Like a roller coaster ride, 2011 saw oil prices climb gradually, only to fall dramatically this last week. Here I offer my thoughts on some of the key contributing factors.
Let's begin with the relation between oil prices and the exchange rate. If the dollar depreciates by 1%, the dollar price of oil would have to go up 1% to keep the price paid outside the United States constant. This is a bit simplistic, one reason being that there is usually some third factor, such as a rise in incomes outside the United States, that is causing a change in both real oil prices and the exchange rate. Different factors affect the two series differently, so one might see a 1% depreciation correspond to an increase in dollar oil prices of more or less than 1%, or sometimes even an oil price decline. Between September 2009 and September 2010, a 1% depreciation of the exchange rate was associated on average with a 1.3% increase in the dollar price of commodities like oil or copper. The dollar rose about 3.5% against the euro between Wednesday and Friday, and the 4.5% decline in the price of copper could be pretty well explained by the exchange rate alone based on the recent correlations (3.5 x 1.3 = 4.5). But something more is involved in the 11% drop in the dollar price of crude oil observed those same two days.
Looking at the broader trend, the price of oil shot up 19% in February and March, during which the dollar depreciated against the euro by only 3%. The exchange rate can account for only a small part of recent movements in the price of oil.
Actual oil price and the value predicted by the exchange rate. Solid line: actual price of West Texas Intermediate (in dollars per barrel), Sep 1, 2010 to May 6, 2011. Dashed line: Sept 1 price times exp(1.3 times change in natural logarithm of exchange rate since Sep 1). Updates graph from Econbrowser Nov 10, 2010.
What I believe should instead be the first place to begin any discussion of recent oil prices is the broader global trend of supply and demand. The graph below plots world oil consumption over the last 15 years. This increased by 7.3 million barrels per day between 2000 and 2005, but by only 1.2 mb/d between 2005 and 2010. But very importantly, consumption by China increased by 1.7 mb/d between 2005 and 2010.
Total world oil consumption, annual, millions of barrels per day, 1995-2010. Data source: EIA.
Please note the necessary implications of this arithmetic: if China is consuming 1.7 mb/d more, but the world as a whole is only consuming 1.2 mb/d more, that means that people outside of China, as a group, have decreased our consumption by 500,000 b/d over the last 5 years. And the first question to ask anybody who claims that the price of oil has been "too high" recently, is, how much of a price increase do you think would have been necessary to persuade consumers outside of China to reduce consumption by a half-million barrels per day over a five-year period?
I've been talking about global consumption, though the EIA also reports data on world production. It's not accurate to conclude that if reported production exceeds reported consumption, then there must be excess supply with the difference going into inventories somewhere. The two series are collected from different sources, and the difference between reported production and reported consumption is often just reflecting errors in measuring the two series. But it's interesting to note that, from production data, it looks as if we're finally lifting above the five-year plateau, with the latest production figures showing significant increases relative to 2005 in countries such as the U.S., Brazil, Russia, Angola, Iraq, Azerbaijan, China, and Canada far in excess of the declines in the North Sea, Mexico, Venezuela, Indonesia, and Saudi Arabia over that period.
Total world oil production and consumption, annual, millions of barrels per day, 1995-2010. Data source: EIA.
We can also look at direct oil inventory data for the United States. The black curve in the figure below plots the average seasonal behavior of U.S. crude oil inventories. The green curve gives the values for 2008. Inventories were significantly below normal in the first half of that year, making it difficult to insist that the price at that time, though rising quickly and very high by historical standards, was higher than it needed to be to keep demand from outstripping supply. By contrast, the orange curve (2011 data) indicates that current inventories are currently well above normal, suggesting that, particularly given the recent production gains, a lower price would likely be consistent with the quantity consumed being equal to the quantity produced.
By. James Hamilton
Reproduced from Econbrowser