Friday, November 22, 2013

Five Myths About the "Carbon Asset Bubble"

  • The idea that efforts to mitigate climate change expose fossil fuel assets to the risk of a bubble-like collapse has attracted some high-profile supporters.
  • However, the notion of a "carbon bubble" depends on questionable assumptions concerning our current knowledge of climate change, the rate of adoption of renewable energy technology, and how such assets are valued.
In their recent Wall St. Journal op-ed, Al Gore and one of his business partners characterized the current market for investments in oil, gas and coal as an asset bubble. They also offered investors some advice for quantifying and managing the risks associated with such a bubble. This is a timely topic, because I have been seeing references to this concept with increasing frequency in venues such as the Financial Times, as well as in the growing literature around sustainability investing.

Although bubbles are best seen in retrospect, investors should always be alert to the potential, particularly after our experience just a few years ago. In this case, however, I see good reasons to believe that the case for a “carbon asset bubble” has been overstated and applied too broadly. The following five myths represent particular vulnerabilities for this notion:

1. The Quantity of Carbon That Can Be Burned Is Known Precisely
Mr. Gore is careful to differentiate uncertainties from risks, which he distinguishes for their amenability to quantification. For quantifying the climate risk to carbon-heavy assets, he refers to the widely cited 2°C threshold for irreversible damage from climate change, and to the resulting “carbon budget” determined by the International Energy Agency (IEA). As Mr. Gore interprets it, “at least two-thirds of fossil fuel reserves will not be monetized if we are to stay below 2° of warming.” That would have serious consequences for investors in oil, gas and coal.

The IEA’s calculation of a carbon budget depends on a factor called “climate sensitivity.” This figure estimates the total temperature change resulting from a doubling of atmospheric CO2 concentrations. The discussion of climate sensitivity in the recently released Fifth Assessment Review of the Intergovernmental Panel on Climate Change (IPCC) sheds more light on this parameter, which turns out not to be known with certainty. Their Summary for Policymakers includes an expanded range of climate sensitivity estimates, compared to the IPCC’s 2007 assessment, of 1.5°-4.5°C with a likelihood defined as 66-100% probability. It also states, “No best estimate for equilibrium climate sensitivity can now be given because of a lack of agreement on values across assessed lines of evidence and studies.”

The draft technical report that forms the basis for the Summary for Policy Makers provides more detail on this. It further assesses a probability of 1% or less that the climate sensitivity could be less than 1°C. That shouldn’t be surprising, since temperatures have already apparently risen by 0.8°C above pre-industrial levels. At the same time, the report indicates that recent observations of the climate — as distinct from the output of complex climate models — are consistent with “the lower part of the likely range.”

In other words, while continued increases in atmospheric CO2 resulting from increasing emissions are widely expected to result in warmer temperatures in the future, the extent of the warming from a given increase in CO2 can’t be determined precisely before the fact. For now, at least, the CO2 level necessary to reach a 2°C increase would be consistent with calculated carbon budgets both larger and smaller than the IEA’s estimate. That means that the basis of Mr. Gore’s suggested “material-risk factor” — as distinct from an uncertainty — is itself uncertain.

2. The Transition to Low-Carbon Energy Is Occurring Fast Enough to Threaten Today’s Investments in Fossil Fuels
There is no doubt that renewable energy sources such as wind and solar power are growing at impressive rates. From 2010 though 2012 global solar installations grew by an average of 58% per year, while wind installations increased by 20% per year. Yet it’s also true that they make up a small fraction of today’s energy production, and that the risks for investors of extrapolating high growth rates indefinitely proved to be very significant in the past.

For further clarity on this, consider the IEA’s latest World Energy Outlook, the agency’s analysis of global energy trends, which was just released on November 12. The IEA projects global energy consumption to grow by 33% from 2011 to 2035 in its primary scenario, which reflects expanded environmental policies and incentives over those now in place. In that scenario, the global market share of fossil fuels is expected to fall from 82% to 76%, but with total fossil fuel consumption still growing by 24% over the period. Only in their “450″ scenario, based on similar assumptions to its carbon budget, would fossil fuel consumption fall by 2035, and then only by 11%.

Moreover, in its April 2013 report on “Tracking Clean Energy Progress,” the IEA warned, “The drive to clean up the world’s energy system has stalled.” This concern was based on their observation that from 1990 to 2010 the average carbon dioxide emitted to provide a given unit of energy in the global economy had “barely moved.” That’s hardly a finding to be celebrated, but it serves as an important reminder that while some renewable energy sources are growing rapidly, fossil fuel consumption is also growing, especially in the developing world — and from a much larger base.

The transition to lower-carbon energy sources is inevitable. However, it will take longer than many suppose, and it cannot be accomplished effectively with the technologies available today. That’s a view shared by observers with better environmental credentials than mine.

3. All Fossil Fuels Are Equally Vulnerable to a Bubble
As Mr. Gore correctly notes, “Not all carbon-intensive assets are created equal.” Unfortunately, that’s a distinction that some other supporters of the carbon asset bubble meme don’t seem to make, particularly with regard to oil and natural gas. The vulnerability of an investment in fossil fuel reserves or hardware to competition from renewable energy and decarbonization doesn’t just depend on the carbon intensity of the fuel type — its emissions per equivalent barrel or BTU — but also on its functions and unique attributes.

The best example of this might be a recent transaction involving the sale of a leading coal company’s mines. What’s behind this wasn’t just new EPA regulations making it much harder to build new coal-fired power plants in the US, but some fundamental, structural challenges facing coal. Power generation now accounts for 93% of US coal consumption, as non-power commercial and industrial demand has declined. This leaves coal producers increasingly reliant on a utility market that has many other--and cleaner--options for generating electricity. That’s particularly true as the production of natural gas, with lower lifecycle greenhouse gas emissions per Megawatt-hour of generation, ramps up, both domestically and globally. Coal accounts for about half of the global fossil fuel reserves that Mr. Gore and others presume to be caught up in an asset bubble.

Compare that to oil, which at 29% of global fossil fuel reserves, adjusted for energy content, still has no full-scale, mass-market alternative in its primary market of transportation energy. Despite a decade-long expansion, biofuels account for just over 3% of US liquid fuels consumption, on an energy-equivalent basis. They’re also encountering significant logistical challenges and concerns about the degree to which their production competes with food. This has contributed to efforts in the EU to limit the share of crop-based biofuels to around 6% of transportation energy. Biofuels have additional potential to displace petroleum use, particularly as technologies for converting cellulosic biomass become commercial, but barring a prompt technology breakthrough they appear incapable of substituting for more than a fraction of global oil demand in the next two decades.

Electric vehicles offer more oil-substitution potential in the long run, though they are growing from an even smaller base than wind and solar energy. Their growth will also impose new burdens on the power grid and expand the challenge of displacing the highest-emitting electricity generation with low-carbon sources.

Meanwhile, natural gas, at 20% of global fossil fuel reserves, offers the largest-scale, economic-without-subsidies substitute for either coal or oil. In any case, it has the lowest priority for substitution by renewables on an emissions basis, and so should be least susceptible to a notional carbon bubble.

4. A Large Change in Future Fossil Fuel Demand Would Have a Large Impact on Share Prices
Although Mr. Gore’s article includes a good deal of investor-savvy terminology, it is entirely lacking in two of the most important factors in the valuation of any company engaged in discovering and producing hydrocarbons: discounted cash flow (DCF) and production decline rates. Unlike tech companies such as Facebook or even Tesla, the primary investor value proposition for which depends on rapid growth and far-future profitability, most oil and gas companies are typically valued based on risked DCF models in which near-term production and profits count much more than distant ones.

At a conservative discount rate of 5%, the unrisked cash flow from ten years hence counts only 61% as much as next year’s, while cash flow 20 years hence counts only 38% as much. Announced changes in near-term cash flow due to unexpected fluctuations in production or margins would normally be expected to have a much bigger impact on share prices than an uncertain change in demand a decade or more in the future.

This is compounded by the decline curves typical of many large hydrocarbon projects. If the first 3-5 years of a project account for more than half its undiscounted cash flows, it won’t be very sensitive to long-term uncertainties, nor would a company made up of the aggregation of many projects with this characteristic. This is even truer of shale gas and tight oil projects, which yield faster returns and decline more rapidly.

I can’t speak for Wall Street's oil and gas analysts, but I’d be surprised based on past experience in the industry if the risk of a 10% or greater drop in global demand for oil or gas in the 2030s would have much of an effect on their price targets for companies — certainly not enough to qualify as a bubble.

5. Fossil Fuel Share Prices Don’t Already Account for Climate Risks
The assertion of a carbon bubble in fossil fuel assets ultimately depends on investor ignorance of climate-response risks, presumably because companies haven’t quantified those risks for them. To the extent the latter condition is true, it represents an opportunity for companies seeking to capitalize on the boom in sustainability-based investing.

However, you needn’t be an adherent of the Efficient Markets Hypothesis for which Eugene Fama was named as a recipient of this year’s Nobel Prize in Economics to realize that thanks to the Internet, average investors have access to most of the same information on this subject as Mr. Gore and his partners. Institutional investors, who make up the bulk of the shareholding for at least the larger energy firms, and the analysts who follow these companies have the resources to access even more information.

Nor is the idea of a carbon bubble exactly new. Mr. Gore didn't create it, and I’ve been following it for a couple of years, as it took over from waning interest in Peak Oil. It’s not an obscure risk, either, in the sense that sub-prime mortgages and credit default swaps were in the lead-up to the failure of Lehman Brothers in 2008. It’s becoming more mainstream every day, although the burden of proof that this risk is mispriced rests with those advocating this view.

Before concluding, a word of disclosure is in order. As you may gather from my bio, I spent many years working with and around fossil fuels, though my ongoing involvement in energy is much broader than that. As a result of that experience, my portfolio includes investments in companies with significant fossil fuel holdings. I strive for objectivity, but I can’t claim to be disinterested. However, neither can Mr. Gore. As a major investor in renewable energy and other technologies through the firm cited in the article and other roles, he has as much at stake in promoting the idea of a carbon bubble — and on a very different scale — as I might have in dispelling it.

The carbon bubble is an interesting hypothesis, even if I don’t yet find the arguments made in support of it convincing. Despite that, I see nothing wrong with investors wanting to track their carbon exposure, consider shadow carbon prices, or ensure they are properly diversified. However, the biggest risk I see that might eventually warrant considering divestment of fossil-fuel-related assets isn’t based on the merits of this analysis, but on the possibility of creating a self-fulfilling prophesy by means of drumming up social pressure on institutional investors. You might very well think that applies to this Wall St. Journal op-ed. I couldn’t possibly comment.

A different version of this posting was previously published on Energy Trends Insider.


Ed Reid said...

We know that CO2 is a greenhouse gas. We know that atmospheric CO2 concentrations are increasing, at least in part as the result of anthropogenic emissions. We know that increased atmospheric CO2 concentrations will tend to lead to increases in global average surface temperatures. We know that the effect of increasing CO2 concentrations is logarithmic, with each increment of CO2 having progressively less effect. We also know that increased CO2 concentrations increase the growth rates of most plants. We also know that current atmospheric CO2 concentration is below the optimal concentration for enhanced plant growth. We do not know the ideal atmospheric CO2 concentration. Therefore, we do not know whether the current increase in atmospheric CO2 concentration is approaching the optimum or moving away from it.

We do not know the current global average surface temperature, though we have an estimate. We do not know the ideal global average surface temperature. We do not know the current contribution of increased atmospheric CO2 concentrations to global average surface temperature. We do not know the current contribution of land use changes to global average surface temperature.

We know that the reported global average surface temperature anomalies are based on adjusted temperature records rather than actual temperature data. We know that the global average surface temperature anomalies reported by the principal reporting groups are different. We know that the reported global average surface temperature anomalies have not changed significantly in approximately 17 years. We know that the output scenarios of virtually all of the currently available climate models did not predict this hiatus; and, create global average surface temperature anomaly scenarios which are significantly higher than the adjusted measured anomalies, no less the anomalies in the underlying data. We also know that the outputs of the various climate models are different, so we know that at least all but one of them are in error; and, that perhaps all of them are in error.

The IPCC has recently proclaimed (with 95% confidence) that more than half of the increase in the global annual surface temperature anomalies is the result of "human activity". A recent draft paper, as yet unpublished,( that, at least in the US, approximately half of the increase in the reported temperature anomaly is the result of adjustments to the underlying data. This suggests that one specific "human activity" (temperature adjustments) might be responsibility for the bulk of the increase attributed to "human activity".

To paraphrase former US SecDef Donald Rumsfeld: There are known knowns, known unknowns and unknown unknowns. Climate science is well endowed with unknowns. It would be prudent to minimize the unknowns before spending $ trillions to revise the global energy economy.

It would also be wise to have demonstrated suitable, economical substitutes for fossil energy sources before taxing, capping or otherwise restricting the use of fossil energy. We have clearly not yet achieved that status.

“You’ve got to be careful, if you don’t know where you are going, because you might end up someplace else.”. Yogi Berra, American philosopher

“Don’t begin vast programs with half-vast ideas.”, Ed Reid, crotchety old curmudgeon

Geoffrey Styles said...

Thanks for your long and measured comment. You've identified some of the key uncertainties in the climate projections upon which so much policy is based. In particular, I'm not sure we know the exact absolute temperature of the earth, either--or that that term has any practical meaning. However, to the extent that the changes we are dealing with represent actual observations from a consistent set of instruments, the trend can be significant even if we don't know the absolute level from which it began.

Unfortunately, that notion has been undermined by relentless fiddling with the data by the folks who are charged with gathering and safeguarding it. NASA's GISS dataset has exhibited major revisions in past temperatures that didn't involve someone traveling back to the 1930s with a time machine and a thermocouple. The output of models depend on enough assumptions as it is, without the "data" input to them depending on still further assumptions.

However, I have to question your use of the word "ideal". I don't see anyone arguing that the pre-industrial level was ideal, either for nature or humanity. The biggest sources of worry seem to be a rate of change that could outstrip the ability of species and ecosystems (and humanity?) to adapt to, and the potential for large-magnitude changes that would clearly be much less ideal for supporting 6+ billion humans, a large fraction of whom live near current seacoasts. A little warmer might be more comfortable than a little colder, but a lot warmer would clearly be worse for humans and the natural systems we depend on.

Ed Reid said...


Depending on which global average surface temperature anomaly you analyze, the reported anomaly is ~0.8C over a period of ~150 years. The beginning of that period roughly coincides with the end of the Little Ice Age; and, with the exception of the Central England Temperature Record (CETR), the beginning of the instrumental temperature record. It is generally, though not universally, agreed that the Medieval Warm Period was as warm as, or warmer than, the current period.

We cannot know whether the current global average surface temperature, whatever it is, is ideal, cooler than ideal or warmer than ideal for animals and/or plants. We do know that the current atmospheric CO2 concentration is less than half of the ideal concentration for plant growth, though we do not know whether it is ideal, lower than ideal, or higher than ideal for humans and animals.

At present, the rate of change of the global average surface temperature anomaly appears to be zero; or, at least, to have approached zero asymptotically. This hiatus is occurring in the face of rapid increases in atmospheric CO2 concentration. The gap between the climate model scenarios and the "adjusted" temperature anomaly is growing. Recent research suggests that the divergence between the model scenarios and the "adjusted" temperature anomalies is the result of unrealistic sensitivity inputs to the models. Other recent research suggests that we have begun a protracted period of global cooling.

Geoffrey Styles said...

It's ironic that the credibility of long-range climate projections hinges on recent trends and what happens in the next few years. Skeptics point to a hiatus in warming since the late 1990s that is apparent even in the NASA GISS dataset, which records that the 2012 temperature "anomaly" of 0.6°C above the 1951-80 average is essentially unchanged since 1998, and 2013 ytd is on the same track.


The IPCC addressed this in its just-published Fifth Assessment Review by indicating that the period in question is too short from which to draw any conclusions, due to natural variability and the sensitivity to beginning and end dates. Even to someone who is convinced of the overall warming trend, this looks like hand-waiving.

Only the passage of time can resolve this discrepancy. If we're still talking about a hiatus 10 years from now, then the IPCC will have to go back to the drawing board. If temps start to rise again and we're consistently 0.2°C or so warmer than today, then the longer-term predictions will start to look more credible again.

Ed Reid said...


We are talking about a long term anomaly of less than 1oC, based on a temperature record collected from a global surface temperature measuring network. The US component of that network consists of measurement installations 90+% of which are subject to measurement errors of >/= 1oC. (

The reporting of anomalies to two decimal places is based on the presumption that, while the temperature measurements are of questionable accuracy, the sensor outputs are stable enough over long time periods to justify that apparent precision. However, we know that the instrumental temperature readings are being "adjusted", so that the reported anomalies are not based on actual data, but rather on "adjusted" temperatures. (

That raises the question of how large the anomaly must be before we can be reasonably sure that it is real.