score Model has issued a warning on 70 occasions. There was only one false warning on Wolters Kluwer in 2009 (false warning error 1.4%; Type-1 error 0.04%). Only one Business Failure remained undetected: WorldCom in 2000 (non-detection error/Type-2 error: 1.4%). The report (LINK) identifies every one of the 70 business failures: name, year and type of failure.
Business Failure is the phase in which a company is forced to take drastic measures: asset stripping, turnaround, forced recapitalization or forced sale, while Chapter 11 and bankruptcy cannot be excluded. The OK-Score Model warns one to three years in advance.
European Rating House, has started the accuracy assessments in 2013. The portfolio then consisted of some 2,000 credit scores. After the first assessment report came out in September 2013, the portfolio has grown by more than 700 credit scores until 2714 at the end of 2016. The error-levels have remained extremely low: 1.4% down from 2.2%.
The official gatekeepers (credit rating agencies and accountants) have extreme difficulties in getting their predictions right:
- Credit rating agencies don’t warn for two-thirds of the defaults, while three out of every four warnings they issue is false. ‘Default’ is a narrow section of the Business Failure spectrum, and credit rating agencies are supposed to warn 0-12 months in advance. This is almost an impossible job.
- Accountants fail to provide investors with reliable warnings for going concern issues. When going concern issues arise, it appears that the accountants had not warned in more than half of the cases. Their false-warning level even exceeds 80%. ‘Going concern’ is a difficult concept, and reports must be issued 0-12 months in advance. This is an extremely tough job.
The conclusion must be that investors should rely on the OK-Score Model rather than on the mandatory reports from credit rating agencies and accountants.