Every time gasoline prices spurt upwards, the fallout is predictable. Consumers allege evil conspiracies of price gouging at the pump. The industrialized countries browbeat OPEC to raise production. Government leaders voice their "concern" on the impact on consumers, while inwardly salivating in a Pavlovian fashion at the prospect of a tax windfall.

The big oil companies gush profits even more fabulous than the norm. And, inevitably, the ideas of Marion King Hubbert make an excursion from the land of tree huggers and greenies to the mainstream.

A geophysicist, Hubbert predicted in 1956 that U.S. oil production would reach a peak between 1965 and 1972. This proved to be a remarkable bit of foresight; the zenith was actually reached in 1970.

Since then, many other would-be petroleum Nostradamuses have tried their hand at predicting a high-water mark for world production. So far, all have proved to be premature as declines in some parts of the world have been made up for by an extra spurt elsewhere.

Hubbert's theory is based on the principle that almost all of the oil harboured within the planet's crust has been identified. It therefore does no good to drill deeper or broaden the search. What's there is there.

This would perhaps explain why, despite greater efforts at exploration, the pace of new discoveries has slowed down sharply over the past 30 years.

His detractors, on the other hand, believe there are still places on the globe where oil might yet be found.

When an oilfield is exploited, it takes a while to find the best places to drill and for equipment to be built up and deployed, so production climbs slowly at first, then grows more rapidly until it hits a maximum, after which it wanes as the most easily available oil has been removed.

This bell-shaped pattern of oil extraction is widely known as the Hubbert curve, and when all the oil pools of the world are taken in aggregate, it is thought by proponents to describe the inevitable peaking of world production. According to them, this peak should occur between 2003 and 2009.

The Hubbertites like to emphasize that theirs is a more scientific approach to estimating, based on geology rather than economics. In fact, they deride the opposing notion that levels of output depend on the size of exploration budgets, new technologies for extraction, or the manipulation of supply and demand.

This is important stuff, because if they are right, the path for investors is clear: buy shares in a few large international producers with lots of reserves and salt them away for the long term.

We know that the world is not going to get over its addiction to petroleum anytime soon, so if there is a permanent and irreversible decline in production year after year, it's a slam dunk. What is left in the ground will become more and more valuable unless substitutes increase in quantity and are competitively priced.

It isn't easy to assess the plausibility of an imminent Hubbert's Peak. Though data on yearly levels of production is probably fairly accurate, reserves are a different story altogether.

A series of severe reductions in proven reserves has rocked Royal Dutch/Shell, costing the CEO and CFO their jobs and resulting in a billion-dollar lawsuit and an SEC investigation.

Here in Canada, new rules spearheaded by the Alberta Securities Commission that include more stringent criteria for the evaluation of proven reserves, have hit many juniors in the oil patch.

Baytex Energy Trust saw its reserves drop by 39 million barrels, equivalent to around $750 million. That's a lot of black gold.

On the other hand, revisions upwards are also not uncommon. ARC Energy Trust saw its reserves bumped up by about eight percent under the new rules.

While Shell was doing its mea culpas, BP disclosed that calculations based on SEC guidelines indicated that its North Sea reserves were 151 million barrels greater last year than published in the company's 2003 annual report.

The change had nothing to do with geological omissions but was simply a matter of the assumed benchmark for the price of oil.

The SEC stipulates that estimates must consider how much oil can be extracted on an "economically viable" basis, and that amount is far different at $15 than at $30 or $45 a barrel.

This method contrasts with that used by most oil companies and the Department of Energy, who do three sets of calculations, one with a high assumption, one with a low, one with best guess.

But the SEC wants only to know based on the current price, in a "mark to market" type of calculation, more like Jerry Maguire screeching, "Show me the money!"

Indeed, the whole issue of scarcity must be seen through the prism of what can be extracted at a reasonable cost.

This is a key point made by Professor John Tilton, a leading mineral economist at the Colorado School of Mines: "Long before the last barrel of oil is extracted from the earth's crust, costs would rise, at first curtailing but eventually completely eliminating demand. In short, what we have to fear is not physical depletion, where we literally run out of mineral resources, but economic depletion, where the costs of producing and using mineral commodities rise to the point where they are no longer affordable."

Aside from economic assumptions, reserve estimates vary as much as the personalities of the people who make them. These gravitate roughly into two ways of approaching the job.

"Under-promise and over-deliver" was Prime Minister Paul Martin's mantra during his successful business career, and this clearly carried through his days as finance minister. Then there is the "You gotta believe!" type of rosy optimism, more reminiscent of Martin's predecessor Michael Wilson, whose budget shortfalls were a lot bigger than Shell's.

Oil, politics and power have always been tightly intertwined, from the days of Andrew Carnegie and John D. Rockefeller through to the Getty and Bush families.

The amount of oil a nation has in the ground is like the size of its armed forces in that it is sometimes beneficial to exaggerate it, while at other times it pays to hide it.

The OPEC cartel is modelled on a much older organization, the Texas Railroad Commission, which controlled the flow of crude from the Lone Star state's oilfields from 1920 until 1970.

So plentiful was the supply that prices were supported by limiting the monthly volume to a fraction of its theoretical maximum. After 1970, when production peaked, the limits became moot and all wells were allowed to produce at full capacity.

OPEC started out following this protocol, but in the late 1970s it began to factor each member's reserves into the calculation of its allowable output.

Almost overnight, many OPEC members made huge upward revisions to reserve estimates in order to boost their output quotas. Was this wishful thinking or an admission that previous numbers had been grossly underestimated? Take your pick.

So if the amount of petroleum in the ground is so murky, do the figures for world production suggest we are reaching the top of a bell curve?

While production for places like Alberta and Texas do show a remarkably smooth curve to a defined peak, this one looks quite different. The main difficulty is that, nine times in the past 23 years, the rate of production has dropped from the previous year.

That being the case, how can we possibly know when the wave has crested for the last time?

Note that after 1979 it took a full 13 years for supply to reach a new high. Or was that in fact a peak in demand?

What would it take to knock oil down hard -- that is, trigger a sharper drop than the "moderation" in prices foreseen by many analysts and which is now way overdue?

It's easy enough to review what happened the last time prices slumped near $10 a barrel, given that it was a mere six years ago. Just pull out an old Contra from that period and you can see the pounding we took on Pioneer Natural Resources. On the supply side, Iraq was finally allowed by the U.N. to restart exports under the "oil for food" program. Kuwait, meanwhile, boosted production to recover from the war. OPEC's efforts to cut production quotas were met with opposition and cheating.

Simultaneously, on the demand side, there was the "Asian contagion," an economic retrenchment that temporarily trashed the fast-growing countries of the Pacific Rim.

A crash back to $10 does seem unlikely, but the potential risk cannot be waved away with impunity. Imagine, if you will, that a couple of years down the road Iraq is more stable and pumping furiously to rebuild the country.

All over the world the boost in capital spending increases production. In the U.S., Alan Greenspan, or his eventual successor, boosts interest rates to contain inflation and precipitates a recession. As oil prices fall, OPEC's solidarity disintegrates.

Countries like Nigeria and Venezuela, which don't enjoy the luxury of being able to institute production cuts the way the Saudis do, pump more to make up for budget shortfalls. Throw in a mild winter or two and suddenly Calgarians have more to worry about than hanging on to Jarome Iginla.

None of which is to say that such a scenario will play out, but only to point out that such a chain of circumstances is not far-fetched and could easily occur, even if the Hubbert hypothesis is by and large correct.

The ecologically minded point out -- quite rightly -- the preposterous juxtaposition between the hundreds of millions of years that it took oil to form on this planet and the fact that humans will probably use it all up in a couple of centuries.

Compared to the gradual depletion that will occur over several decades, supply and demand dance on a time line of only a few years. Less than a blink of the eye from a geological point of view, but for the patient investor, plenty of time to ride the cycles and profit from them.