Sunday, September 07, 2014

The more uncertain we are, the more careful we should be

It is a staple of apologists for the chemical and fossil fuel industries to say, "We have no proof that what you are talking about is dangerous." Let me restate that in probabilistic terms: "We are highly uncertain about the harm of what you are talking about."

When stated in probabilistic terms, uncertainty about harm becomes much more alarming. Nassim Nicholas Taleb has added to a working paper which I discussed last week entitled "The Precautionary Principle: Fragility and Black Swans from Policy Actions." As I suggested in last week's piece, climate change is an obvious candidate for the precautionary principle because climate change involves the risk of systemic ruin.

In his addendum Taleb explains that climate change deniers who criticize climate models for their uncertainty don't have the slightest clue what that implies. Rather than suggesting that we should ignore such models, the uncertainty suggests that we should be even more diligent about mitigating climate change since the high uncertainty means, probabilisticly speaking, that we have larger exposure to catastrophic outcomes.

Statistically, this is explained as an increase in the scale of the distribution which leads to an increase in the size of the tails associated with the probability curve. It means that the system we are dealing with is MORE fragile and thus more subject to catastrophic outcomes. Tails represent rare, but highly impactful events and in this case, a ruinous result. If rare becomes a lot less rare (fat tails), then the risk of ruin is greatly increased.

Let's look at other cases where the risks are likely to be greater than widely assumed. U.S. Department of Energy forecasts for energy supplies, particularly oil and natural gas, are treated as gospel by companies and governments across the globe. But the caveats that the Energy Department includes show that such forecasts are highly fragile, indeed, the entire oil and natural gas supply system is fragile in that it may not deliver what we want it to deliver. Here is a sample from an Energy Department discussion of forecast model uncertainty in reference to the presumed renaissance in U.S. oil and natural gas resulting from the exploitation of deep shale resources:

Estimates of technically recoverable tight/shale crude oil and natural gas resources are particularly uncertain and change over time as new information is gained through drilling, production, and technology experimentation. Over the last decade, as more tight/shale formations have gone into production, the estimate of technically recoverable tight oil and shale gas resources has increased. However, these increases in technically recoverable resources embody many assumptions that might not prove to be true over the long term and over the entire tight/shale formation. For example, these resource estimates assume that crude oil and natural gas production rates achieved in a limited portion of the formation are representative of the entire formation, even though neighboring well production rates can vary by as much as a factor of three within the same play. Moreover, the tight/shale formation can vary significantly across the petroleum basin with respect to depth, thickness, porosity, carbon content, pore pressure, clay content, thermal maturity, and water content. Additionally, technological improvements and innovations may allow development of crude oil and natural gas resources that have not been identified yet, and thus are not included in the Reference case.

It turns out that a forecast that many people assume is all-but-certain is admittedly quite shaky according to its creators. The forecast is fragile and subject to catastrophic failure with the possibility that actual production will be significantly below the forecast in the next few decades. A similar failure in current optimistic worldwide oil and natural gas forecasts would carry grave consequences if we were to make no preparations for a surprise on the downside.

But precisely because so many people believe these optimistic oil and natural gas forecasts to be facts rather than speculation, THEY MAKE NO PREPARATIONS FOR AN ALTERNATE AND POSSIBLY DISASTROUS OUTCOME! And, that is the problem with forecasts that are widely accepted and used for planning and policy purposes.

Keep in mind that a forecast is nothing but a model of something over time in the future. It isn't and cannot be based on actual knowledge of the future. Most models, particularly financial and resource models, really only extrapolate the past into the future which is actually a naive approach.

As for genetically modified crops, we are told that there is no evidence of harm from ingesting these crops. Long-term animal feeding studies have been made all-but-impossible by the companies that own the patents to the seeds. So, the lack of evidence is partly intentional. But LACK OF EVIDENCE IS NOT THE SAME AS LACK OF RISK. We did not have any evidence that the drug thalidomide, used to treat morning sickness in pregnant women, would cause deformities in human fetuses. But that didn't mean there was no risk. Wikipedia notes: "At the time of the drug's development, scientists did not believe any drug taken by a pregnant woman could pass across the placental barrier and harm the developing foetus." Why investigate risks to the fetus when you already believe there are none! Likewise, if you've already decided that genetically modified foods pose no greater risk than traditional foods (as the U.S. Food and Drug Administration has), you will not investigate the risks.

Finally, we see a don't-worry-be-happy attitude about the worldwide electrical grid. We have known for many years about the threat of an electromagnetic pulse or EMP. Such a pulse, if widespread enough, could bring down the electrical grid worldwide. Because so many processes in the modern world, especially information technology, must have continuous inputs of electricity, the result would indeed be a wipeout for modern civilization since there is currently no capability to repair such widespread damage. (For more on this, see my piece "Solar storms, EMP and the future of the grid.")

An EMP can be produced by a high-altitude explosion of a nuclear weapon. This is why much military hardware has been shielded from EMP. But it is the natural source of EMPs, the Sun, that should concern us more. If a solar storm similar to the Carrington Event, which hit Earth in 1859, were to occur today, we would very likely face ruin.

What we don't know is the frequency of such high-intensity storms hitting the Earth. And, it is our very ignorance which subjects us to heightened risk. Not knowing whether we might have thousands of years or just a few to prepare for such an event creates that risk. Once again, lack of evidence is not the same as lack of risk.

Our assumptions may be wrong. Our observations or data faulty. Our information woefully incomplete. Knowing this we would be wise to put large margins of safety into the systems we build and the practices we initiate. We would be wise to forego new practices that clearly run the risk of widespread and systemic ruin. Instead, we all too often close our eyes to risk; and, all too often the reason is immediate profit for a few. Meanwhile, the rest of us suffer the consequences.

Kurt Cobb is an author, speaker, and columnist focusing on energy and the environment. He is a regular contributor to the Energy Voices section of The Christian Science Monitor and author of the peak-oil-themed novel Prelude. In addition, he has written columns for the Paris-based science news site Scitizen, and his work has been featured on Energy Bulletin (now, The Oil Drum,, Econ Matters, Peak Oil Review, 321energy, Common Dreams, Le Monde Diplomatique and many other sites. He maintains a blog called Resource Insights and can be contacted at


zarathrustra said...

The more desperate the system becomes to maintain itself the more it will gamble on increasingly risky strategies;Not unlike a Drug addict moving from gambling to mugging to breaking and entering to murder.
The psychosis of the Entropic Death Machine.

John Clark said...

It seems counterintuitive to assume uncertainty should lead to caution. Quickly this arrives at any uncertainty leading to complete inaction. And the world is not deterministic.

Analysis paralysis being another term for the same concept.

Yes, resource estimates are uncertain. This is irrelevant in terms of development. The oil and gas industry certainly understands that every well drilling "investment" might lead to ZERO return on that investment. This is not a surprise.

But it doesn't stop development, because it is the histogram of results that matter, not a single result. So the uncertainty is COUNTED ON, and KNOWN, in developing these types of resources.

Scientists even estimate them in this way, to quantify the range of uncertainty.

But this uncertainty does not, in and of itself, have ANYTHING to do with being careful, not being careful, or anything resembling the type of caution you imply.

Kurt Cobb said...

John Clark seems to miss my point entirely. It is not the average result which concerns me. It is the extreme result that is hidden from us in the probablistic structure of the universe. I'm not talking about prospecting for resources, I'm talking about building a society around resources which are finite,the extent of which might be much less or even possibly much more than we think. The truth is hidden from us. If we destroy our civilization because we guess wrong, this is not an acceptable risk in my view. See my previous piece called "Ruin is forever." If systemic ruin is a probability, even a small one, for any practice, then as you continue that practice in the long run the probability of ruin approaches one. It's just basic statistics.