The M-Score Explained

by: Red Team Analytics

Summary

The M-Score is a research-backed analytical tool shown to successfully detect earnings manipulation as well as predict stock returns.

Many of the existing public resources on the M-Score include inaccurate or misleading information, either utilizing older versions of the model or displaying a misunderstanding of its application.

Each of the variables making up the M-Score convey important information; understanding the meaning behind their values makes it a much more useful tool.

A company scoring highly on each variable of the M-Score would be one that is growing extremely quickly, experiencing deteriorating fundamentals, and adopting aggressive accounting practices.

In recent years, companies flagged as earnings manipulators by the M-Score have performed nearly 10% worse on average than non-flagged companies.

Introduction

In 1998, as Enron traded around $48/share, a group of business students at Cornell University's Johnson Graduate School of Management studied the company for a class project. Long before Enron peaked out at $90/share before plummeting to bankruptcy, the Cornell students rated Enron a "Sell" and wrote that "the 8-variable Beneish model shows that Enron may be manipulating its earnings." Needless to say, their findings were almost universally ignored.

In addition to Enron, the Beneish model has successfully predicted many of the more infamous and unexpected fraud cases involving public companies, including Adelphia Communications, Global Crossing, Qwest Communications International, and Cendant Corporation.

Developed by Dr. Messod Beneish, a professor of accounting at the Kelley School of Business in Indiana University, the Beneish model is a data-driven analytical tool designed to detect the likelihood of earnings manipulation. Further research has demonstrated that the Beneish model can also predict stock returns.

Unfortunately, many of the articles that have been written about the model include inaccurate or misleading information, either utilizing older versions of the model or displaying a misunderstanding of its application. This article is meant to explain the Beneish model and its output (the M-Score) in plain terms to allow investors to make the best use of this valuable tool.

The Formula

The Beneish model is an equation that produces an M-Score utilizing data readily available from any company's GAAP financial statements (NOTE: the Beneish model does not apply to financial services companies). According to Beneish's peer-reviewed research,[1] the higher a company's M-Score, the more likely it is that the company is manipulating its earnings. Specifically, when considering the entire market, the Beneish model classifies a company as an earnings manipulator if its M-Score is greater than -1.78. (NOTE: Several websites and articles refer to an M-Score threshold of -2.22 or mention an alternative 5-variable formula rather than the 8-variable formula described below; these refer to older versions of Dr. Beneish's model and are not supported by his current research).

M-Score = -4.84 + (0.920×DSR) + (0.528×GMI) + (0.404×AQI) + (0.892×SGI) + (0.115×DEPI) - (0.172×SGAI) + (4.679×ACCRUALS) - (0.327×LEVI)

Where

DSR (Days' Sales Receivables index) = (Current Year Receivables ÷ Sales) ÷ (Prior Year Receivables ÷ Sales)
GMI (Gross Margin Index) = (Prior Year Gross Margin) ÷ (Current Year Gross Margin]
AQI (Asset Quality Index) = (Current Year Noncurrent Assets except PPE ÷ Total Assets) ÷ (Prior Year Noncurrent Assets except PPE ÷ Total Assets)
SGI (Sales Growth Index) = Current Year Sales ÷ Prior Year Sales
DEPI (Depreciation Index) = [Prior Year Depreciation ÷ (Depreciation + PPE)] ÷ [Current Year Depreciation ÷ (Depreciation + PPE)]
SGAI (Sales, General, and Administrative expenses Index) = (Current Year SGA ÷ Sales) ÷ (Prior Year SGA ÷ Sales)
ACCRUALS (Accruals Index) = (Income Before Extraordinary Items - Cash from Operations) ÷ Total Assets
LEVI (Leverage Index) = [(Current Year Long-Term Debt + Current Liabilities) ÷ Total Assets] ÷ [(Prior Year Long-Term Debt + Current Liabilities) ÷ Total Assets]

The Variables Explained

Each of the variables in the Beneish model convey important information; understanding the meaning behind their values makes the Beneish model a much more useful tool. It should be noted that each component of the model is supported by a significant body of academic research into factors that can signal financial misrepresentations and fraud.

Each variable (except for ACCRUALS) is constructed so that a value of 1 is neutral; whereas any value above 1 makes it more likely that the firm will be an earnings manipulator and any value below 1 makes it less likely.

DSR (Days' Sales Receivables index) = (Current Year Receivables ÷ Sales) ÷ (Prior Year Receivables ÷ Sales)

A value for DSR > 1 indicates that the company's receivables are growing as a percentage of its sales year over year; the higher DSR is over 1, the more receivables have grown compared to sales. This can suggest either deteriorating economic conditions at a company (i.e. the company is incentivizing sales by extending more lenient credit terms to customers) or aggressive accounting (the company is recognizing revenue before it is earned). The relatively new practice of factoring or securitizing receivables can undermine the usefulness of this variable, so investors must be aware if a given company engages in this technique and adjust their data accordingly.

GMI (Gross Margin Index) = (Prior Year Gross Margin) ÷ (Current Year Gross Margin]

A value for GMI > 1 indicates that the company's gross margin is deteriorating year over year; the higher GMI is over 1, the more drastic the deterioration. Financial research shows that public companies with declining profit margins are more likely to manipulate their earnings.

AQI (Asset Quality Index) = (Current Year Noncurrent Assets except PPE ÷ Total Assets) ÷ (Prior Year Noncurrent Assets except PPE ÷ Total Assets)

A value for AQI > 1 indicates that the company's noncurrent assets (such as goodwill, intangibles, and other items of uncertain long-term value) are increasing as a percentage of all assets year over year; the higher AQI is over 1, the more these soft assets of uncertain quality have grown compared to the more tangible current assets. Financial research shows that this decrease in asset quality can be the result of excessive expenditure capitalization (deferring costs), and/or can be a sign of deteriorating fundamentals at the company.

SGI (Sales Growth Index) = Current Year Sales ÷ Prior Year Sales

A value for SGI > 1 indicates that the company's sales are growing year over year; the higher SGI is over 1, the greater the growth in sales. This variable is likely the most counterintuitive to the typical investor who is used to viewing increased sales strictly as a positive. Financial research suggests that the managers of public companies with very high sales growth may feel pressure to maintain a high level of sales growth, even if that means manipulating earnings to do so. Taken alone, high sales growth is by no means a negative; however, when a company with high sales growth also scores poorly on other variables measured by the Beneish model, it can be a significant warning sign to investors.

DEPI (Depreciation Index) = [Prior Year Depreciation ÷ (Depreciation + PPE)] ÷ [Current Year Depreciation ÷ (Depreciation + PPE)]

A value for DEPI > 1 indicates that the company's effective depreciation rate has slowed year over year; the higher DEPI is over 1, the more the depreciation rate has decreased. This suggests that the company has increased its estimate of its assets' useful lives, thereby slowing the recognition of expenses and (perhaps artificially) increasing the appearance of income.

SGAI (Sales, General, and Administrative expenses Index) = (Current Year SGA ÷ Sales) ÷ (Prior Year SGA ÷ Sales)

A value for SGAI > 1 indicates that the company's SGA expenses as a percentage of sales are increasing year over year, representing declining administrative and marketing efficiency; the higher SGAI is over 1, the more the SGA expenses have grown as a percentage of sales. Financial research shows that public companies may be more likely to engage in earnings manipulations to cover up deteriorating operational performance.

ACCRUALS (Accruals Index) = (Income Before Extraordinary Items - Cash from Operations) ÷ Total Assets

ACCRUALS measures the difference between accounting profit (or loss) and cash profit (or loss); the difference is then divided by total assets in order to make the variable comparable between companies of different sizes. Financial research shows that companies that generate large accruals (accounting-only profits) are likely to be engaging in earnings manipulation.

LEVI (Leverage Index) = [(Current Year Long-Term Debt + Current Liabilities) ÷ Total Assets] ÷ [(Prior Year Long-Term Debt + Current Liabilities) ÷ Total Assets]

A value for LEVI > 1 indicates that the company became more leveraged year over year; the higher LEVI is over 1, the more its leverage increased. An increasingly leveraged company tightens its debt constraints and predisposes it to manipulate earnings.

The Formula Explained

As explained by Dr. Beneish, a company scoring highly on each variable of the Beneish model would be one that is "growing extremely quickly, experiencing deteriorating fundamentals (as evidenced by a decline in asset quality, eroding profit margins, and increasing leverage), and adopting aggressive accounting practices (e.g., receivables growing much faster than sales, large income-inflating accruals, and decreasing depreciation expense). Such a company would receive a larger M-Score, and should be considered a likely earnings manipulator.

Can the M-Score Predict Stock Returns?

After developing and verifying the model, Dr. Beneish released additional research in 2013 finding that the M-Score not only predicts earnings manipulation, but stock returns as well.[2] For the time period studied (1993-2010), high M-Score companies that were flagged by the model as earnings manipulators performed an average of 9.9% worse than lower M-Score companies. In 1995, 1999, 2000, and 2001, the non-flagged companies beat the manipulators by an impressive 16.5%, 28.2%, 32.2%, and 16% respectively. It is important to note that these size-adjusted return spreads are based purely on whether or not a company was flagged as a manipulator by having an M-Score higher than -1.78; no further industry or equity analysis or judgment was applied. When combined with other forms of analysis, the M-Score's predictive power can be even more remarkable.

Red Team Analytics' Utilization of the M-Score

Our firm utilizes the M-Score as a key initial screening tool when analyzing any potential investment. Although a company that manipulates its earnings can certainly perform well in the short term, we do not view such companies as viable long-term investments. While we do utilize Dr. Beneish's recommendation to consider any firm an earnings manipulator with an M-Score larger than -1.78, Red Team Analytics also finds the relative M-Scores between companies to be useful in comparing investments. This remains true even if neither firm (or both) is flagged as a manipulator.

Additionally, in keeping with Dr. Beneish's 2013 research, Red Team Analytics pays careful attention to companies with low scores on the ACCRUALS variable but earning high M-Scores overall. Beneish found that low-accrual companies (companies with high earnings quality according to their accrual ranking) displayed a significant spread in returns between those with high and low M-Scores. Specifically, Beneish found high M-Score but low-accrual companies earned an average of 19.8% worse returns than low M-Score low-accrual companies.[3]

Finally, the M-Score is an important component of the O-Score, another tool developed by Dr. Beneish. Red Team Analytics utilizes a modified version of the O-Score (the "Red Team O-Score") to compare potential investments within a particular industry.

Conclusion

The M-Score is an incredibly valuable tool for any investor. Unfortunately, even though the original research papers upon which the M-Score is based are available to the public, there is still much confusion about how to use and interpret it. Perhaps of most concern are those individuals and entities which report M-Scores for companies without providing the values of the individual variables and analyzing the meaning behind them. Armed with a better understanding of the model, investors can utilize the M-Score to identify potential earnings manipulators as well as better predict stock returns.


[1] Beneish, Messod D. "The Detection of Earnings Manipulation." Financial Analysts Journal 55.5 (1999): 24-36.

[2] Beneish, Messod D.; Lee, Charles M.C.; and Nichols, D. Craig. "Earnings Manipulation and Expected Returns." Financial Analysts Journal 69.2 (2013): 57-82.

[3] Ibid.

Disclosure: I/we have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours.

I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.