The Horizon Accident, Peak Oil and the Law of Diminishing Returns

Includes: COSWF, CVX, DNR, SU
by: Jonathan Bernstein

A number of investment newsletter writers would like the peak oil theory to go away. As their argument runs, the worldwide annual rate of oil production will not peak because higher prices will induce the invention of enhanced technologies and bring forth new oil supplies. Some of these folks even seem to believe that higher oil prices will allow U.S. oil production to rebound, though no one comes out and states that.

Before I deal with this “cornucopian” argument, kindly let me review my understanding of what peak oil is. The Shell Oil geologist M. King Hubbert predicted in 1956 that U.S. onshore oil production would peak in 1970, because he believed that annual oil production in a given region follows a bell curve over time. Annual production would expand, peak, and then eventually decline. “Hubbert’s Law” says that when about half of a reserve’s original oil in place has been lifted, annual production rates fall.

As it happened, U.S. oil production peaked at 9.6 million barrels per day in 1970, precisely when Hubbert said it would, and except for the temporary bounce in the early 1980s from Prudhoe Bay, it has been falling ever since. In 2008 we produced 4.95 million barrels per day, the first dip below the 5 million mark. Here are the figures:

U.S. Field Production of Crude Oil(Thousand Barrels per Day)
Decade Year-0 Year-1 Year-2 Year-3 Year-4 Year-5 Year-6 Year-7 Year-8 Year-9
1850’s 0
1860’s 1 6 8 7 6 7 10 9 10 12
1870’s 14 14 17 24 30 33 25 37 42 55
1880’s 72 76 83 64 66 60 77 77 75 96
1890’s 126 149 138 133 135 145 167 166 152 156
1900’s 174 190 243 275 320 369 347 455 488 502
1910’s 574 604 609 681 728 770 822 919 920 1,037
1920’s 1,210 1,294 1,527 2,007 1,951 1,700 2,112 2,469 2,463 2,760
1930’s 2,460 2,332 2,145 2,481 2,488 2,723 3,001 3,500 3,324 3,464
1940’s 4,107 3,847 3,796 4,125 4,584 4,695 4,749 5,088 5,520 5,046
1950’s 5,407 6,158 6,256 6,458 6,342 6,807 7,151 7,170 6,710 7,054
1960’s 7,035 7,183 7,332 7,542 7,614 7,804 8,295 8,810 9,096 9,238
1970’s 9,637 9,463 9,441 9,208 8,774 8,375 8,132 8,245 8,707 8,552
1980’s 8,597 8,572 8,649 8,688 8,879 8,971 8,680 8,349 8,140 7,613
1990’s 7,355 7,417 7,171 6,847 6,662 6,560 6,465 6,452 6,252 5,881
2000’s 5,822 5,801 5,746 5,681 5,419 5,178 5,102 5,064 4,950

And note well, the crude oil price in 1970 was about $1.35 per barrel, depending on the grade you pick. The current price is over fifty times higher — okay, call it “only” 10 times higher if you adjust for inflation. Despite the higher prices, U.S. production has decreased year after year almost without relief. I haven’t heard the cornucopians explain that.

Of course there will always be some oil left in the ground. Hubbert did not say that U.S. or worldwide production would go to zero any time soon. He simply said that once the peak was passed it would no longer be possible to produce at the same annual rate.

After Hubbert succeeded in forecasting the U.S. peak, other geologists tried to forecast the worldwide production decline curve. There is a range of predictions regarding the date when worldwide production will peak. Some experts like Princeton professor Jeffery Deffayes even say we’ve already peaked. But within the universe of respected energy specialists, nearly all agree that world production will peak at some point (sounds obvious, doesn’t it?). Even Cambridge Energy Research Associates (CERA), which is known for optimistic production forecasts, agrees with this statement. They just predict a later peak than most others do: Around 2040. It’s hard to be more sanguine than CERA but the cornucopians do try.

The cornucopian argument has two glaring weaknesses. First, Mother Nature doesn’t care how much money we humans pay one another for oil. The oil that is in the ground now is the amount we get to enjoy. Higher prices make it economic to drill for oil in remoter regions and to develop new extraction methods but they don’t change the amount of oil in place.

The second weakness is a bit more subtle. Though peak oil is a geological theory, it also follows from a bedrock principle of economics, the law of diminishing returns. That principle is associated with 19th century economist David Ricardo.

Ricardo declared that in agriculture, there were diminishing returns to cultivation. Farmers would first till the most fertile land, and as population grew farmers would begin working less productive land, until farming the most marginal acreage became uneconomic. The crop yields wouldn’t justify the effort needed to till such parcels, even if the farmer added additional labor, fertilizer, or irrigation.

Diminishing returns show up in oil just the way they do in agriculture. A concept called the Energy Return on Energy Invested (EROEI) illustrates the diminishing returns to oil exploration effort through the history of the oil industry. EROEI measures the energy content of an oilfield divided by the energy needed to get the oil.

In the 1930s, U.S. oil production is now estimated to have yielded an EROEI of about 100; The number dropped to 30-ish in 1970 and to the range of 11-18 by the year 2000. Think about what happened when Colonel Drake drilled the first oil well in Pennsylvania, or when Chevron (NYSE:CVX) made its early Saudi Arabia discoveries. All they had to do was stick a shallow well in the ground and oil flowed freely to the surface. You’d expect a very high yield on prolific reserves that demand next to no effort to get.

As one might expect, prospectors found the easy oil first, and as one drills deeper, EROEI falls. But even in the life of what was once an easy oil field EROEI tends to fall over time. As an oil field matures, natural pressure falls until oil no longer flows to the surface without a push. Oil companies extend the life of oilfields by injecting water or carbon dioxide to force oil toward the well bore. Another enhanced oil recovery (EOR) technique is directional drilling, where the angle of the well bore is set to maximize the length of pipe that has direct contact with the layer of oil bearing rock being tapped.

EOR requires additional energy expenditure and so it reduces EROEI. This increased energy expenditure increases the denominator in the oilfield’s EROEI ratio. Eventually the oilfield owner can get nothing more out of the reservoir. He has passed the point of diminishing returns like Ricardo’s farmer.

Higher prices have made resources such as the tar sands of Canada and Venezuela economic to exploit even though their EROEI is only about 5, quite a comedown from Saudi Arabia or East Texas in their glory days. Tar sands deposits are not even oilfields, at least as one normally thinks of oilfields. Typically they are more like strip mines, where companies like Suncor (NYSE:SU) and the Canadian Oil Sands Trust (OTCQX:COSWF) dig up a solid called bitumen. They load the bitumen into special, giant trucks, crush the big pieces into small pieces, separate out the sand and debris, and then run the separated product through an “upgrader.” The early stage processing requires substantial inputs of water and heat. Only then can the crude be refined into products such as gasoline, heating oil, and jet fuel.

The Horizon drilling rig and its tragic explosion illustrate diminishing returns in another way. One needs elaborate and expensive equipment to drill for oil that is miles below a seabed that is itself thousands of feet below the ocean’s surface. These rigs take lots of energy and steel to make. Horizon even needed to expend energy by continuously running powerful thrusters to keep it in position. It wasn’t feasible to anchor the rig to the ocean floor in that depth of water. And in difficult locations like this, as we have been reminded, workers die when things go really wrong. So while higher prices make deepwater drilling feasible, one gets less usable energy per unit of effort than from conventional on shore oilfields, especially when all costs are factored in. Ricardo would understand all too well.

No one doubts that higher prices promote EOR, and EOR extends the life of aging oilfields. EOR also enables midsized outfits like Denbury Resources (NYSE:DNR) to work over largely depleted fields that the oil majors no longer care to bother with. I’ll also grant that higher prices make that business economic. Nor does anyone doubt that higher oil prices encourage people to substitute coal or gas for oil, to increase energy efficiency and develop new energy sources.

But if you want to say that U.S. oil production has not peaked, you have to go really far out on a limb. You have to argue that we will return to, and surpass, the halcyon days when the U.S. could produce ten million barrels per day. Good luck selling that story.

So in my view, the total oil supply is not only finite, it gets harder to obtain with each passing year. And that’s why I continue to like oil companies with long-lived reserves that are also located in politically stable countries. I particularly like the major players in Canada’s tar sands; of course I liked them better at May 2009 prices and said so at the time, here. But if you have to own some stocks, these are the kinds of companies that will do well long term even if the market’s correction deepens. If stocks recover from last week’s turbulence and keep going up, so much the better.

Dislosure: Long SU