WORK EXPANDS TO FILL

the time available. That's Parkinson's Law, familiar to anyone with a long memory-or a cynical mind. Its author, C. Northcote Parkinson, was a British historian intent on poking fun at the perpetually overburdened civil service. In the 1950s and '60s, Parkinson wrote several tongue-in-cheek books about economic theory, and this month's column is an update on another of his insights.

My subject is capital spending, the kind of nuts-and-bolts expansion that ought to be good news. In the short run, it is. When a company announces a new plant or a major investment project, its shares typically move up. Longer term, however, growth in capital spending is a leading indicator of poor stock performance.

To understand why, consider Parkinson's Second Law: Expenditures rise to meet income. It's easy to see how this applies to individuals. When you get a raise, you're apt to buy a larger house, a designer dress or a faster car whether you need those things or not. Politicians also tend to find new uses for money as fast as it comes in. Could those hard-nosed folks who run the nation's businesses behave the same way? Yes. The more money a company makes, the more likely its managers will build new plants and buy more equipment, even if the potential return doesn't justify the expense. Economists noticed this relationship about 20 years ago, and it has puzzled them ever since. Corporate executives, with all their training in project finance and present-value analysis, shouldn't behave as if money burns a hole in their pockets.

Several recent studies highlight what I call the capital spending anomaly. If you buy shares in companies that are cutting back on spending and bet against the shares of companies that are on an expansion binge by selling them short, you can turn a tidy profit.

Before I get to details, it's important to be clear about definitions. Capital spending is an accounting term, a measure of additions to fixed assets. These can range from a few Aeron chairs to a $2 billion semiconductor fabrication plant. Both the chairs and the factory wear out gradually, so the cost is not an immediate expense. It goes on the balance sheet, in an account called "property, plant and equipment," which is gradually reduced by charges against income otherwise known as depreciation.

Two points are worth remembering. First, not all capital spending adds to capacity or improves productivity. A $6,000 shower curtain gets the same accounting treatment as a room full of automated machine tools. Second, capital spending is different from investment. That's an economists' term, and it includes things such as research and development, which accountants require companies to expense immediately. Build a factory and the cost goes on the balance sheet. Equip a lab and the cost comes off the income statement.

One of the first papers to explore the link between capital spending and stock returns appeared last year in the "Journal of Financial and Quantitative Analysis." The authors, Sheridan Titman of the University of Texas and two colleagues, looked at stocks during the period from 1973 to 1996. No complex math: They simply measured how each company's most recent capital spending (taken from the annual cash flow statement) related to its average capital spending for the previous three years. Then they created model portfolios based on capital spending growth.

The results are ammunition for anyone who thinks managers waste shareholders' money. The top quintile (the 20% of companies with the greatest capital spending growth) trailed the market by 1.5%age points annually. Meanwhile, the bottom quintile beat the market by an average of a percentage point. That may seem like a small difference. But it's statistically powerful and the numbers get bigger in later studies. What's more, the performance gap may persist for as long as five years.

In a paper that will appear in the "Journal of Finance" later this year, Christopher Anderson of the University of Kansas ranks companies based on two-year capital spending growth. His results show a gap of about six percentage points a year between the big spenders and the penny pinchers. John Wei, who now teaches at Hong Kong University of Science and Technology and was a coauthor of Titman's paper, has his own follow-up study. Using a more complex measure of capital spending, he generates gains of better than eight percentage points annually by buying companies with low capital-spending growth and shorting those with high growth.

If this seems hard to swallow what economists call counterintuitive consider another strand of research. Marianne Bertrand at the University of Chicago and M.I.T.'s Sendhil Mullainathan tracked nearly 25,000 bids on leases to drill in the Gulf of Mexico between 1963 and 1999. The bidders, mostly large energy outfits, tended to pay higher prices when they had more to spend. And no surprise to Parkinson those more costly tracts were not more productive. The most popular theory about why companies make such wasteful investments is "empire building," management's desire to run a larger organization. In the offshore bidding situation, however, Bertrand and Mullainathan, who's now at Harvard, propose an alternative explanation called "quiet life." Rather than working to find new things to do with their cash, executives simply pay more to take the path of least resistance.

The accompanying table is my attempt to piggyback on these ideas. As usual, I looked only at companies with a market value above $500 million. Like the academics, I eliminated financial firms, which tend to have negligible capital spending. Then I ranked spending as a percentage of assets for the most recent fiscal year compared with the average for the previous three years. I focused on the bottom 10% of the list.

That gave me 200 companies where current capital spending is at least 50% below the three-year average. I cut this list to 60 by focusing on stocks that pay dividends on the theory that giving money to shareholders is likely to improve efficiency. I got my eight finalists by applying two value yardsticks: I wanted companies that are cheaper in book-value terms than industry medians, but growing faster.

Check the capital expenditure numbers in the fifth column above. While spending was down sharply at my companies, it also declined for the Standard & Poor's 500, though not nearly as much. Note, too, that while capital budgets may not be growing at my companies, several are still big spenders. At SkyWest, for example, spending was 13% of total assets last year, three times the market average.

Otherwise, the firms have little in common. American Italian Pasta, the low-cost noodle producer, with 25% of the market, is waiting for the Atkins diet to fizzle. Black Hills is a South Dakota utility that's unusually diversified. Ditto Journal Communications, a Wisconsin media company with newspaper, broadcasting and telecom businesses. Donnelley and SPX have new management teams eager to shake things up.

PerkinElmer is a veritable supermarket for anyone who buys scientific instruments. Royal Caribbean boasts an enviable niche in the cruise business, buoyed by favorable demographics and high barriers to entry. And SkyWest with more than $500 million in cash is a nonunion provider of regional air services for larger carriers.

Capital spending is in a slump at each of these companies. If my take on the academic research is correct, that may lead to superior stock performance. Why? Maybe the folks in charge aren't out to build empires or live the quiet life. They might simply be trying to make a buck for shareholders.