Elliptical Research Contribution 2008.5 - supplemental:

Accounting for the Time Delay Between Oil Drilling and Production

Supplement to
Why “Drill, Baby, Drill!” is Not a National Energy Policy


Timothy D. Kailing
Elliptical Research
September 2008
As stated in the main article, empirically, for United States oil production there is a 4-6 year lag in the
relationship between oil drilling and oil production. We can account for this by using the methods in the main
article but adding a lag to the production. Rather than simply relating:

                             
 (oil production at time t per rig at time t) vs. (the number of rigs at time t)

Instead, we graph the relationship of

                             
 (oil production at time t+5 per rig at time t) vs. (the number of rigs at time t)

Where t is in years in all cases. Doing this gives the following graph:
Again, a power law relationship is apparent in the data. Looking at this data logarithmically, as before gives the
following relationship:
Note that the relationship for the main (older) data in blue has strengthened a bit: the variance explained has
increased to 94%, compared to 91% for the case not accounting for the lag. Though the lag was determined by
other methods, this supports the case for about a five year lag found empirically from drilling to production. (One
important point: by no means does this mean that such a five year delay is somehow a universal lag, or even the
most common lag, between drilling and production in individual cases of oil exploration and development; it just
means that, empirically, it is the average effective lag seen in the overall data.)

For the recent data in red one might naively have some hopes, when comparing to the delay-agnostic analysis in
the main article. While the relationship is still strongly negative, the slope is a little less negative (-1.02
compared to -1.30) and the R-squared is notably lower (0.71 compared to 0.95). Perhaps the relationship is
breaking down?

Unfortunately it does not appear to be the case. By adding the five year lag, we lose the last five years of data,
because while we know the number of rigs drilling these last five years, we don’t yet have production for 2008-
2013. However, it turns out that we can make a reasonable estimate the next five years of production
independently of the relationship between drilling and production. Here’s how. If we graph the production data
vs. time since 1985 we find that, despite the great variation in oil prices and drilling activity lately, the relationship
has been very simple over these last 30+ years:
A simple model of exponential decay in production fits these data very well, explaining over 98% of the variation
in domestic oil production. We then use this model to project the next few years. These projected data are
shown in orange on the graph.

If we take these production projections, which were produced independently of any drilling information, and use
them to project the production from the drilling we know from the last five years (including an annualized
adjustment of the drilling in 2008 so far) we get:
With the projected data included, the relationship is again very strong, with rig number explaining an impressive
94% of the oil per rig data. And, like before, the relationship is strongly negative, with the steepest slope of all at
-1.35.

Using projected data always makes conclusions conditional. But in this case it is simply that unless the
exponentially decaying U.S. production turns sharply up in the next few years—of which there is no indication,
despite the incentive of recent high prices—there is no reason to expect that drilling will have any significant
impact at all. It is important to keep in mind that even the most optimistic interpretation of the relationship of
drilling and production, including excluding all projected data, still gives a profoundly pessimistic assessment of
the feasibility of significantly increasing domestic oil production by means of even an enormous increase of
drilling activity. And, as the main article points out, whether doing this is in the United States’ long term interest is
highly questionable.

If we are to solve our domestic energy problems, it must be through other means than increased drilling.





Back to the main article on the relationship of oil drilling and oil production
Timothy D. Kailing is the principal at
Elliptical Research.

He wears another hat as an
advocate for the benefits of early
literacy in children; in this effort he
authored the book:
Native Reading: How to Teach Your
Child to Read, Easily and Naturally,
Before the Age of Three.
Copyright © 2008 Timothy D. Kailing