Speculation & Oil Price Fluctuations
A more plausible theory is however that the cause is a low short-term price elasticity for demand and supply. The flip side of a low price elasticity is that changes in demand (or supply) not caused by price movements will cause very big price fluctuations.
If say, 5% of demand suddenly for some reason (say a sharp cyclical downturn) disappears and say that a moderate price decline will only cause an increase in supply or decrease in demand of a few tenths of a percentage points, then it is clear that in order to avoid surpluses, prices must decline really dramatically.
That was exactly what we saw both in the great run up in the price of oil to early July 2008 and in the dramatic (75%) price decline that we saw until December 2008.
How then do we know if a certain price fluctuation is due to speculation or other factors? The answer to that question is: by looking at inventory changes. In order for speculators to increase/lower prices they must increase/lower inventory levels. The reason for that is that if they take long positions in futures contracts they are left with two choices. Either they unwind that long position in which case they will lower the price to the level where it would have been if they hadn't taken that long position in the first place, or they'll have to keep the delivery they paid for as inventory. Meaning that in order for speculation to affect the price, inventories must increase.
If you look at the evidence (For the original data go here and then click "Complete History XLS) , you can see that in the months before the July 2008 peak, inventories fell significantly(about 50 million barrels below year ago levels), while in the coming 5 months inventories rose sharply (shifting to being nearly 50 billion above year ago levels). Since then they have been more or less flat adjusted for seasonal patterns. This means that speculators helped reduce the price increase that we saw until July 2008, and then helped reduce the price decrease until December 2008. The increase in price we've seen after that was by contrast purely driven by underlying supply and demand movements.
So while Paul Krugman was right noticing that inventories are now higher than a year ago, he completely misses the point. That inventory build up was happened not during the latest price rally but during the preceding price decline. Meaning that it was during the dramatic price decline that speculation helped raise prices, not during either the preceding or following rallies. Adjusting for likely base effects, inventories have if anything declined during the recent rally.