Few irregularities detected by these models ever rise to criminal activity and in fact most red flags are subsequently corrected and reflect major internal changes of a particularly company that can cause such irregularities to show up in the forensic algorithms.
Top forensic algorithms that are used to detect fraud and bankruptcy do not show any evidence of financial irregularities in Infosys’ reported numbers from 2013 to 2019, a US-based Seeking Alpha analyst said in an interview with Moneycontrol on October 23.
These models include Altman Z-Score, Beneish M-score, Ohlson O-score probability and the Montier C-score. JD Henning is an investment adviser, fraud examiner and certified anti-money laundering specialist with more than 25 years experience in trading and investing stocks. He is the founder of Value & Momentum Breakouts.
“The time period evaluation of these forensic and value algorithms does not show significant or prolonged risks of financial irregularities that may relate to the whistleblowers' complaints,” JD Henning, an investment advisor with V & M Breakouts said in the Seeking Alpha article.
Here are the excerpts from the interview:
Q. How long have you tracked Infosys? Are you an investor, do you have Infosys shares?
I do not have any shares or any association with Infosys, and I have no intention to buy their shares at this time. I have only focused on Infosys in the past week as I regularly analyse whistleblower claims against companies to see if any clear public financial data can support or refute these allegations.
Q. What do you make of Infosys’ image in terms of corporate governance?
Based on the reports and allegations, it appears that there are credible claims against Infosys that are impairing the company's corporate governance image. The SEC is starting a probe on Infosys over the whistleblower claims, and we will soon know more about the severity and credibility of these allegations. An investigation is warranted to settle these allegations and restore credibility, if possible, that the company is being run properly.
Q. Based on the numbers you analysed and the signals you picked from their reported financial data, what would you summarise? Could there be something wrong indeed?
My summary is in the conclusion of my published article, and I have specified in it that there are some areas that would be worthwhile for further investigation. The models do not rule out all forms of potential malfeasance and I do not know the specific nature of the whistleblowers' claims nor the exact time periods where the alleged wrongdoing may have occurred. Without more information at hand, it is not a good practice for a fraud examiner to speculate further on what may match the complaints that he has not seen. As more information is released, readers can apply it to the scores and values provided across seven years of data to see how well any irregularities may fit the data.
Q. The four models you applied, how well can they hold? How reliable are they?
To the best of my knowledge, all the models, except for the Montier-C model, have been peer-review tested by financial scholars for many years and are widely trusted. The Beneish M-score in particular has been very effective in identifying and confirming major fraud activities in publicly traded firms like Enron and Worldcom long before any wrongdoing was identified and prosecuted.
These models are accepted and frequently applied by certified fraud examiners like myself, who use algorithms to narrow in on areas with irregularities and questionable scores for further analysis. They are more effective by being used in combination to cover as many data points as possible to help validate scores and detail certain areas of financial reporting that may be in doubt. As I stated in my article, these algorithms are not fool-proof and can miss fraudulent activity, but they do significantly enhance the probability of detection.
Q. Have you applied similar models to understand any other past accusations or whistleblower complaint for any other company? Can you share the details ?
Yes, I regularly apply these models as a certified fraud examiner (CFE) and anti-money laundering specialist (CAMS) to calibrate and test the algorithms as different whistleblower claims arise in the financial markets.
Most recently I ran tests on claims against General Electric and Disney that are linked in my article and detail potential areas of concern. In the case of GE, I did not find strong evidence of wrongdoing or in the magnitude of over $30 billion as alleged by the whistleblower. I did find that there were some significant financial concerns for GE which are being aggressively addressed to shore up operations and improve the health of the business under the relatively new CEO.
Regarding Disney, the claims of $6 billion in revenue manipulation did not bear out in the analySed financial models in any time period and were far too large to be missed if the claims were completely accurate.
More details on the analysis of those top companies are in the published articles linked in my Infosys article. So far, the accusations on the different companies that I have previously examined have not measured up to their original claims, but some have shown some financial risk for the companies that have subsequently been addressed and corrected with aggressive management changes.Few irregularities detected by these models ever rise to criminal activity and in fact most red flags are subsequently corrected and reflect major internal changes of a particularly company related to mergers, acquisitions, accounting changes, spin-offs, and other structural business activity that can cause such irregularities to show up in the forensic algorithms.Get access to India's fastest growing financial subscriptions service Moneycontrol Pro for as little as Rs 599 for first year. Use the code "GETPRO". Moneycontrol Pro offers you all the information you need for wealth creation including actionable investment ideas, independent research and insights & analysis For more information, check out the Moneycontrol website or mobile app.