Analyzing Gradient Analytics

Surely you’ve heard of Gradient Analytics by now, as it’s all the Wall Street media outlets have been talking about lately. In case not though, let’s recap. Gradient is a research firm based in Scottsdale, Arizona. They focus on forensic accounting analysis of public companies.

The latest issue that has been getting all the attention has to do with negative reports Gradient issued on Biovail (BVF), a poorly run Canadian drug company. Biovail is suing Gradient claiming their inaccurate and misleading reports caused a 50 percent drop in BVF stock during 2003 and 2004. News broke that Steve Cohen, hedge fund manager at SAC Capital and Gradient customer, actually requested a report from Gradient that focused on negative aspects of Biovail’s business. SAC was interested in shorting Biovail shares, or perhaps they already were short at the time of request. Gradient did produce a report, and subsequently distributed it to the rest of its customers.

On the surface, this kind of thing looks very suspicious. If SAC Capital was short BVF and asked Gradient to write a negative report on it so they could profit from their short position, it certainly appears that so-called “independent” research is far from independent. Even still, a Gradient report is not responsible for a 50 percent drop in BVF stock, as the company contends. The company’s financial results account for the drop.

In response to the lawsuit, Gradient has said that SAC was requesting a follow-up to a previous report it published about Biovail on its own. If that is true, and SAC was not ghost-writing these reports with false information (as some are contending), then it is hard to reach the conclusion that any law was broken. Unless Gradient and/or SAC knowingly disseminated false information in order to profit from existing short positions, this issue seems to be blowing way out of proportion.

Most are screaming for disclosure of these types of relationships. As a result, Gradient should add some fine print at the end of its reports saying that they might have been published based on a client’s request, not because it was the research firm’s original idea. I have no problem with such disclosures, but don’t think for a second that it will change anything.

Sell side research now discloses how many buy, sell, and hold ratings they have on all of their investment banking clients, but we still see mostly buy ratings on stocks of banking clients. I recently read a report from a boutique firm specializing in healthcare stocks. They had a buy rating on the small cap biotech company I was reading about. In fact, at the end of the report they disclosed that their research analysts cover 11 companies with whom their firm has a banking relationship. All 11 stocks are rated “buy”.

How much stock should investors put on Gradient’s research anyway? Last week shares of Rackable Systems (RACK) dropped 6 bucks temporarily after Gradient’s computer model spit out a negative red flag about increasing inventories. They postulated this was a sign of poor earnings quality.

Unfortunately for Gradient’s clients, the computer program wasn’t able to research Rackable’s business model. Had it done that it would have learned that Rackable, much like Dell (DELL), builds product only after it is ordered. So, all of its inventory has already been sold, thereby making increasing inventory levels a signal of stength in the business, not the opposite as Gradient concluded.