Iran should provoke the U.S. right up to the point where the marginal benefit from an additional unit of provocation - higher oil revenues from a rising risk premium - just equals the marginal cost:
Troubled Waters Over Oil, by James Surowiecki, The New Yorker: The past few months haven’t been easy for Iran’s President, Mahmoud Ahmadinejad. His refusal to halt Iran’s uranium-enrichment program led the United Nations to impose sanctions... Inflation in Iran has exploded... In the country’s recent municipal elections, Ahmadinejad’s political allies were crushed, and clerics and lawmakers have begun criticizing him in public. Worse still, through the second half of 2006 the price of oil tumbled almost thirty per cent, a disaster for an economy as dependent on oil revenue as Iran’s. ... And then the Bush Administration said that it had authorized U.S. troops to detain or kill any Iranians found to be working with the Iraqi insurgency, and dispatched a second aircraft-carrier group to the Persian Gulf, sparking rumors that a military strike against Iran was in the works.
This latest confrontation with the U.S. should have been the capper... Strangely, though, it may instead have brought about an upturn in his fortunes. Soon, oil prices started to rise, jumping twenty per cent in just two weeks. As a result, the Iranian regime suddenly has an extra twenty million dollars or so to spend every day, a windfall that will help Ahmadinejad to placate his critics and solve some of his country’s more pressing economic problems.
The jump in oil prices wasn’t entirely a geopolitical phenomenon—the cold snap in the U.S. was also a big factor—but it was driven in part by an increase in what oil traders call the “risk premium.” ... In the current confrontation between the U.S. and Iran, ...[there is a] a perverse set of incentives: whenever the U.S. says things that make a military conflict with Iran seem more likely, the price of oil rises, strengthening Iran’s regime rather than weakening it. The more we talk about curbing Iranian power, the more difficult it gets.
It’s hard to measure the risk premium exactly, but most estimates suggest that in the past couple of years ... it has accounted for somewhere between ten and twenty dollars on each barrel of oil. ... Ten months ago ... when Iranian leaders were talking about their progress in enriching uranium, and were threatening to attack Israel in response to any U.S. attack, the price of oil rose to more than seventy-five dollars a barrel. The economic consequences of this are not trivial; in the past few years, the inflated risk premium has given Iran tens of billions of dollars that it would otherwise not have had.
This helps Ahmadinejad enormously, because Iran has made huge commitments to government spending that can be kept only by relying on oil revenue. ...
The persistence of the risk premium means that Ahmadinejad ... has an economic incentive to say confrontational things that spook the oil market. But the effect of his pronouncements is limited, because traders know that self-interest is likely to keep Iran from doing anything that would cut off the supply of oil. What really keeps the risk premium high is the American penchant for public responses to Iran’s provocations. So cooling down the martial rhetoric—even if we plan to take military action eventually—would likely bring oil prices down for a time, making Iran weaker. ... Talking tough may look like a good way of demonstrating U.S. resolve, but when tough talk makes our opponent richer and stronger we may accomplish more by saying less.