Peak oil demand

Jan. 2, 2018
The past couple of years have seen a shift from concerns that oil supply would soon peak to a recognition of the potential for oil demand to peak and decline, which would have a significant impact on the value of oil companies' equities, comparing it to the sharp downdraft in US coal company stocks. Unfortunately, there remain some serious questions and uncertainties that will be addressed in this article.
A threat to oil company valuations?

Mike Lynch, Strategic Energy & Economic Research, Mass.

The past couple of years have seen a shift from concerns that oil supply would soon peak to a recognition of the potential for oil demand to peak and decline, which would have a significant impact on the value of oil companies' equities, comparing it to the sharp downdraft in US coal company stocks. Unfortunately, there remain some serious questions and uncertainties that will be addressed here.

Discussion of the topic is not new. A decade ago, the Rocky Mountain Institute and its doyen Amory Lovins released "Winning the Oil Endgame," which argued that a combination of vastly improved automobile efficiency and cellulosic ethanol would result in oil demand peaking (coining the term "negabarrels"). Later, high oil prices suggested that demand would grow much slower in the future.

More recently, peak oil demand has become a popular topic, with groups ranging from the environmental advocates Carbon Tracker Initiative and analysts like Bloomberg New Energy Finance foreseeing the potential for a peak in oil demand in the near future—perhaps in a decade or less. The driving factors cited vary include:

• The potential for battery technology advances, allowing electric vehicles to make large inroads in vehicle markets;

• A breakthrough in biofuels technology;

• Autonomous on-call vehicles that might reduce vehicle ownership and miles driven;

• Climate change policies that would require that large amounts of fossil fuels be left in the ground.

Although they end up at peak oil demand, these threads all have different assumptions, causes, and probabilities, but peak oil demand will have different impacts on oil reserve values depending on the types and locations. Oil sands tend to be produced slowly, for instance, while conventional oil declines rapidly.

Similarly, the impact of technological advances on petroleum prices will depend on the nature of the advance. For example, better batteries could make renewable energy more competitive with natural gas turbines, especially by requiring less backup power. Climate change policies might primarily affect coal consumption, or even benefit natural gas producers, while efforts to reduce energy poverty could see displacement of noncommercial energy by petroleum.

Belief in a Peak

Many pundits have made a seamless transition from arguing about peak oil supply to predicting peak oil demand by simply admitting that the former was proved incorrect, especially by the shale oil revolution. The reality is more complex. Shale oil played a role in preventing supply from peaking, but only a subsidiary one: from 1998, when "The End of Cheap Oil" appeared in Scientific American, petroleum production has increased by nearly 20 mb/d, of which only 4 mb/d was shale oil.

Indeed, peak oil theory was always pathological science. The claim that any given country's oil production followed a Hubbert curve and could thus be reliably predicted is false: most nations' production does not resemble a Hubbert curve. Similarly, the argument that production in a field, nation, or region could be extrapolated to the endpoint once in decline was also false: a peak in any given area can be followed by increasing production, and sometimes a new peak.

Peak oil advocates' supply modeling and resource estimates were based on the mistaken presumption that geology determined production trends and changes in fiscal terms or political access to resource basins were irrelevant. This is why so many peak oil forecasts proved wrong, as predicted in Lynch (1995). But why then did the idea have such strong appeal?

The basic concepts—oil is finite, depletion raises costs—appeared sound enough that few examined the actual methods used to produce the forecasts behind peak oil. It is noteworthy that the idea had the greatest attraction amongst the environmental community, who trumpeted the basic argument without examining the fundamental concepts.

Real world peaking

But whereas there has never been a global peak in production due to scarcity for any non-renewable resource, a peak in demand is actually possible. Granted, three millennia after the end of the Bronze Age, we still use large amounts of copper (not so much for weapons anymore), but coal is rarely used to heat homes in developed nations despite its abundance.

On the other hand, there is no doubt but that many of those touting peak oil demand do so because it suits their particular ideology. It is amusing to see's Joseph Romm confidently argue for an early peak for oil demand, since a few short years ago he was similarly secure in the belief that peak oil supply was imminent, as witness his 2009 comments "…I have blogged endlessly on the painfully obvious reality that we are at or near the peak oil supply." Now, he admits, "The idea of peak oil supply…is dead."

This demonstrates once again how many pundits adopt a stance and cherry pick some data or citations to support their views without actually knowing much about the situation. Consumer (of punditry) beware.

The primary two arguments supporting the case of peak oil demand are that climate change policies and the technological revolution in transportation will act to suppress demand. There is an obvious overlap between the two.

Climate change

The Carbon Tracker Initiative has argued that, generally speaking, 80% of current fossil fuel reserves cannot be consumed between now and 2050 if the world is to meet the 2 degrees Celsius increase ceiling. Clearly, this is aspirational rather than analytical. The Kyoto Accords were greeted with great fanfare, yet only 50% of the targets have been met.

And the presumption restricting fossil fuel consumption will be the only policy is obviously flawed. Although most forms of carbon sequestration are currently uneconomic, in the future it might prove as viable as, say, electric vehicles. Similarly, geoengineering holds significant potential, with some approaches already appearing to be economically viable.

And the impact on oil demand appears to be exaggerated. Studies of necessary GHG reductions such as the IEA's see as little as 18% coming from the transportation sector, with 53% from power generation and buildings, where oil is not usually a major component of demand. And finally, since electric vehicles are a particularly expensive way of reducing GHG emissions, a reduction in support for electric vehicles is quite likely, rather as Spain, Germany, Japan and others have curtailed financial support for expensive renewable energy.

Technological change

Technology clearly appears to be advancing rapidly, particularly in the fields of electronics, and many activists others are embracing it as the solution to climate change. The main impact on oil demand would be from lithium-ion battery progress, but ride-sharing and biofuels appeal to some as potential game-changes that could affect oil demand.

For biofuels, it is simply enough to note that for 20 years, promises of imminent breakthroughs have not been fulfilled, and while some great advance could occur at any time, the odds suggest that there will be no major change in production costs in the next decade. Indeed, as a Newsweek article put it, "several major companies including Shell and ExxonMobil are seemingly abandoning their investments in this environmentally friendly fuel." Jatropha, seen a decade ago as a source of biodiesel, has been largely abandoned and U.S. cellulosic ethanol production, hailed as a savior by Richard Lugar James Woolsley two decades ago, remains at roughly 1 tb/d or 0.1% of ethanol supply.

Autonomous vehicles could progress rapidly into the market, but are more likely to increase vehicle-miles-traveled rather decrease them. Ride-sharing has great appeal for technophiles, but the evidence that it will change vehicle ownership, let alone the type of propulsion, is minimal. Groups like ReThinkX, who argue that private vehicle ownership will plummet in the next decade, sales falling to zero by 2024, appear to be wildly optimistic given that large cities like New York and Tokyo with expensive parking and insurance and dense mass transit and taxi systems still have high levels of car ownership, roughly one per two households, demonstrating there is a significant consumer demand for the convenience of car ownership.

While ride-sharing of autonomous battery electric vehicles will probably grow in major urban centers, the market penetration in the US as a whole seems unlikely to be significant for at least a decade. Further, the more ready availability of transport might mean an increase in miles traveled, rather than a decrease.

Battery electric vehicles (BEVs) get the most attention as a technology with the potential to suppress petroleum demand. Already, there are roughly two million BEVs in use and sales have been growing exponentially. As Table 1 shows, a number of groups have projected that BEVs will be competitive with conventional vehicles within the next decade, based on their projections of falling costs for lithium-ion batteries.

Of course, there have been many previous instances where electric vehicles were touted as on the verge of acceptance among consumers, only to prove false. In the 1990s and 2000s, many hailed the progress in both electric and fuel cell vehicles, predicting early market penetration, only to see them completely fail to win consumer acceptance.

Which doesn't mean that electric vehicles will never be ready for market. Battery costs have come down and performance has improved, but with the thin market until recently, estimates of cost trends are hardly reliable. To date, most sales appear to be spurred by hefty government incentives.

However, performance is also a major factor in consumer choice. Engine performance is as good as or better than ICE vehicles, but inconvenience remains a barrier. Range anxiety is a very real thing, in part because drivers are advised against charging the battery more than 80% or discharging it below 20% on a regular basis, meaning an estimated 200-mile range is actually more like 140 miles. In cold weather, the range drops further, at least 5%.

And although many BEV advocates deny it, recharging remains significantly more inconvenient than refueling an ICE vehicle. Use of a Tesla Supercharger cuts the time from as much as eight hours to as little as 30 minutes, although other manufacturers' vehicles can't use them. Still, the difference between five minutes for 300 miles of range in an ICE and 30 for "up to" 170 is significant.

Consumers are being asked to spend substantially more money for a product that delivers much poorer services. The evidence of consumer preference is particularly clear where subsidies have been reduced or ended. In Georgia, the state withdrew a $5,000 per vehicle subsidy and sales dropped by 80%. It would seem that, as with renewable energy, governments are beginning to regret the benefits and subsidies offered electric vehicle buyers, especially as sales have risen and the payouts become more substantial. Enthusiasm for BEVs looks like irrational exuberance.

Peak demand and equity prices

The most extreme case for lower oil company equity prices comes from Carbon Tracker Initiative which states that "…up to 80% of declared reserves owned by the world's largest listed coal, oil and gas companies and their investors would be subject to impairment as these assets become stranded." Further, they believe that "greater than 50% of the value of an oil and gas company resides in the value of cash flows to be generated in year 11 onwards." The implication is that oil company equities should be discounted by 25% or more.

This does not appear to conform to the general understanding of discounted cash flow in the oil industry. With the usual industry practice of a discount rate of 10%, about 65% of a constant cash flow would be received in the first ten years. And the decline rates of reserves translate into an additional discount factor: if 10% of proved reserves are produced in a given year, the depletion or natural decline rate is 10%, which means that production and cash flow will decline by 10% per year (assuming constant prices) and the discounted cash flow would decline by 20% per year. In that case, about 90% of the value of the reserves would be received in the first ten years.

Table 2 shows recent production rates for a number of countries and locales, which highlights the importance of differentiating types of production. The clear lesson is that synfuels such as tar sands have very low decline rates, while offshore reserves tend to be highest. (Shale oil generally has very even higher decline rates, but company data will show a mix of new, fast-declining well and older wells with more stable production.) Table 3 indicates how much of the total discounted cash flow will be produced within 10 years, given different decline rates.

So, only for oil sands production and some conventional onshore oil should there be any significant discounting of reserve values due to expectations of weak oil demand, and even that is problematical. In most cases, the majority, if not the bulk of the value, of proved, developed reserves is recovered within a decade and events beyond that period are all but irrelevant for current reserve and equity values.

Prices and demand

And of course, economics could play a role in reducing oil demand. A reduction in subsidies in oil producing countries could reduce oil demand slightly, but moves in that direction have been slow, while large consumers like China and India have only modest subsidies. So even a complete termination of product price subsidies would only moderate world demand trends slightly.

A bigger effect would be from rising crude oil prices, as many predict (Figure 1). Unfortunately, Lynch (1992) showed that long-term oil price forecasts have tended to be overly optimistic or bullish for many years, due to a belief that depletion must send costs rising over time. But this was countered by Adelman (1986): "Diminishing returns are opposed by increasing knowledge, both of the earth's crust and of methods of extraction and use. The price of oil, like that of any mineral, is the uncertain fluctuating result of the conflict." In both the 1970s and 2000s, political disruptions of supply that sent prices soaring were misinterpreted as "new paradigms" which set new, higher floors on prices, and that appears to be the case now.

Forecasted price of $100/barrel and more are not impossible to reach, but would primarily require major producers to hold their oil off the market for long periods. Climate change policies or expanded electric vehicle sales would tend to depress the price, not raise it, which have a positive effect on demand.

Impact of peak oil demand on price

Bloomberg New Energy Finance's claim that a loss of 2 mb/d of demand because of rising electric vehicle market shares would cause a price collapse similar to 2014 is simplistic. From 1980 to 1985, oil demand dropped by only 2 mb/d, but was about 20 mb/d below trend. Given projections that future oil demand will grow by 1-1.5 mb/d per year, a reduction (as BNEF suggests) of 300 tb/d per year is significant, but remains only one part of the supply/demand equation. Saudi Arabia, now producing 10 mb/d, could easily accommodate a 2 mb/d cut, and US shale oil production growth has dropped from 1 mb/d per year to 0.3-0.5 mb/d per year, more than offsetting the implied BEV effect. Similarly, Iraqi production could potentially grow by 0.3 mb/d per year, or be flat for an extended period.

Thus, an aggressive electric vehicle adoption rate could put pressure on oil prices, but there are much greater threats to price levels.

The potential for wasted investment

Carbon Tracker Initiative has argued that meeting Paris Climate Accords targets means that 43% of capital expenditures in oil between now and 2025 are unneeded. As argued above, it is questionable that there would be much impact from peak oil demand on current assets, but could the threat to future investment be more serious?

Interestingly, those who argue that long-term investment programs should be guided by current market assessments are also quick to point out how bad projection of renewable investments have been. More tellingly, the oil industry invested billions in the early 1980s in response to "expert" predictions of future market tightness and rising prices (after they had just tripled), with both the Carter and Trudeau Administrations insisting that markets were too myopic to recognize the need for synfuels and Arctic oil resources, respectively.

Which brings back the question of decline rates: if the world loses 4-5 mb/d of production capacity every year, how hard would it be to respond to demand that is weaker by 10% of that amount due to the adoption of BEVs? The raw numbers imply that this is not a challenge, but of course, in the real world, different behavior by OPEC NOCs, independents, and majors could make this much more complex in practice.


Expectations that world oil demand will peak within the decade are based primarily on assumptions about electric vehicle battery technology and consumer choice that do not appear well founded. The odds that fossil fuel use will come screeching to a halt likewise seems extremely optimistic. But the simple fact is that the rapid decline in discounted cash flow from current reserves implies that the impact on equity values for petroleum producers will be quite small.

That said, long-lived assets such as oil sands and stranded natural gas fields are more at threat than shale oil or offshore reserves, whose production tends to decline the fastest. Those with aggressive expectations for BEV or climate change policies should adapt their investment strategy accordingly.
About the author

Michael Lynch graduated from M.I.T. and has researched energy economics for four decades, with publications in six languages. He was a researcher at M.I.T. and chief energy economist for DRI-WEFA, and has served as president of the US Association for Energy Economics. He currently is president of Strategic Energy & Economic Research, blogs for, and recently published The Peak Oil Scare and the Coming Oil Flood (Praeger).

T1: Predictions of Parity between BEVs and ICEs

Battery price



















































T2: 2016 decline rates
























T3: Share of production and discounted cash flow in 1st 10 years (out of 35 years total)

Oil Sands




Depletion rate





Produced 1st 10 years





DCF 1st 10 years