Jim Hamilton on what determines prices in oil markets:
...The Wall Street Journal carried this account last week:
Oil prices dropped more than $3 in less than a minute late in the trading day on Monday, just as trading volume spiked. The move also dragged down prices of gold, copper and even the euro.
"Traders were looking like deer in the headlights," said Peter Donovan, a floor trader... "I called four different desks, and they all said, 'we don't know.' " ...
The move sparked talk of an erroneous trade—called a "fat-finger" error in industry parlance—or a computer algorithm gone awry.
Fat finger or no, there was an even bigger drop on Wednesday...Those who doubt that oil prices are determined solely by fundamentals would naturally ask, what aspect of the supply or demand for oil could have possibly changed in the course of less than a minute last Monday? The obvious and correct answer is, there was no change in either the supply or the demand for physical oil over the course of that minute. The minute-by-minute price of a NYMEX contract is determined by how many people are wanting to buy that financial contract and at what price, not by how much gasoline motorists burned in their cars that minute. But since changes in the price of crude oil are the key determinant of the price consumers pay for gasoline, doesn't that establish pretty clearly that the whims or fat fingers of financial traders are ultimately determining the price we all pay at the pump?
In one sense, the answer to that question is yes-- last week's decline in the price of crude oil will soon show up as a lower price Americans pay for gasoline. But here's the problem you run into if you try to carry that theory too far. There are at the end of this chain real people who burn real gasoline when they drive real cars. And how much gasoline they burn depends in part on the price they pay-- with a higher price, some people use a little bit less. Not a lot less-- the price of gasoline could change quite a lot and it would take some time before you could be sure you see a response in the data. That small (and often sluggish) response is why the price of oil can and does move quite a bit on a minute-by-minute basis, seemingly driven by forces having nothing to do with the final users of the product.
But if the price of oil that emerges from that process turns out to be one at which the quantity of the physical product that is consumed is a different amount from the physical quantity produced, something has to give. Indeed, the bigger price drops we saw on Wednesday followed news that U.S. inventories of crude were significantly higher than expected ...