All around the country, a new airport craze is in full swing. Denver International - the first major airport built since Richard M. Nixon was president - is finally running more or less smoothly. Big airports coast-to-coast are expanding rapidly. And new ones are being constructed, including speculative gambles such as MidAmerica Airport in St. Clair County, Ill., 25 miles from St. Louis, and Northwest Arkansas Regional Airport, near Fayetteville.

So many military bases are closing that airfield dreams are popping up everywhere. In Southern California, which already has four major airports, plans are being made to create at least four more out of closed bases.Could we be overdoing it? If you build it, is it possible that they won't come?

Well, yes. The whole role of airports in economic development appears to be more complicated and harder to nail down than anyone thought. On the one hand, communities want airports in order to stimulate economic activity. On the other hand, merely having an airport may not be enough. "In a deregulated airline industry," says Roger Cohen, a spokesman for the Air Transport Association, "airlines fly to markets, not airports."

Far from being a "field of dreams" issue, airport construction may well be a chicken-and-egg question involving the airport and the surrounding market. If you want to promote growth, which do you subsidize: the chicken or the egg?

Much of the airport jostling these days derives from typical American decentralization. Commercial air travel is required to operate as a seamless system under the guidance of the Federal Aviation Administration. But when and where to schedule the flights is up to the airlines. And where the airports are is up to the local communities, at least partly. Right now, cities, metropolitan areas and states all over the country have decided it is in their interest to subsidize airport construction and expansion.

In some cases - as with the pending expansion of Los Angeles International Airport - the market for air travel is so extensive that there probably will never be too much capacity. But in other cases, as with MidAmerica, airports turn out to be little more than another way for communities to fight it out.

Lambert St. Louis International Airport, for example, is small and outdated, but it is a well-established destination for Southwest, Delta, United, American and its flagship airline, TWA. Lambert is undergoing a $2.6 million expansion that is scheduled for completion in seven years.

Meanwhile, just across the state line, St. Clair County in Illinois has constructed MidAmerica at a cost of $300 million. It is fully operational, but has no commercial business and no scheduled flights. County officials say they aren't worried because they've budgeted for a five-year, ramp-up period. But there's little question that the Lambert-MidAmerica battle is a less-than-ideal (if typical) solution to a regional problem. After fighting for decades over which state would get the new St. Louis airport, Missouri and Illinois decided they would each build one - even though there was little evidence of sufficient demand to support two.

Local leaders often say that building airports is a local decision, and so there's nothing wrong with communities wasting their own money if they feel like it. But many airports are constructed with substantial amounts of federal money, and one could argue that federal priorities should prevail.

In the case of MidAmerica, some $150 million, or fully half the construction budget, came from the FAA as a result of the fact that Illinois won the lobbying battle in Washington. Because MidAmerica is a "joint use" operation with Scott Air Force Base, the Pentagon chipped in another $60 million. But this did not stop Missouri from using Lambert's established position to raise revenue from passengers and from Missouri taxpayers to expand Lambert - thus going into competition with the federal government's decision to funnel money to Illinois rather than Missouri.

Maybe the lesson in all this isn't that if you build something, people will come or even that you need to decide between the chicken and the egg. Maybe it's just that in a decentralized system such as the one we have in the United States, you can't stop communities from taking big gambles even if they're destined to lose. It's not likely that both Lambert and MidAmerica will come out winners in the end - but it would have been impossible to talk either side out of taking the gamble in the first place.