Well, a futures contract is a bet about the future price. It has no, zero, nada direct effect on the spot price.
He can't mean that.
Think of the foreign currency market. If speculators bid up the future price of Japanese yen, then the spot price of Japanese yen will go up. And you won't see any particular pattern of inventories among currency dealers. The inventory issue is much closer to a red herring than to the decisive empirical data that Krugman maintains it to be.
Krugman and I should both listen to the podcast with Tyler Cowen, linked in this post. Cowen's point is that when you disagree with someone, you tend to assign a probability that you are wrong that is too low.
Although Krugman and I differ on political issues, I cannot think of anything political in this argument. It seems purely technical.
My views on the oil market are almost the exact opposite of Krugman's. I believe that the futures price has to be the key determinant of the spot price. Because oil is a non-renewable resource, the oil market has to reflect expectations for demand and supply over the entire future time horizon, and those expectations ought to be embedded in futures prices.
Another way to think about futures markets is that they substitute for a central planner in the oil industry. If we had a central planner, he would have to decide how intensively to search for new oil reserves and how quickly to extract the oil from known reserves. Absent a central planner, the futures market sends those signals.
Go through the following thought-experiment. Suppose that the U.S. government were to stop filling its Strategic Petroleum Reserve (SPR) and perhaps even sell some of its existing reserve. However, suppose also that the U.S. government were to buy an equal amount of oil in the futures market.
If Krugman is right, and all that matters is that spot supply of oil, then this should lower the current price of oil. If I am right, then the effect on oil prices would be approximately zero.
I did not say which futures contract that the government speculates in. Suppose that there is a futures contract dated one week from now. Surely, if the government sells oil today and commits to buying an equal amount one week from now, there is essentially no effect on price. But Krugman's logic seems to force him to say that the spot price of oil will plummet when the oil from the SPR hits the market.
So this is an instance where I think that the probability that I am wrong is low.
I agree with Krugman that blaming oil speculators for the high price of oil is unhelpful. The politicians make it sound as though there has been a sudden outbreak of greed among oil speculators. Instead, there has been a change of expectations about future supply and demand. From what I can tell, there was no real news to cause this change in expectations. Either speculators were badly wrong six months ago or they are badly wrong today. It is more likely that they were wrong six months ago, but the probability that they were closer to correct then is far from zero.