When George W. Bush ran for president two years ago, it was with the help of some high-profile supporters at the statehouse level: Eight of the 10 most-populous states were governed by Republicans. In 2004, it won't be quite the same. Democrats took a serious bath in the elections last month, but they managed to climb out of the tub clutching the statehouse in half of the Top 10 and, even more remarkably, in 13 of the Top 20. As of next January, more Americans will be living under Democratic governors than under Republicans--at least 20 million more, as a matter of fact.

We all know why this is important: The more governorships a party controls, especially in the large states, the more of an advantage it has in seeking the White House. All the experts agree on it.

And all of them seemed to be repeating it before and after Election Day. "To have governors in these statehouses, in these large electoral states, is crucial," declared Democratic National Committee chairman Terry McAuliffe. "The Midwest will be critical," conceded Republican consultant Scott Reed. "It will be tougher to win with Democratic governors setting the agenda in those states."

Just how much difference can a governor make for his party's presidential nominee? Most commentators seem pretty much in agreement on this as well: 2 to 3 percentage points. This sounds plausible, until you stop and reflect that this is their answer to almost any question about the importance of any factor in elections. What can you expect to gain from a get-out-the-vote effort? 2 to 3 points. How much bounce can you expect from a strong debate performance? 2 to 3 points. What's the likely impact of a presidential visit on an off-year Senate race? 2 to 3 points.

If I were a little more cynical, I might suspect that the experts had hit upon a foolproof way to sound perceptive without actually knowing much. A couple of percentage points is a small enough number to seem plausible to almost anyone, and just precise enough to be convincing. That's not a criticism, of course. I may want to try this ploy myself sometime soon.

The beauty of such prognostications is that they are very hard to disprove. But they do submit to at least a rough kind of historical analysis. Just for fun, I tried it. When the last presidential election was held, in 2000, Republicans held the governorship in eight of the nation's 10 largest states. Bush won three of them--if you include Texas, where he himself was the governor, and Florida, where nobody even knows what happened. If you don't include Texas and Florida, Bush carried exactly one out of the GOP's Big Eight.

An aberration, you might say. But then I looked at 1996. The list of megastates was slightly different at that time, but Republicans still controlled the governorship in eight of them. Bob Dole carried one. If this is the kind of juice governors possess in presidential politics, it's been running a little thin lately.

Admittedly, the methodology isn't foolproof. Perhaps one should focus on individual states rather than the entire map at once. Bush carried Ohio in 2000 by a little more than 150,000 votes. A portion of this might possibly be attributable to the "governor's bonus" afforded him by the presence of Republican Bob Taft in the statehouse. Then again, it might not. At some point, the whole subject becomes so fuzzy as to cease being very interesting. About all you can say with confidence is that it helps to have the governor on your side if he is your brother, the result is extremely close and he is in charge of counting the votes. Otherwise, you shouldn't expect too much.

What's interesting is how journalists, consultants and other experts seem determined to find some connection between governors and presidential politics, however tenuous it might be. Second only to the "governors bonus" theory about presidential elections is the "new faces" theory. You hear this one all the time. After all, the theory goes, four of the last five men to win the White House emerged from gubernatorial ranks. Take a good look at the incoming governors of this year, and you will see the presidential and vice presidential nominees of tomorrow.

Well, not tomorrow, exactly. Some time in the distant future would be more like it. Bill Clinton made it to the White House in 1992, but that was 14 years after his first election as governor of Arkansas. Fourteen years also elapsed between Ronald Reagan's initial gubernatorial victory in California and his victory for president. Even Michael Dukakis had to wait 14 years before being nominated to head the Democratic ticket--albeit unsuccessfully--in 1988. (Seems to be something in the 14-year apprenticeship, doesn't there? Maybe 2004 will be the year for Pete Wilson.)

George W. Bush made it to the White House after six years running Texas, but he was able to do that because he was a Bush, not because of anything he did as governor. In the entire 50-plus years of postwar history, only Jimmy Carter's election in 1976 provides any real evidence for the "new faces" doctrine. Carter can legitimately be said to have built on his record and contacts in one term as governor of Georgia to vault himself quickly onto the national political scene. But that's a slender thread to hang a cliche on.

It's even more slender when you consider governors as potential vice presidents. In the past 80 years, only one sitting governor has been elected vice president. That was Spiro Agnew in 1968. Before Agnew, you have to go all the way back to Calvin Coolidge. Stories about the new crop of governors producing fresh material for the next national ticket are no doubt flattering to the governors themselves. But they hardly have any historical basis at all.

Well then, if governors don't determine presidential results in their states, and they don't make quick, dramatic leaps into presidential politics, what makes gubernatorial elections worth devoting so much space to? Here's an answer that's so simple it tends to be overlooked: Governors are important not because they are electoral power brokers, or because they are future presidents, but because they make policy. They make national policy. It's arguable that they have been making more of it lately than Congress has.

Over the past decade, as Congress has mired itself ever more deeply in partisan gridlock, creative governors have moved in to fill the policy vacuum, sometimes out of ambition, sometimes out of sheer desperation.

After Congress failed to enact a health care program in the early 1990s, Vermont under Howard Dean began to enact one on a piecemeal basis that eventually grew to take in much of the middle class as well as the uninsured poor. Maryland's Parris Glendening all but invented the concept of Smart Growth, sold it to his legislature and then pushed it onto the national agenda effectively enough to generate new laws in states all over the country. The one legislative action of the 1990s most often cited as a genuine congressional accomplishment-- passage of the welfare reform act of 1996--was, in fact, a derivative of experiments launched in Wisconsin under Tommy Thompson, and then tried in other places and judged to work.

It would be easy to come up with a much longer list. The 1990s produced a large crop of statehouse entrepreneurs in both parties who not only wanted to influence the national agenda but knew how to go about it. Virtually all of them are gone now, or will be next month, either through voluntary retirement, term limits or, in the case of Georgia's Roy Barnes, electoral defeat.

So if the era of gubernatorial activism is to continue, it will have to be in the hands of a new generation of activists. As it happens, a huge new generation is about to be sworn in--24 of them, which would be the largest number of new governors to come in to office together in decades. They are the gubernatorial equivalent of the Baby Boom generation.

And they are being born into a situation far different from the one their predecessors were familiar with. Dean, Glendening, Thompson and their contemporaries had budget surpluses and significant amounts of cash to play with; the incoming Class of 2003 will be confronting the biggest state fiscal crisis in decades. Nearly every state is burdened by a tax system that doesn't generate enough revenue to meet its costs, and a combination of fixed spending commitments, foremost among them Medicaid, that are impossible to repudiate. And virtually all of these new governors campaigned for the office on promises not to raise taxes, or to cut spending in the few big categories where they have discretion, such as the schools.

But who's to say that austerity won't be as good a crucible for creative policy as prosperity. Faced with the need to break with precedent and find new ways to live within their means, at least a few members of the statehouse Baby Boom will turn out to be genuine innovators. And the experiments that succeed will force their way into national debate. That's the way things work now: Ideas flow up from the state capitals, not down from Washington.

And that's what makes governors interesting and makes their elections worth paying attention to. Treating them as adjuncts to presidential campaigns is not only useless--it misses the crucial point about how the American political system currently works.