The recent surge in oil prices to roughly $50 a barrel teaches some useful lessons. One is that surprises happen. A year ago futures contracts predicted today's price would be $25. A second is that the economy has grown less vulnerable to oil "shocks." Compared with 1973, we now use almost 50 percent less energy for each dollar of output. New industries (software, theme parks) need less than the old (steel, chemicals). But the largest lesson is depressingly familiar. Americans won't think realistically about oil. We consider cheap fuel a birthright, and when we don't get it, we whine -- rather than ask why or what we should do.

If prices rise, we blame a conspiracy of greedy oil companies, OPEC or someone. The reality is usually messier. Energy economist Philip Verleger Jr. attributes the present price run-up to massive miscalculation. Oil companies and OPEC underestimated global demand, particularly from China. Since 2001 China's oil use has jumped 36 percent. This error led OPEC and companies to underinvest in new production capacity, he says. In 2002 the world had 5 million barrels a day of surplus production capacity; now it has little. Unexpected supply interruptions (sabotage in Iraq, civil war in Nigeria) boost prices.

Verleger says prices could go to $60 next year or even $80 if adverse supply conditions persist. No one really knows. Analyst Adam Sieminski of Deutsche Bank thinks prices may retreat to the low $30s in 2005. A slowing Chinese economy could weaken demand. But the uncertainties cannot obscure two stubborn realities. First, world oil production can't rise forever; dwindling reserves will someday cause declines. And, second, barring miraculous discoveries, more will come from unstable regions -- especially the Middle East.

We need to face these realities; neither George Bush nor John Kerry does. Their energy plans are rival fantasies. Kerry pledges to make us "independent" of Middle East oil, mainly through conservation and an emphasis on "renewable" fuels (biomass, solar, wind). Richard Nixon was the first president to promise energy "independence." It couldn't happen then -- and can't now. The United States imports about 60 percent of its oil. A fifth of imports come from the Persian Gulf. Even if we eliminated Persian Gulf imports, we'd still be vulnerable. Oil scarcities and prices are transmitted worldwide. The global economy -- on which we depend -- remains hugely in need of Persian Gulf oil.

Bush's pitch is that we can produce our way out of trouble. No such luck. Drilling in the Arctic National Wildlife Refuge, with possible reserves of 10 billion barrels, might provide 1 million barrels a day, or 5 percent of present U.S. demand. Fine. But the practical effect would be to offset some drop in production elsewhere. American oil output peaked in 1970; it's down 34 percent since then.

A groundbreaking study from the consulting company PFC Energy illuminates our predicament. The world now uses 82 million barrels of oil a day; that's 30 billion barrels a year. To estimate future production, the study examined historical production and discovery patterns in all the world's oil fields. The conclusion: The world already uses about 12 billion more barrels a year than it finds. "In almost every mature [oil] basin, the world has been producing more than it's finding for close to 20 years," says PFC's Mike Rodgers. That can't continue indefinitely.

The study is no doomsday exercise. Rodgers says that future discovery and recovery rates could be better -- or worse -- than assumed. With present rates, he expects global oil supply to peak before 2020 at about 100 million barrels a day. Whatever happens, the world will probably depend more on two shaky regions: the Persian Gulf and the former Soviet Union. The Gulf now supplies a quarter of the world's oil; PFC projects that to rise to a third in a decade.

Although the future is hazy, what we ought to do isn't. We need to dampen oil use, expand production and -- if oil prices recede -- significantly increase the Strategic Petroleum Reserve. These steps can't end our vulnerability to global price surges or the effects of a catastrophic loss of oil supplies from, say, war or terrorism. But they can reduce it. Most important, Americans should curb gasoline use. The Energy Information Administration reports the following: Gasoline represents about 45 percent of U.S. oil demand; since 1991 the explosion of SUVs and light trucks has meant no gains in average fuel mileage efficiency; and over the same period, typical drivers travel almost 1,000 miles more annually.

We should be promoting fuel-efficient vehicles, particularly "hybrids." Combining gasoline and electric power, they get 20 percent to 40 percent better mileage than conventional vehicles, says David Greene of the Oak Ridge National Laboratory. They also cost from $3,000 to $4,000 more than conventional cars, he says, mainly because they have two power sources. But Greene plausibly asserts that if production expanded, the cost gap would shrink. The way to expand demand would be to adopt a gasoline tax of $1 to $2 a gallon. Americans would know that fuel prices would stay high. They'd have reason to economize.

The tax should be introduced over five to 10 years to give drivers and auto firms time to convert.

Of course, a fuel tax is a political showstopper. It isn't in Bush's or Kerry's plan. They promote hydrogen-powered cars. These sound great but -- given the technical obstacles -- won't become widespread for many years, if ever. This captures our choice: taking modestly unpleasant preventive steps; or running greater future risks by clinging to our fantasies. History favors our fantasies.