Must a windfall profit tax be a one-time affair?

Continuing on Monday's windfall-profits-taxes-are-hard theme, here's Australian economist Andrew Leigh (via Mark Thoma):

From an economic standpoint, the strongest argument for a windfall tax is that it has the potential to be non-distortionary. A one-off windfall tax levied on past profits should not change firms' behaviour, since it does not affect future costs and prices.

Leigh is (or was; his essay was published back in June) writing about coal mining companies in Australia, not oil companies in the U.S., but the economic forces at work are pretty much identical:

The miners would ... be quick to contest the claim that a windfall tax can have little economic impact on their future decisions. And it is true that as soon as mining companies hear of the tax, their decisions will change. As a result, surprise windfall taxes are more economically efficient than anticipated ones. This may be one issue on which a full and robust public debate does not lead to a better outcome.

Companies are also likely to raise the spectre of repeated raids on their revenue. Having been levied once, what is to stop a windfall tax being imposed again? To counter this, the government must be clear that the present minerals price increases are a once-in-a-lifetime event, and so is the windfall tax. The less credible this claim, the more the tax will deter future investment in the sector.

By that reasoning, the U.S. has already used up its once-in-a-lifetime windfall tax allotment (although the 1980-1988 tax was levied too late and hardly picked up any windfall at all). Once in my lifetime, at least. To quote economist Thomas Schelling, testifying in 1979 before a Congress:

I wish it were possible to tax away today's and yesterday's windfall profits without causing any anticipation that we may do the same thing next year, and the year after, and ten years from now. But you cannot forever treat bygones as bygones without people anticipating that you'll do it again.

Obama has been deliberately vague about his windfall tax plans, but the main windfall tax bill in Congress doesn't even pretend to be a one-time affair. It calls for the tax to be imposed whenever "the adjusted taxable income of the applicable taxpayer" is higher than the "reasonably inflated average profit for such taxable year," unless the company is spending lots and lots of money on exploration, alternative fuels, etc. And what's the "reasonably inflated average profit"?

the average of the adjusted taxable income of such taxpayer for taxable years beginning during the 2002-2006 taxable year period (determined without regard to the taxable year with the highest adjusted taxable income in such period) plus 10 percent of such average.

A recurring windfall tax on oil profits seems to make economic sense only if you believe oil is basically over: There will be no major new discoveries or drilling/refining breakthroughs, and thus no significant cost to discouraging exploration and investment. The investment behavior of the big oil companies (who give more money back to their shareholders than they spend on exploration) would seem to indicate that they share this view at least partly. But is further discouraging oil exploration and development good policy? It would be, I guess, if we were just a nudge or two away from a glorious alternative fuel future. Are we?