Nearing the Old High?

Whenever the price of crude oil breaks its previous record in nominal prices, we see comparisons to the all-time high of the early 1980s, adjusted for inflation. When it passed $80 the other day, I saw several such comparisons. Only one problem: they differed by more than a dollar per barrel among them. That's more than a rounding error, considering that price data are given to the penny, and the GDP deflator commonly used for such calculations goes out three figures beyond the decimal point--five when you realize it's quoted as a percentage. What could be causing these disparities, and what is the true, inflation-adjusted oil price high? That turns out to be a decidedly non-trivial question, and there's really no single correct answer. The best I can come up with is around $91/barrel, which is much lower than recent estimates over $100/barrel that probably rely on the consumer price index.

Explaining why this isn't nearly as simple as it looks requires a bit of background on the different grades of oil, how the market works now versus how it worked in the early 1980s, and what we mean by inflation adjustment. This isn't intended as an economic dissertation, however, so I'm going to paint it in broad strokes:

Start with oil, itself. There are probably more different types, or grades, of it than there are rock bands with funny names. The grade the media generally refer to is West Texas Intermediate (WTI), a generic mix of light, sweet domestic crude oil streams (plus a few specified import grades) with sulfur content below 0.42% and API gravity (a measure of density) between 37 and 42 degrees. While it has been the primary benchmark crude oil type for the US and the world for a very long time, it's not particularly representative of what most refineries actually run, though it was more representative 25 years ago.

Now consider the market. Today, the WTI price usually refers to the settlement price for the front-month futures contract on the New York Mercantile Exchange. Unfortunately for anyone trying to compare current and historical oil prices, the NYMEX only started trading crude oil in 1983, a couple of years after the price had peaked. To compare WTI before 1983, you must look at "first purchaser" wellhead prices. The best publicly-available proxy for those that I could find is "posted prices", which are not actual transaction prices, but price schedules published by companies soliciting offers to sell them oil. To complicate matters further, actual transactions typically occur at a premium to postings, referred to as "P-plus". For example, "XYZ agrees to buy 10,000 barrels per day of WTI at Cushing, OK for the month of June 1987 for P-plus 50 cents." As quaint as that sounds in an era of real-time electronic trading, I understand that a fair amount of crude is still transacted on a P-plus basis, for various reasons.

Next consider what we mean by inflation. How relevant is the commonly-used consumer price index to the wholesale price of a primary industrial commodity? Not very, even if oil isn't just any commodity. I'm a lot more comfortable using the GDP deflator, which is a very broad measure of the impact of inflation on nominal prices across the entire economy. It might be equally legitimate to argue for using some other index, such as nominal GDP per capita or relative share of GDP.

The highest posted price for WTI during the first energy crisis was $39.50/barrel from April-July 1980. (While there's better data available for imported crude oil in the same period, trying to equate that to current WTI requires all kinds of additional assumptions, and I'd prefer to avoid those.) Applying the ratio of GDP deflators to the posted price yields $87.92 in 1Q2007 dollars. If we assume that actual transactions in the peak month of 1980 were probably done at P-plus $1.00--about as much as the market would take before competitive forces pushed the postings up--then we get to $90.14. But that's still a wellhead price, so we'd need to add something for gathering, handling and transportation to arrive at a figure that equates to NYMEX WTI at Cushing, OK. Call it a buck, and we're at $91 and change.

If I've done my sums properly, we're closer to the all-time high oil price than some analysts are suggesting. The puzzle, of course, is how prices can be this high for this long without putting the economy into a deep recession, such as we saw in the previous energy crisis. Part of the answer is found by translating that July 1980 oil price based on its relative share of the 1980s GDP, compared to today's, rather than using the deflator. On that basis, oil at $80 still has a long way to go to match its equivalent of $187/barrel in 1980.

Update, 11/7/07: The Washington Post described similar difficulties in coming up with an inflation-adjusted all-time high, reporting different results from three groups, the IEA, CERA, and the Energy Information Agency of the US DOE. The EIA's figure comes closest to mine at $93.48. They applied the same GDP deflator, though they used a different starting point, the average monthly refiner acquisition price in Jan. 1981. CERA used WTI postings, as I did, but inflated at the CPI to arrive at $99.04.