by Mike Kimel
The Effect of Oil Prices on Oil Drilling in the U.S.
Oil markets have changed dramatically in the past couple of decades or so. Except for a few years following the second Oil Embargo – prices got as high as $60 (in 2005 $) a barrel in 1981 – real prices have tended to be below $25 a barrel through about 1999. Conversely, 2003 prices have been higher than that – in some years quite a bit higher. Now, there are all sorts of explanations for this big change we’ve observed over the past few years, ranging from Peak Oil to the war in Iraq to the rise of the BRICs to market manipulation, but that’s the point of this post…
Instead, I want to look at the relationship between the price of oil and the number of oil rigs, and how that relationship has changed over the last couple of decades or so. Oil rigs, of course, are the machines that dig oil wells; once a well is completed and has begun production, the rigs are removed and either go into storage or move on to drilling another well. Data on the number of oil rigs in operation in the United States used in this post comes from Baker Hughes. Regular readers know I normally do not use data from private sources, but Baker Hughes data are as close to “official” as possible, as the figures you’ll find the Dep’t of Energy’s website on rigs originate with Baker Hughes. Rig count data comes out weekly and begins in mid-1987. I’ve taken annual averages beginning in 1988. Price data are annual averages from Table 5.18 of the 2010 Annual Energy Review put out by the Department of Energy. That data runs through 2009.
Now, a few details. Some time toward the end of the last millennium and the first few years of this one, there was a revolution in the drilling of oil (and natural gas). Two new technologies, hydraulic fracturing and horizontal and/or directional drilling, changed everything. Hydraulic fracturing is the fine art of pumping sand and water mixed with small amounts of some fairly toxic chemicals at high pressure to break apart some types of rock formations (usually shale) in which oil (or gas) is trapped. And the other thing available now are rigs that don’t just drill straight down, but instead can drill sideways once they reach the desired depth.
There is no fine line we can point to and say: this is the point when these two technologies became widespread. Instead, based in part on the numbers, I’m just going to say that until about 1998, those technologies were rarely used in the US, but after 2002 they were in widespread use. So… let me put up two graphs. The first one shows the relationship between the rig count and the price of oil from 1988 to 1998, and the second shows the same relationship between 2002 and 2009.
(A few comments to the statistically oriented… yes, I know that a single equation regression is nothing more than a correlation, but this was for illustrative purposes. And before you mention autocorrelation, take a look at the graph again and think of exactly what would change if I did correct for it.)
So what does all this mean? A few comments:
1. The relationship between prices and rig count exists because as world prices rise, U.S. producers have an incentive to drill more.
2. In the first period, for every dollar increase in the price of a barrel of oil, on average 25 rigs were added in the U.S. In the second period, for every dollar increase in the price of a barrel of oil, on average only 3.5 rigs were added.
3. Part of the difference noted in 2. is just due to the fact that rigs are so much more efficient today than they were a decade and a half ago.
4. Another part of the difference noted in 2. is that there are only so many resources available to install new rigs in the U.S.
5. Yet another explanation for the difference in 2. may be price volatility; given price fluctuate so much these days, prices today aren’t as indicative of prices in six months or a year as they used to be.
6. Drilling for oil is a capital-intensive and risky operation. The relationship observed in comment 2. might be even greater were it not for the low interest rates prevalent in the second period.
7. The fit is much better (i.e., the relationship between price and rigs is much tighter, as there are fewer points far off the line) in the second period.
8. Do 2. and 7. indicate that perhaps oil drillers are becoming “more professional”?
9. Should this serve as a bit of an automatic stabilizer on price volatility? In other words, do the volatile oil prices reduce volatility in oil output, which in turn might reduce price volatility?
10. One other thought, only semi-related, and I’m not sure how it fits. In the oil market, you can get a lot of price volatility with even a small change in output. If world output falls by, say 1%, there are a lot of users without that many good substitutes (at present) willing to bid up the price on the marginal unit.
And one last thought…. does any of this say anything, one way or the other, about the notion of Peak Oil?
A few notes. First, full disclosure – I am not authorized to speak on its behalf, nor do I necessarily see the big picture, but I believe the company I work for would benefit from increased regulation of hydraullic fraccing. Second, the idea of looking for a relationship between prices and the means of production of a similar commodity came from Craig Truesdell. I’ve found Craig’s insight seems to provide useful intuition in a lot of markets.
Cross-posted at the Presimetrics blog.