Over the past quarter century, Enron, Lehman Brothers, General Electric, and Supermicro have all been notorious for financial mismanagement, especially the malpractice of “cooking the books.” Despite this, such deceptions often go undetected until they have cost investors billions. Analysts have thus been compelled to develop methods to detect companies employing dubious or outright illegal accounting strategies to obscure their underperformance.

Urooj Khan, an accounting professor and Deloitte & Touche Centennial Faculty Fellow at Texas McCombs, has introduced pioneering research offering a novel and more effective method for assessing what he terms “earnings quality”. His innovation, the earnings quality score (EQSCORE), has demonstrated remarkable success in surpassing the capabilities of existing models at pinpointing firms potentially engaged in accounting misconduct.

The strength of EQSCORE lies in its comprehensive approach. Khan highlights that, unlike other models, EQSCORE considers an extensive array of accounting red flags and integrates considerations such as governance issues—for instance, the potential bias of an insider board chair—along with other discrepancies identified by financial auditors.

Khan’s research genesis was aimed at closing the disconnect between the theoretical models of earnings quality discussed in academic circles and the practical applications used by Wall Street. This led to the creation of a more robust model that reflects a wider spectrum of factors influencing company performance and the accuracy of financial reports.

In developing EQSCORE, Khan collaborated with Venkat Peddireddy of the China Europe International Business School and Shiva Rajgopal of Columbia University. They meticulously analysed 613 detailed reports from a leading private research firm highlighting various companies involved in potentially illegal or questionable financial reporting. These reports revealed critical metrics and “signals” used to identify 230 companies, with the majority of indicators related to accounting practices like questionable revenue figures and profit margins, as well as concerns related to corporate governance and auditing.

The team identified 51 red flags from these analyses, which formed the basis of EQSCORE. These included 31 accounting variables, eight about board characteristics, and 11 related to audit issues. To validate EQSCORE, the researchers applied it to a different dataset—companies targeted by Securities and Exchange Commission (SEC) enforcement actions. While the SEC’s dataset does not encompass all instances of suspect corporate accounting, it does capture the most egregious cases, particularly given the SEC’s constrained budget, which necessitates selecting cases with substantial evidence of manipulation.

EQSCORE was developed using data from companies subjected to SEC enforcement actions between 2004 and 2009. Its efficacy was then tested against a sample of companies charged by the SEC from 2010 to 2016. The results were compelling; EQSCORE successfully predicted 71% of the accounting years investigated by the SEC, a significant improvement over the next best model, which only predicted 55%. Additionally, it missed only 29% of the years flagged by the SEC, compared to 45% for the competing model.

Khan is optimistic that EQSCORE could garner interest from the SEC, which has been developing its model since 2012 to detect fraud and accounting irregularities. The model could also be an invaluable tool for private investors, aiding them in identifying companies that tread close to, but do not necessarily cross, the legal boundaries of accounting practices.

To illustrate the potential benefits for investors, the researchers devised a stock market strategy based on EQSCORE. They simulated investment scenarios where they bought stocks of companies with low EQSCOREs and shorted those with high scores, betting that stocks with high scores would decline. Over seven years, this strategy yielded average annual returns 7.8% higher than expected.

In conclusion, Khan asserts that EQSCORE significantly enhances the ability to predict fraudulent activities and effectively identifies firms that resort to aggressive accounting tactics to misrepresent their proper financial health, mainly when they are in a state of decline.

More information: Urooj Khan et al, Earnings quality on the street, Contemporary Accounting Research. DOI: 10.1111/1911-3846.12975

Journal information: Contemporary Accounting Research Provided by University of Texas at Austin