If someone tells you they can predict the likelihood that your company will face a securities-related class action lawsuit over the next year, should you believe it? Los Angeles-based Audit Integrity says it can do just that.

The company claims it has developed a model that it says can predict securities-related class action lawsuits of public companies using a variety factors, the main one being its own “Accounting & Governance Risk” score.

The company, founded in 2001, spent nearly three years developing its proprietary AGR model. The AGR, which is based on more than 600 factors, measures forensic accounting risk—the risk that companies’ reported financial numbers misrepresent their actual results, making them unreliable.

Gottlieb

“AGR does not measure fundamental risk,” says Ophir Gottlieb, Audit Integrity director of client services. “Fundamental risk is all based on one core underlying assumption that all numbers are correct. AGR doesn’t measure that. We measure the risk that those numbers are misrepresentative.”

Now Audit Integrity has gone a step further, by developing a model that it says can predict class action litigation. The CAL model is based on a company’s AGR score, its AGR volatility over a two-year period, its market cap, industry, and stock decline over the last 12 quarters.

Audit Integrity rates about 7,000 U.S companies as well as 400 American Depository receipts. Ratings are updated quarterly, when new filings are made. In the interim, the company also tracks high-risk events, such as negative earnings surprises and divestures. It updates its clients on a weekly basis on the companies for which they have requested information.

“The inputs to the model are what the companies disclose to the public,” says Gottlieb. But the data isn’t limited to what’s included in a company’s financial statements; it includes annotations from the footnotes, information from MD&A, as well as an internal annotation process for restatements and litigation. “The quantitative modeling is the value add,” says Gottlieb. However, he also notes that the company does not apply qualitative analysis to its data. “The model is totally objective,” he says. “We don’t put our own analytical human judgment on it. The customers have to do that.”

Conservative Equals Safe

According to Gottlieb, Audit Integrity doesn’t track whether a company’s management may be unethical; rather, the firm simply analyzes whether a company’s accounting policies appear more “aggressive” than others. “We can’t say ‘This company is fraudulent,’” says Gottlieb. “But we can tell people if a company is behaving like previously fraudulent companies.”

In fact, he claims that companies that the firm rates as less aggressive have yet to be the subject of class action litigation. “In the last six months not a single company we’ve rated as conservative has been sued,” says Gottlieb.

NEXT IN COURT?

The ten highest one year litigation probabilities as modeled by Audit Integrity's Litigation Model.

Company

AGR Rating

One YearLitigationProbability

Xilinx

Average

13.38%

Celgene

Aggressive

13.27%

QUALCOMM

Aggressive

13.23%

Maxtor

Very Aggressive

13.14%

Lucent

Average

13.09%

Infineon Technologies

Average

12.61%

Sears Holdings

Very Aggressive

12.54%

Boston Scientific

Aggressive

12.46%

Pfizer

Aggressive

11.35%

Motorola

Very Aggressive

10.60%

Source: Audit Integrity.

That isn’t to say that the model never misses. “Every once in a while, the model can be wrong,” he says. “That’s the natural existence of quantitative models—you need human judgment on top of it.” The other caveat, according to Gottlieb, is that the model doesn’t predict product risk. “We don’t predict product risk. You can’t predict product lawsuits based on accounting.”

Gottlieb says the model can predict the risk of class action litigation over a three-month period, six months, nine months and one and two years. He adds, “We only predict SEC class actions where there’s either guilt or a settlement, not frivolous lawsuits.”

According to Gottlieb, a year is the most commonly used timeframe. “It’s generally the most useful, since D&O insurers generally write one-year policies,” he says. The company claims that D&O insurers are among its clients, as are ratings agencies, auditors, investors and some corporations; Forbes Magazine used Audit Integrity’s AGR ratings to help it choose its 2005 “Best Managed Companies in America.”

A Grain Of Salt

But should companies really care if they get flagged as being at high risk by AI’s litigation model?

Maybe, according to some securities class-action experts.

Tabak

“Looking at the steps they’ve gone through, it seems to be fairly reasonable, but people shouldn’t take the rating so literally,” says Dave Tabak, senior vice president at NERA Economic Consulting. “All models have a reasonable degree of error.”

NERA developed its own D&O frequency and severity model for U.S. public companies, which it says helps calculate the likelihood that a company will face a future securities class action suit and the possible severity of a suit.

Tabak points out that Audit Integrity’s rate of probability of litigation for the same company varied widely over the different time periods. Says Tabak, “When results bounce around like that, you have to take it with a grain of salt.”

However, he adds that companies should think of such ratings as a preliminary check up at the doctor’s office.

“It might be useful to raise some flags,” says Tabak. “If you a have a bad score, you should try to figure out why—see whether the model just isn’t a good fit for your company or whether there really is something there.”

Another reason Tabak says companies might want to look at their rating is simply the possibility that others—such as investors and D&O insurers—may be looking at them. “If investors are looking at this and using it in making their decision on whether to buy and sell a stock, it could impact a company’s stock price,” he says. “And if insurers look at this to help them determine what type of premium to charge for D&O insurance, companies might want to pay attention,” adds Tabak.

However, Tabak cautions, “The rating focuses very much on accounting practices. If you look at the most recent batch of litigation, most of it related to accounting practices. If the focus [of litigation] shifts to something else, it’s not clear whether this model will work.”

Roberts

According to Lyle Roberts, a securities and litigation expert at Wilson Sonsini Goodrich & Rosati, services like Audit Integrity’s may be limited in both audience and relevance. “The people who would and should care are most likely D&O insurers,” he says. “For companies, the ratings may be interesting, but they may not be as directly behavior modifying.”

In addition, he warns that there is a chicken-and-egg conundrum when considering predictive models. “It’s a bit of self-fulfilling prophecy,” says Roberts. “If they can convince enough people that they can accurately predict [litigation] and they have a lot of users, it would be a helpful tool.”

In addition, Roberts warns that there are myriad factors that impact whether a securities class action is ultimately brought against a company, and many of those “intangible” factors are not readily accessible for predicting behavior. "Computer modeling may assist you in figuring out the likelihood of a suit,” he says, “but there are a lot of difficult-to-predict factors.”