Earnings Per Share? Phooey!

Ivan Feinseth eschews traditional gauges -- and his picks have trounced the S&P

Ivan P. Feinseth, director of research at Manhattan broker-dealer Matrix USA, has slapped so many stocks with a "Strong Sell" rating that you'd think he has an ax to grind. Even after the stock analyst scandals of the 1990s, most Wall Streeters are loath to call a bad stock when they see one: In total, they recommend that clients get out of just 7% of the stocks they cover. But Feinseth rates 40% of his universe of 3,100 stocks -- roughly 1,200 companies -- "Sell" or "Strong Sell."

Feinseth insists it's nothing personal. Rather, his computer model, which forces him to rank the best and worst relative values in the Russell 3000 index, makes him do it. Given that a third of that benchmark's stocks were down this year through May 26, Feinseth's 40% "Sell" percentage seems a trifle more realistic than the rest of the Street's.

Skeptics point out that by focusing primarily on data, he misses out on the nuances and "body language" that other analysts pick up on as they study companies and their executives. Feinseth doesn't seem to care. "We don't talk with the companies," he says, swiveling around to four computer monitors brimming with charts in Matrix' modest, disheveled midtown office. "What we're doing is a deep analysis of the numbers."

What makes Feinseth's analysis unique is its fixation on something other than the bottom line. He says traditional corporate earnings are vastly overrated as an investment gauge. The question, he says, is whether a company is creating real, hard value for its shareholders. "Earnings per share," he says, "is just an accounting measure of profit, when the real key to the stock price is economic profit." In other words, earnings are in the eye of the beholder. Companies manage analysts' expectations quite carefully -- so they can beat them. But they can't fudge economic profit, which Feinseth defines as the extent of a company's return above its cost of capital.

That barometer, which anchors a proprietary model that includes 12 other factors, has helped Feinseth trounce the market since Matrix' late 2004 launch. Through mid-May, Matrix' 25-to-60-stock "Focus List" of favorites (yes, he recommends stocks as well -- 1,200 in all) has gained 12% in 2006, compared with 4% for the Standard & Poor's 500-stock index. Last year the list registered a gain of 33%, vs. 5% for the S&P. Overall, since its September, 2004, inception, the list has done more than triple the market's return.

It would consume every hour of the trading day for someone to act on the thousands of Matrix' frequently changing calls. Still, Feinseth believes this is a trader's market where he can deliver excess return only by managing his ratings in real time. "Even the greatest stocks," he says, "are not buy and hold. You need to understand the right time to be in and out."


Lately, Feinseth has been focusing on Microsoft Corp. (MSFT ), whose shares are at a multiyear low. Wall Street doesn't see much hope, at least in the near term. But Feinseth's analysis suggests a screaming "Buy." The reason: The company's enviable 39% return on capital dwarfs both its borrowing costs and the 4% or so that its $35 billion cash hoard is earning in the bank. It makes perfect sense for Microsoft to take advantage of that yawning spread by aggressively acquiring its own shares. "They should theoretically take out debt to buy back stock," he says. "They'd only be increasing their own buying power." Feinseth, naturally, is confident his reasoning will prove correct.

In the meantime, he says 30-employee Matrix is seeing brisk business. It earns its keep through an institutional client list of mutual funds, state pensions, and hedge funds. The firm also has a private client group that manages accounts that average $1 million, collecting an advisory fee on assets.

Matrix' marketing material is grandly rendered. "Accounting-based models of valuation should be buried right alongside Communism," reads one packet. "If the Berlin Wall can come down, so too can the assertions that stock prices are based on reported accounting earnings." Nothing nuanced about that.

By Roben Farzad

    Before it's here, it's on the Bloomberg Terminal.