In Fraud, Many Hands Make Light Work

Implications: This article is based on “Control Overrides in Financial Statement Fraud” by Robert Tillman and Michael Indergaard of St. John’s University, available at http://www.theifp.org/research%20grants/tillman_final_report.pdf The study examines 834 companies that filed financial restatements between 1997 and 2002. Of those 374 (45 percent) were accused of securities fraud and subject to shareholder suits, SEC enforcement action or both. Of those resulting in legal action seven individuals on average were implicated holding a variety of senior positions, including board members. management and auditors  

Analysis: Methods

The study is worth reading because it shows the mass of information available, and point to many research opportunities that are ignored or mentioned and not explored.

The original paper is 94 pages long, and has other information of interest. With the information available it could have done more than address the state aim of: “…understanding how senior executives were able to commit financial statement fraud despite the presence of numerous control agents who, in theory, should have been able to detect and prevent those fraudulent activities.” (p38).

The first step was creating a Large-N drawn from 1997-2002 for companies from a

GAO sample of firms filing restatements consisted of 919 restatements announced by 845 companies. The GAO study “focused on financial restatements resulting from accounting irregularities but excluded those required by changed GAAP, organization or Class action suits filed against those firms with restatements. The sample was further broken down by and reduced by these data points:

-SEC action against those firms - 166

-Companies with class action suits - 348

-Companies with either/or - 374

There are problems with sample selection that should have been addressed and weren’t:

-Restatements are not a necessary indicator of fraud. Further there can be fraud when no restatement takes place that is the amounts were not material. As a note, materiality, at the time was generally held to be 5% of income.

-The sample was further limited to those firms with restatements that suffered class action suits, and SEC action.

The problem is limited the sample based on other people’s, in this case GAO’s, on initial sample selection. The description could have been clearer based on the following:

-Are restatements limited to amended filings or do they include restatements of prior year statements. The distinction is important because it ties to severity of the problems.

-Are the restated items cash or non-cash. This distinction is more important now than before. Non-cash items are less interesting than cash ones for the market and analysts.

-If restatements trigger class-action suits and SEC investigations, what about those that happened when there was no restatements? It would have been useful to look at other instances of legal action that were not associated with filing changes.

In summary, limited a study of fraud to items deemed by a third party to be fraudulent, is very limiting. Why not look at the universe, and then look at other indica besides financial restatements. 42% of the companies had class action suits and 20% had SEC actions. A better method would have been to start with the original group, and check that statistically, or problems.

For the Small-N sample, the authors chose 17 firms from the 166 subject to SEC investigation, which varied greatly in size. For this small group, the authors reviewed SEC documents and class action suit with an emphasis on individual and group relationships.

A better plan might have been to ignore restatements entirely and selected a Small-N from all suits or all SEC actions. By going from restatements to SEC suits, they ignored instances where there was wrong doing which had not been monetized as of yet.

Results

Everyone ‘knows’ that big frauds have a lot of people involved. This study actually puts numbers to the people and positions in the Large-N part of the study. In the 374 class action suits, there were 865 defendants and for the 166 SEC actions, and 2, 281 for the 348 class actions: five and six per case respectively.

New Economy firms comprised 38% of all the firms in the Fraud Sample and only 27% of the firms in the Non-Fraud Sample. In this case New Economy includes not only computers, but energy and telecommunications. The authors do not note that the latter two had undergone radical deregulation and speculative pressures at that time. The main finding was that the largest losses related to New Economy companies.

The important findings relate to the number of people involved and their relationship. Not only were the companies involved, but other companies, banks (actually broker dealers) law and accounting firms, made the cut. In the sample accounting firms were named 71 times, and when they were named the losses were bigger, mirabile dictu!

As noted, it was corporate finance departments named were involved in IPOs and accounted for the largest group of outsiders named: again no particular surprise.

Among company personnel, well the issue came right back to “Tone at the Top.” CEOs were in the box 89.7% of the time with CFOs covering at 78.3%. Board members were a distant third at 40%. Thus, as we might expect, fraud is done by those with power and opportunity.

In the Small-N sample the authors explored the relational dynamics of fraud. Executives at both the middle and senior management levels had the ability to override controls even when acting alone. More noteworthy was the fact that they rarely were alone: in 12 of 17 instances, CEO colluded with others. Further, clusters of senior managers actually normalized corruption.

In the Small-N Sample, found three types of fraud:

-Isolated frauds had six cases, with only one CEO involved and no CFOs. These show how senior management can act alone by restricting information.

-Frauds by senior management cliques with three of the six total combining to manipulate financial results.

-Boundary crossing frauds were similar to cliques, except that they involved recruiting lower level managers.

As for auditors, whether there was collusion or not, in ten or the seventeen cases, no resistance was offered. Oddly enough, Sarbanes-Oxley puts great weight on senior management attestation of results, but it must be noted that auditors run a much greater reputation risk than management but seems to have no problem being involved with fraudulent results.

I think this paper is of interest because it shows the levels of collusion in frauds over a number of years. There are some weaknesses and areas where further exploration is necessary:

-There is a big difference between a fraud that takes company assets and one that misrepresents company worth. This paper notes the difference but doesn’t show how the two are different.

-The sample is skewed because it starts with restatements, then to a subset including suits and SEC actions. Most theft frauds are so small that the losses are not material. One large enough to be so, is quite uncommon or the company very small.

The study does provide some useful information, but I believe the sample selection was flawed and it did not exam enough individual companies nor discuss enough of the specific techniques used.