06/01/2012 03:06 pm ET Updated Aug 01, 2012

SEC Should Use Wall Street Technology to Detect Corporate Crime

In my article last week on funding for the Securities and Exchange Commission ("Wall Street Reform: Taking a Knife to a Gunfight"), I advocated increasing the budget for the agency to better enable it to police Wall Street. Now I want to present a specific idea about how the SEC could improve its performance and effectiveness dramatically.

Quantitative funds or "quants" use sophisticated computer programs to analyze and predict the behavior of financial markets. Broadly speaking, the quant system is made up of: 1) Input Engine to gather relevant data; 2) Forecasting Engine which processes the data using mathematical algorithms and applies probabilities to predict the future, and; 3) Recommendation Engine which applies heuristics (experience-based rules of thumb) to fine-tune the results of the Forecasting Engine.

In plain English, quants use high speed computers to first collect mountains of data -- including stock market data like prices, volume, volatility; company operating data like earnings, debt levels and other performance metrics, and even macroeconomic data -- and then to "crunch the numbers" to make predictions about the future. They basically do what stock market analysts do but on a much larger scale and with many more data points. Obviously, when the markets are extremely volatile, the results can be unreliable; but the basic principle behind these systems is solid.

My proposal is that the SEC should develop and use a similar computer-based system to identify corporate crime.

The biggest hurdle to catching corporate crime is a lack of manpower to process the reams of financial information that are generated by the staggering number of companies in the United States every day in order to detect any meaningful patterns. Or, to put it another way, there aren't enough people in China to analyze all the data that the private sector generates, and even less to identify smoking guns in the mix. Looking for a needle in a haystack would be a gross understatement.

By using mathematical models and the power of computers, that work could be done in a fraction of the time it would take humans to do it and would enable the SEC to red flag the most likely culprits for closer examination. It would be more comprehensive, possibly more accurate and definitely more efficient. What makes this possible is that financial crime leaves a trail, no matter how well hidden, and with enough data points and smart analysis, coupled with known patterns from previous cases, the SEC can uncover those trails quickly and automatically.

Sadly but fortunately, there has been enough fraud, insider trading and other white collar crime in the United States over the past few decades to enable the SEC to create templates for major types of crimes and to tag the telltale signs of those crimes. From the fraudulent accounting at Enron, malfeasance at WorldCom and the subprime mortgage scam to Bernie Madoff's Ponzi scheme, risk mis-management at JPMorgan Chase and now investment banking irregularities connected with the Facebook IPO, the corporate landscape is full of useful examples and criminal benchmarks... If analyzed carefully, these cases and a database of templates based on them can be used to make intelligent guesses for the future -- an electronic instinct if you will.

The key challenge here would be getting enough data to make analysis useful and that depends largely on Congress. Sarbanes-Oxley has made big public companies more transparent, making it easy to get information; for hedge funds and private companies the task is more difficult but not impossible, especially with reporting requirements now being considered for even highly secretive businesses. Moreover, the most destructive corporate crimes are not the ones committed by small companies involving a few thousand dollars but the more egregious ones committed by the largest players and running into billions -- those are the ones that damage our entire system and which must be curbed for the sake of the economy.

A computerized system would no doubt be expensive to create and take time to set up, but would dramatically reduce the manpower required for financial monitoring by automating a lot of the most basic tasks. The SEC's resources could then be reserved for more specialized functions such as individual investigations and enforcement. In the long run, the cost savings for the government would be tremendous, not to mention the benefit in terms of greater effectiveness in identifying fraud.

So, will all this eliminate corporate crime? No, but more efficient law enforcement will make transgressors think twice. A system like the one I have suggested would be the equivalent of a thousand cops on the streets, discouraging people from committing crimes and making the neighborhood safer.