Three years ago the stock market valued Starbucks (SBUX) at close to $25 billion. It was an ambitious price. The Seattle coffee chain, founded in 1985, had just reported fiscal 2005 results. Its market value worked out to 51 times profit and almost four times sales. Still, Wall Street analysts pointed out reasons for optimism. Sales at longstanding Starbucks stores had improved 8% in the fourth quarter, a sign that more store openings were warranted. Management had plans for plenty of them: 1,800 locations over the following year, an 18% expansion. A menu expansion was in the works, too. After several quarters of tests in Seattle and Washington, D.C., hot breakfast sandwiches were deemed ready for a broader rollout.

The past three years haven't been kind to stocks in general. The S&P 500 index, of which Starbucks is a member, has lost a quarter of its value. But Starbucks has lost more than two-thirds. It's not just that the hot sandwiches have flopped. (Management said in January it would discontinue them.) Fiscal 2008 results, reported late Monday, mostly describe what happens when premium pricing and overzealous expansion meet a sharp souring of consumer mood. Sales at longstanding stores shrank 8% in the company's fourth quarter. Management completed the first third of a planned 600-store closing. Restructuring charges, along with a sales decline against high fixed costs, shrunk quarterly profit by 97% from a year earlier.

For the full year, earnings per share fell to 71 cents (ignoring 28 cents of "restructuring" and "transformation" charges) from 87 cents a year earlier. Management didn't project 2009 results exactly, but offered a matrix of possible outcomes. All foresee further declines in sales at longstanding stores. Remarkably, none envisions a shrinking of earnings per share (so long as another 12 cents per share in restructuring charges is ignored). With sales drops ranging from 2% to 7%, earnings per share should be up 27% to flat, the company says. Analysts say such a feat is possible with a further closing of underperforming stores, efficiency gains at remaining stores and layoffs of non-store workers.

An hour into Tuesday's trading Starbucks was valued at about $7 billion. That's about 14 times fiscal 2008 earnings and 0.7 times sales. Investors who three years ago were asking how much higher Starbucks shares could climb are now wondering whether they've bottomed out. I called the stock too expensive at $24 in October 2007, saying it would be a better deal at $17. When it hit that price in March I thought it fairly priced, but called on the company to spend part of its sizable cash flow on a dividend. Stock prices in general have deteriorated since March, but Starbucks has done worse. On Tuesday it fell below $10.

Improvements are still afoot, as they have been since founder Howard Schultz returned to the role of chief executive in January. Single-cup brewing machines are meant to increase freshness and reduce waste. Chain-wide server retraining ought to reduce instances of negligent frothing. Gooey pastries, whose calorie counts have stunned my fellow New York City residents now that a local rule requires their listing, will soon be complemented by healthier options, like oatmeal and a fruit and nut bar. Loyalty cards give frequent customers a modest price break.

That said, I can't bring myself to re-recommend the stock right now. There are macroeconomic reasons. Consumer confidence in both the U.S. and U.K. recently fell to the lowest scores on record. The $5 daily latte, long the financial planner's favorite illustration of savings waiting to be found, seems due for a backlash. There are microeconomic reasons. Coffee competitors McDonald's (MCD) and privately owned Dunkin' Donuts look likely to attract Starbucks deserters in a slowdown, thanks to lower prices. There are subjective reasons. Regardless of price, I prefer both McDonald's coffee and Dunkin' Donuts coffee to Starbucks, which I often find over-roasted and smoky. That's one customer's opinion, but I'm a coffee enthusiast who roasts his own beans at home. And my opinion is shared by Consumer Reports, which last year gave McDonald's coffee top marks and found Starbucks "burnt and bitter." Dunkin' Donuts has recently run television ads saying it beat Starbucks in a blind taste test. Of course, the counter-case to all this is that so many customers can't be wrong about Starbucks. The counter to the counter-case is White Zinfandel wine.

More than anything else, though, there's a fiduciary reason for staying clear of the stock. Starbucks doesn't provide a return. That is, the dividend I mentioned in March still hasn't materialized. The company says it will spend most of its free cash over the next year to pay down short-term borrowings, a worthy enough goal. But in a down market buyers are far more attracted to plump dividend yields than to modest price/earnings ratios. They know that in a long slump quarterly cash payments can make all the difference. For example, between the first time the Dow Jones Industrial Average hit 8,000 on July 16, 1997, and the most recent time, Oct. 10, dividends made the difference between a dog like General Electric (GE) losing 13% and making 13%, and between a mediocre performer like Consolidated Edison (ED) returning 22% and returning 123%. Until Starbucks makes good with payments, I'd pass.

Starbucks isn't the only mature cash generator I've badgered of late for not coughing up dividends. I reached a similar conclusion on videogame seller GameStop (GME) in mid-October. It has since fallen another 12%. If you're interested, here's a list of some other offenders, produced using SmartMoney's stock screener.

Recipe

• Trailing 12-month sales greater than $1 billion
• Price/free-cash-flow ratio below 25
• Last quarter sales growth less than 10%
• No dividend