Fixed View Of Resource Limits Creates Undue Pessimism

M. A. Adelman, Michael C. Lynch
Massachusetts Institute of Technology
Cambridge, Mass.

Ever since the human race began to dig for minerals, there has been worry about exhaustion. In theory, the more we remove from a limited stock, the less remains, and the harder it becomes to locate and extract additional amounts. Exhaustion forces us to roam farther and dig deeper, going from good deposits to bad and from bad to worse. Hence costs and prices must rise.

But in the long run, prices of minerals have if anything decreased. Some powerful force is at work to offset depletion and to make them easier and cheaper to find and develop. The limits, if any, seem to be getting more distant.

Many in industry and in government continue to think oil and gas are a big exception. They think the limits are known: proved plus probable plus undiscovered reserves. We are allegedly getting closer to those limits all the time. Reserves are not being replaced fast enough, or "discoveries" are lagging. The numbers are supposed to be scary.

For many years, we have been warned:

Most of the world's oil is in a few large fields, now nearly all of them found. Discoveries are the lowest in decades.
Investment is focused on more intensive development of existing fields, so the industry is running out of new discoveries for future development.
Oil companies are not replacing their reserves.
The decline rate is high in areas like the Gulf of Mexico and the North Sea, and investment is not offsetting it.
Technological advances are important but have only limited application. They cannot create oil which does not exist.
The lag time for development of new fields is 8-10 years, so no supply surprises are likely.
Costs are low because the service industries are depressed. Rig rates, for example, don't cover replacement costs, and eventually costs must rise.

Each of these propositions has some truth to it. They all point to the resource barrel being scraped harder, to rising costs, and lower output. Moreover, in the last 15-odd years, inflation-adjusted oil prices have fallen about 75%. The output drop should have been tremendous if depletion were raising costs.

For many years now, nearly every forecast has been: an early peak, then in 3-5 years decline in virtually every place but the Persian Gulf. Instead, oil output has actually grown-especially in areas outside the Organization of Petroleum Exporting Countries with the least resources.

Repeatedly, the forecasts are revised with a higher and later peak. International Energy Agency reported last year: "The Peak Recedes Again.^"1 Consistent undershooting means an inherent bias.² A "repeating surprise" should be an oxymoron.

These estimates of declining reserves and production are incurably wrong because they treat as a quantity what is really a dynamic process driven by growing knowledge. To know the limit to oil reserves and output, we must first predict future earth science and technology. This is impossible. Repeated attempts to do it have generated repeated bad estimates.

The three primary supply forecasting methods have the same source of error.

The bell-shaped curve

M. King Hubbert contributed the famous bell-shaped curve (Fig. 1) [31830 bytes].³

The area under the curve equals total ultimately recoverable resources (URR). Fit a curve to the historical production data on the left-hand side of the zero line (the present), and read off future production on the right. The curve is appealing, but everything to the right side of the zero line is being assumed.

Hubbert correctly predicted that U.S. crude oil output would peak in 1970. But was it the result of resource exhaustion, or of cheaper imported oil now freely available?

For the U.S., Hubbert in 1974 estimated URR at 170 billion bbl. Production to date has already been 170 billion bbl, proved reserves are 20 billion bbl, and annual accretions above 2 billion bbl. Discoveries continue. Output in 1996 was about twice Hubbert's forecast.

U.S. Geological Survey current estimates of URR are 250 billion bbl. Hubbert estimated 1,250 billion bbl for the whole world; USGS now says 2,400 billion bbl.4 The right-hand side keeps getting bigger.

For U.S. natural gas, Hubbert forecast a 1975 production peak at 14 tcf/year. In 1996, production was 20 tcf and rising. The delay was due to 30 years of misguided price control and end-use regulation. He estimated 850 tcf URR; for 1986, USGS estimated 1,403 tcf URR.⁴

The Hubbert Curve has been used to predict falling Russian production.⁵ In fact, Former Soviet Union output stagnated then fell because of a distorted economic system. When and if the FSU governments provide stable laws and taxes, there will be fresh investment to provide new reserves and capacity, especially in the Caspian area. No reforms, no big new reserves.

Using estimated URR to predict discoveries or cost trends⁶ is topsy-turvy. URR is the end result of cost changes brought about by changes in knowledge. Masters, et al., forecast non-OPEC, ex-U.S., and ex-U.S.S.R. on a high, low, or middle path. The high would peak in 2001, at 24 million b/d. Actual 1996 production was already 26 million b/d and rising.

Field-by-field forecasts

Some forecasters have aggregated existing fields plus expected development, usually assumed as small because of long lead times.

In 1986 Petroconsultants did a detailed analysis of non-OPEC reserves and warned that a production decline in these countries was "imminent" and "unstoppable" and would occur well before 1990. (The Soviet Union, however, would be stable.)⁷

This was not only wrong, it was the contrary of the truth. Ten years later non-OPEC produced 15% more; outside the U.S., 35% more. Over the last 5 years non-OPEC has met more of world demand growth than has OPEC.

Curcio⁸ (1989) and Cambridge Energy Research Associates⁹ (1991) also were much too low. North Sea underestimates are too numerous to mention. A recent IEA report¹⁰ now forecasts offshore expansion worldwide. The Gulf of Mexico, after a long decline to 740,000 b/d in 1991, made 1 million b/d in 1995 and is expected to make 2 million b/d by 2000.

Two factors appear to drive the forecasts off track. First they assume unrealistically high production decline rates.¹¹ Second, they ignore branch, satellite, or improvement projects. Some of these acorns become oaks. It is time to look at the relation between reserves and production.

Proved reserves

In a given reservoir, the proved reserve is determined by three facts: peak production, the decline rate, and the economic limit, where production stops.¹² The area under the curve is the proved reserve. It is an inventory, paid for up-front by drilling and nondrilling investment.

(Fig. 2) [30392 bytes] traces development of reserves of Oseberg field off Norway. Not many fields are as big as Oseberg, but every one has a declining production curve, which is usually bumped up by at least three kinds of incremental investment. One type improves the recovery factor. Second, more knowledge of local geology points to nearby reservoirs. Third, as the network of pipelines and supply and service centers grows more dense, more marginal prospects become profitable, and investment in them adds to reserves. Onshore drilling cost in the U.S. and Canada has long been half or less of cost elsewhere. The thickening network is responsible for much of the North Sea growth.

(Fig. 3) [26478 bytes] shows 20% of 1995 U.K. production was from fields discovered before 1980 but not developed until after 1990.

In 1970, BP found the Forties field, rating it at 1.8 billion bbl of proved reserves. By 1995, additional investment had added many small and large fields to the "Forties system," which had already produced 3.6 billion bbl and had remaining proved reserves of 2.8 billion bbl.10 It is a safe bet that system reserves will continue to grow.

The results in the U.S. are recorded in over half a century of the Oil & Gas Journal's annual reserves estimates. The total of past production plus "remaining recoverable reserves" keeps expanding.

Nehring's 1978 estimates¹³ purport to allow for this, but in 22 out of 39 cases the most optimistic estimate of field URR falls below what had already been produced and booked by 1996 (Fig. 4) [24111 bytes]. The mistake is not mere underestimation of URR. It is in the concept of URR as a fixed amount, rather than a dynamic variable. And the error shows up in the increased "remaining reserves."

Reserves as inventory

Proved reserves are inventory, paid for up-front. Table 1 [70065 bytes] shows how proved reserves have turned over. OPEC countries are special: They restrain development to keep up the price, and some may mark up reserves to justify higher quotas.

Non-OPEC countries started with 29 billion bbl. Since 1944, they have used up 433 billion bbl and at end-1995 had 231 billion bbl "remaining."

The reserve statistics outside North America are not high quality. Numbers often look odd; one suspects inertia, ignorance, or creative accounting. Comparing reserves in two countries is iffy; so is comparing reserves in the same place at two points in time. But over decades, in a large group, gross additions to reserves is accurate, because it is mostly cumulated output.

Returns to drilling

A third forecasting approach is to make econometric estimates of returns to drilling. Most such studies conclude that there are falling returns.^{14 15}

But outside the U.S. and Canada, published oil reserve numbers are not good enough for such close comparisons. Gas reserve numbers are weaker. Lumping oil with gas makes things still worse. And the most carefully articulated models (those of Canadian Energy Research Institute) have repeatedly been revised upward, showing a pessimistic bias at work.

Decreasing discoveries

There are some widely varying estimates of the peak discovery year, some going back quite far. A more cautious estimate is: in the 1970s.¹⁰

Yet somehow the annual additions of gross proved reserves have kept increasing. It does not add up because, first, "discovery" statistics are weak. As fields grow, nobody knows for a long time-if ever-what was found in a given year. Second, boundaries of "pools" and "fields" are often arbitrary. Third, if for all those years new-found fields were getting smaller, deeper, more faulted, etc., the oil industry would be forced to squeeze the old fields ever-harder. Development costs per barrel would be hit by a double whammy and would rise very steeply. But they have declined.

Ten years ago, when it was expected that depletion would drive up U.S. natural gas prices by 5%/year above inflation, we concluded that "adequate supplies are likely in the United States for the next decade, almost without regard to price.^"16 Similar analysis of more scanty data made us conclude that ample gas reserves to serve Europe could be created and produced at lower prices than ruled then.15

That was not wrong. In 1989 a survey of producing countries outside the U.S. showed how costs had kept decreasing from 1955 through 1985.¹⁷ A forecast of growing non-OPEC supply was the basis for expecting falling real oil prices in the l990s.¹⁸ The study was called "heretical"¹⁹ then. Now it does not look far-fetched.

Field size curves

In any area, field size curves show a few very large fields and a long decreasing tail of smaller ones.

A thinner or fatter tail makes a huge difference for the area under the curve. The only good bet is that the tail will get fatter over time and stretch farther to the right because it pays to invest in knowledge and in networks to prove up the smaller accumulations. Over time, it seems to cost less to find a given amount of oil in many small fields than to grope around for one big one.

Reserves types

For distinguishing among "probable," "possible," and "undiscovered" reserves, the best suggestion, we think, was made in 1969 by the geologist Lewis Weeks²⁰: Treat them as ordinal, or comparative. "A potential resource estimate is aimed, first and foremost, at efficiency in spending the exploration dollar; that is, it is an indicator of the area or areas in which that dollar may best be spent." He warned such a number was not comparable to shelf inventory: proved reserves. The warning is needed now.

Once we get past proved reserves, we are in a different ball game. Instead of estimating current inventory, we are forecasting the building of future inventories, mostly in undiscovered fields. But we don't know the future investment needed to create those reserves, nor future prices to induce them. The forecasters are doing economics without knowing it, and assuming future data.

Resources and reserves

There are huge amounts of hydrocarbons in the earth's crust, but the knowledge of how to find and develop them profitably grows in unpredictable ways.

Shale oil was promising a century ago and has gone nowhere. Ditto for coal liquefaction: even under apartheid in South Africa. Geopressured brine and frozen hydrates hold huge quantities of methane, but reserves are zero until and unless cost comes far down.

As late as 1952, the President's Materials Policy Commission thought offshore oil exploration might be useful in case of war.²¹ In time this "unconventional source" became "mature."

The North Sea and the Gulf of Mexico were considered over the hill. Only very recently it became plain how much lower is the investment needed to create a reserve barrel there. In the North Sea, credit must be shared among the thickening network of pipelines; subsea completions and other new methods; and more rational tax systems, all giving higher inducements to invest.

In the Gulf of Mexico, oilmen are beginning to "peer through" some of the salt sheets underground. This enlarges the hunting area. In new and old areas, new seismic methods let oilmen "see" many more deposits. With directional drilling, they can reach them. In deeper water depths, they move from the continental shelf onto the continental slope to see how much payoff there is on softer, less consolidated, more permeable sands. Some of the new Gulf of Mexico fields have well flow rates several times the published averages for Saudi Arabia.

Future additions to proved reserves depend on how important those cost changes turn out to be. Yet in mid-1992, gulf mobile rigs were 60% idle, and major oil companies "had practically abandoned the area."²²

Another unconventional source now coming in from the margin is in various types of heavy oils, bitumens, oil sands, etc.: huge amounts in place, a negligible fraction in proved reserves. But recently companies have announced projects to create proved reserves in some areas of Canada and Venezuela.

We do not know the potential of these latest but not last chapters in the story of technology opening new areas for investment. Reserves will continue to be created out of a resource whose limits we do not know and will never know.

Uncertainty vs. speculation

Proved reserves are estimated assets, inventory renewable by constant fresh investment. But probable-possible-undiscovered "reserves" are not assets. They are implicit forecasts of investment, therefore of future technology, which nobody knows.

A Petroconsultants estimate that there remains 1,100 billion bbl²³-or any other number-is the same old mistake. Because the concept of a fixed limit is wrong, the predicted famine always fails. It is time to give up the concept.

Mankind has been getting smarter a bit faster than it has been forced into deeper and farther-off deposits. No law says this must continue forever.

Near-term indicators are favorable. Stable or lower development costs per unit of new reserves-added also indicates stable discovery costs. However, as noted in this journal, less of this information is being published.²⁴

Another indicator is the value of oil and gas reserves in the ground. If it were getting more expensive to find and develop reserves, the value of barrels already owned would rise. But this has not happened either.²⁵

The oil industry has always been in a tug-of-war between depletion and knowledge. It takes endless effort and investment to renew and expand reserves. But resource limits are a phantom. One should not fear what is not there.

References

International Energy Agency, "The Peak Recedes Again," 1996.
Lynch, Michael C., "The analysis and forecasting of petroleum supply: sources of errors and bias," Energy Watchers VII, ICEED, 1996.
Hubbert, M. King, "Nuclear Energy and the Fossil Fuels," Drilling and Production Practice, 1956.
Masters, et al., "World Petroleum Assessment and Analysis," paper to the World Petroleum Congress, Stavanger, Norway, 1994.
Wattenberger, Robert A., "Oil Production Trends in the Former Soviet Union," Advances in the Economics of Energy and Resources, ed. by John R. Moroney, JAI Press, Vol. 8, 1994.
Masters, C.D., Root, D.H., and Attanasi, E.D., "World Oil and Gas Resources-Future Production Realities," Annual Reviews of Energy, Annual Reviews Inc., 1990.
Petroconsultants, quoted in OGJ, Oct. 20, 1986, p. 22.
Curcio, Edgardo, "Oil Supply Prospects in the 1990s from non-OPEC Producing Countries," Energy Supply in the 1990s and Beyond, proceedings of the 11th Annual International Conference, International Association for Energy Economics, Washington, D.C., 1989.
Esser, Robert, "World Oil Supply through the Year 2005," Cambridge Energy Research Associates, as summarized in Petroleum Intelligence Weekly, July 15, 1991.
International Energy Agency, "Global Offshore Oil Prospects to 2000," 1996.
Lynch, Michael C., "An Omitted Variable in OECD Oil Supply forecasting," in Energy Supply/Demand Balances: Options and Costs, Proceedings of the International Association for Energy Economics, 12th Annual North American Conference, Washington, D.C., 1990.
Bradley, Howard B., ed., "Petroleum Engineering Handbook," Society of Petroleum Engineers, 1987, Chapts. 36, 40, 41.
Nehring, Richard, "Giant Oil Fields and World Oil Resources," R-2284-CIA (Palo Alto, Calif.: Rand Corp.), June 1978.
Ivanhoe, L.F., "Free world discovery indexes, 1945-81," OGJ, Nov. 21, 1983.
Reinsch, Anthony E., Considine, Jennifer I., and McKay, Edward J., "Taxing the Difference: World Oil Market Projections, 1994-2009," Canadian Energy Research Institute, Calgary, 1994.
Adelman, M. A., and Lynch, Michael C., "Supply Aspects," in Final Renort of International Natural Gas Trade Project, MIT Energy Laboratory, Vol. 1 (of 3), 1985-86, p. 47.
Adelman, M.A., and Shahi, Manoj, "Oil development-operating cost estimates 1955-1985," Energy Economics, Vol. 11 (1985), pp. 193-208.
Lynch, Michael C., "Oil Prices to 2000: The Economics of the Oil Market," Economist Intelligence Unit, London, May 1989.
Petroleum Economist, September 1989.
Weeks, Lewis G., "Offshore Development and Resources," Journal of Petroleum Technology, April 1969, pp. 377-385.
President's Materials Policy Commission, "Resources For Freedom," 1952.
World Oil, Feb. 2, 1993, p. 87.
Letter to Economist, (London), Aug. 9, 1994, p. 10.
Beck, Robert J., "Info-sizing", OGJ, Dec. 2, 1996, p. 21.
Adelman, M.A., "The Genie out of the Bottle," MIT Press, 1995, Chapt. 2; Adelman, M.A., Watkins, G.C., "The Value of United States Oil and Gas Reserves," MIT-CEEPR Working Paper 96-004, forthcoming in Advances in the Economics of Energy and Resources, Vol. 9.

The Authors

M.A. Adelman is an economist and member of the Center for Energy & Environmental Policy Research at Massachusetts Institute of Technology. He is an MIT professor emiritus.

Michael C. Lynch is executive director, Working Group on Asian Energy and Security, at MIT's Center for International Studies. He holds combined SB-MS degrees in political science from MIT and since the mid-1970s has conducted a variety of studies related to international energy matters.