by Judith Curry
So, is climate change a local or global threat? Are we risking global ruin?
How to frame a response to the climate change problem requires deeper examination of the actual nature of the risk/threat. A new paper by Nassim Taleb provides some food for thought.
The Precautionary Principle (with Application to Genetic Modification of Organisms)
Nassim Taleb, Rupert Read, Raphael Douady, Joseph Norman,Yaneer Bar-Yam
Abstract (excerpt). The precautionary principle (PP) states that if an action or policy has a suspected risk of causing severe harm to the public domain (affecting general health or the environment globally), the action should not be taken in the absence of scientific near-certainty about its safety. Under these conditions, the burden of proof about absence of harm falls on those proposing an action, not those opposing it. PP is intended to deal with uncertainty and risk in cases where the absence of evidence and the incompleteness of scientific knowledge carries profound implications and in the presence of risks of “black swans”, unforeseen and unforeseable events of extreme consequence.
Introduction: We believe that the PP should be evoked only in extreme situations: when the potential harm is systemic (rather than localized) and the consequences can involve total irreversible ruin, such as the extinction of human beings or all life on the planet. The aim of this paper is to place the concept of precaution within a formal statistical and risk-analysis structure, grounding it in probability theory and the properties of complex systems. Our aim is to allow decision makers to discern which circumstances require the use of the PP and in which cases evoking the PP is inappropriate.
The complete manuscript can be downloaded [here].
This paper was quite controversial in the twit-o-sphere, owing to its conclusions about GMOs. Here I focus on the general framework of Taleb’s argument, and how it might be used in the climate change debate.
Here are some excerpts:
Decision making and types of risk
Traditional decision-making strategies focus on the case where harm is localized and risk is easy to calculate from past data. Under these circumstances, cost-benefit analyses and mitigation techniques are appropriate. The potential harm from miscalculation is bounded.
On the other hand, the possibility of irreversible and widespread damage raises different questions about the nature of decision making and what risks can be reasonably taken. This is the domain of the PP.
The PP is intended to make decisions that ensure survival when statistical evidence is limited—because it has not had time to show up —by focusing on the adverse effects of “absence of evidence.”
The purpose of the PP is to avoid a certain class of what, in probability and insurance, is called “ruin” problems. A ruin problem is one where outcomes of risks have a non-zero probability of resulting in unrecoverable losses. The large majority of variations that occur within a system, even drastic ones, fundamentally differ from ruin problems: a system that achieves ruin cannot recover. As long as the instance is bounded, there may be some hope of reversing the misfortune. This is not the case when it is global.
Precautionary considerations are relevant much more broadly than to ruin problems. For example, there was a precautionary case against cigarettes long before there was an open-and-shut evidence-based case against them. Our point is that the PP is a decisive consideration for ruin problems, while in a broader context precaution is not decisive and can be balanced against other considerations.
PP is not risk management
We can identify three layers associated with strategies for dealing with uncertainty and risk.
- The first layer is the PP which addresses cases that involve potential global harm, whether probabilities are uncertain or known and whether they are large or small.
- The second is risk management which addresses the case of known probabilities of well-defined, bounded gains and losses.
- The third is risk aversion or risk-seeking behavior, which reflects quite generally the role of personal preferences for individual risks when uncertainty is present.
JC comment: global harm does not necessarily imply ‘ruin’ (see below). Risk aversion seems the best characterization for proposed solutions to the climate change problem.
Ruin is forever
A way to formalize the ruin problem in terms of the destructive consequences of actions identifies harm as not about the amount of destruction, but rather a measure of the integrated level of destruction over the time it persists.Our focus here is on the case where destruction is complete for a system or an irreplaceable aspect of a system, a point that does not allow recovery.
JC comment: Should global warming be considered as a ‘ruin’ problem? Paul Ehrlich argues ‘yes’ [link]; color me unconvinced. In many ways, the risk of climate change is an aggregate of the risks faced by individual regions. The warming is not uniform over the globe, and with projected global warming, there are both ‘winners’ and ‘losers’ (see my Testimony). Another relevant question is whether a future ice age – occurring naturally – would be considered as ‘ruin’? The impacts of a future ice age are arguably more severe than doubling or even tripling CO2 (even if you believe IPCC projections). Not to mention that CO2 warming would delay a future ice age. I’m not really seeing AGW as a ‘ruin’ problem, i.e. a ‘catastrophe’. And finally the issue of ‘recovery’ is a time-scale issue (what is ‘forever’?) – e.g. the ice sheets in an ice age will eventually retreat. This whole issue of ‘ruin’ ties back to the issue of what actually constitutes ‘dangerous’ climate change. See these previous posts on the issue of ‘dangerous’:
Scientific methods and the PP
The idea of precaution is the avoidance of adverse consequences. This is qualitatively different from the idea of evidentiary action (from statistics). In the case of the PP, evidence may come too late.
In an evidentiary approach to risk (relying on evidence-based methods), the existence of a risk or harm occurs when we experience that risk or harm. In the case of ruin, by the time evidence comes it will by definition be too late to avoid it. Nothing in the past may predict one fatal event. Thus standard evidence-based approaches cannot work.
More generally, evidentiary action is a framework based upon the quite reasonable expectation that we learn from experience. The idea of evidentiary action is embodied in the kind of learning from experience that is found in how people often react to disasters—after the fact. When a disaster occurs people prepare for the next one, but do not anticipate it in advance. For the case of ruin problems, such behavior guarantees extinction.
JC comment: This clarifies the conflict between ‘lukewarmers’, who seem mainly data driven and don’t see danger (in favor of risk management), versus alarmists who argue for the PP to avoid possible catastrophe or ruin (as inferred from climate model simulations).
Distinguishing global and local risks
Since there are mathematical limitations to predictability of outcomes in a complex system, the central issue to determine is whether the threat of harm is local (hence globally benign) or carries global consequences. Scientific analysis can robustly determine whether a risk is systemic, i.e. by evaluating the connectivity of the system to propagation of harm, without determining the specifics of such a risk. If the consequences are systemic, the associated uncertainty of risks must be treated differently than if it is not. In such cases, precautionary action is not based on direct empirical evidence but on analytical approaches based upon the theoretical understanding of the nature of harm. The essential question is whether or not global harm is possible or not. Theory enables generalizing from experience in order to apply it to new circumstances. In the case of the PP, the existence of a robust way to generalize is essential.
JC comment: However, a systemic global risk does not necessarily imply ‘ruin.’ Climate models are the establishment’s tool for understanding the nature of the harm from atmospheric CO2. Are climate models a robust way to project the future harm? I would say no. Using empirical evidence (e.g. in context of the energy balance models) would seem to be more robust. Black swan events are still missed by both methods, which is why I have argued for alternative methods for scenario generation:
- Worst case scenario versus fat tail
- Broadening the portfolio of climate information
- Alternative approach to assessing climate risk
Fat tails and fragility
To figure out whether a given decision involves the risk of ruin and thus warrants the use of the PP, we must first understand the relevant underlying probabilistic structures. There are two classes of probability distributions of events: one in which events are accompanied by well behaved, mild variations (e.g. Gaussian or thin tails), and the other where small probabilities are associated with large variations that have no characteristic scale (e.g. power law or fat tails).
Given a series of events, in the case of thin tails the sum is proportional to the average, and in the case of fat tails a sum over them may be entirely dominated by a single one. In thin tailed domains harm comes from the collective effect of many, many events; no event alone can be consequential enough to affect the aggregate. In fat tailed domains of risk harm comes from the largest single event.
JC comment: Fat tail and thin tail risks is a good way of looking at extreme events. I effectively adopted this approach in projecting future hurricane damage in Central America and the Carribbean [link].
Limitation of top-down engineering in complex environments
In considering the limitations of risk-taking, a key question is whether or not we can analyze the potential outcomes of interventions and, knowing them, identify the associated risks. Can’t we just “figure it out?” With such knowledge we can gain assurance that extreme problems such as global destruction will not arise. Planning fails due to the inability to anticipate the many conditions that will arise.
JC comment: This relates to the problem of confusing a wicked problem with a tame problem, see these previous posts:
Precaution as policy
When there is a risk of ruin, obstructionism and policy inaction are important strategies, impeding the rapid headlong experimentation with global ruin by those with short-term, self-centered incentives and perspectives. Two approaches for policy action are well justified. In the first, actions that avoid the inherent sensitivity of the system to propagation of harm can be used to free the system to enable local decision-making and exploration with only local harm. This involves introducing boundaries, barriers and separations that inhibit propagation of shocks, preventing ruin for overly connected systems. In the second, where such boundaries don’t exist or cannot be introduced due to other effects, there is a need for actions that are adequately evaluated as to their global harm. Scientific analysis of such actions, meticulously validated, is needed to prevent small risks from causing ruin.
JC comment: This is an argument for dealing with the risk via increasing resilience and anti-fragility. For previous posts on this topic, see:
What is not justified, and dangerous, are actions that are intended to prevent harm by additional intervention. The reason is that indirect effects are likely to create precisely the risks that one is intending to avoid.
JC comment: This is the key issue in the climate policy debate – is the ‘cure’ (i.e. CO2 emissions reduction and associated economic hardships) worse than the ‘disease’ (i.e. warmer temperatures)? A few previous posts on this topic:
- A precautionary tale: more sorry than safe
- Why the decision to tackily global warming isn’t simple
- Uncertainty, risk and (in)action
- Decision making under uncertainty: the dog and the frisbee
When existing risks are perceived as having the potential for ruin, it may be assumed that any preventive measure is justified. There are at least two problems with such a perspective. First, localized harm is often mistaken for ruin, and the PP is wrongly invoked where risk management techniques should be employed. When a risk is not systemic, overreaction will typically cause more harm than benefits, like undergoing dangerous surgery to remove a benign growth. Second, even if the threat of ruin is real, taking specific (positive) action in order to ward off the perceived threat may introduce new systemic risks. It is often wiser to reduce or remove activity that is generating or supporting the threat and allow natural variations to play out in localized ways.
Preventive action should be limited to correcting situations by removing threats via negativa in order to bring them back in line with a statistical structure that avoids ruin. It is often better to remove structure or allow natural variation to take place rather than to add something additional to the system.
JC comment: in the climate change debate, I interpret this statement as arguing to increase resilience and anti-fragility.
When one takes the opposite approach, taking specific action designed to diminish some perceived threat, one is almost guaranteed to induce unforeseen consequences. Even when there appears to be a direct link from a specific action to a specific preventive outcome, the web of causality extends in complex ways with consequences that are far from the intended goal. These unintended consequences may generate new vulnerabilities or strengthen the harm one is hoping to diminish. Thus, when possible, limiting fragilizing dependencies is better than imposing additional structure that increases the fragility of the system as a whole.
JC comment: In the climate policy debate, I regard this as arguing against new energy policy restrictions or engaging in geoengineering.
Putting this post together was very interesting for me, since Taleb’s paper integrates a lot of things that I have been writing about, and this afforded me an opportunity to integrate them (not to mention read some of my old posts).
What does Taleb have to say about climate change? Not much, although this statement does appear in the paper:
The more uncertain or skeptical one is of “scientific” models and projections, the higher the risk of ruin, which flies in the face of the argument of the style “skeptical of climate models”. Hence skepticim about climate models should lead to more precautionary policies.
Ouch. Perhaps Taleb should read my recent post:
A salient issue is that it is difficult to apply these concepts to individual risks without some understanding of the domain itself (this seems to be the source of the problems with Taleb’s conclusions re GMOs, which is admittedly an issue with many shades of gray).
I have attempted to apply Taleb’s ideas to the climate change problem (although I may be hampered by lack of expertise on the the topics that Taleb writes about). The key insight that I feel that I have developed from this exercise is that urgent global CO2 mitigation is justified only if we are facing ruin, and can’t be justified by an economic cost-benefit argument. The potential for adverse and unintended consequences are just too great.
This brings us back to ‘ruin’, ‘catastrophic’, and ‘dangerous’. The AR5 WGII Report made an important step forward and clarifying this issue, my comments:
. . . the message that I am getting is there is a great deal of uncertainty in the attribution and future projections of climate change impacts, and that the threats on the timescale of the 21st century are not existential. The ‘message’ has shifted from documenting dire impacts, to finding solutions that integrate with broader societal challenges.
The global nature of CO2 forcing does not imply global risks; the risks are local/regional, and the aggregation of these risks does not imply the potential for ruin.
The enshrinement of the Precautionary Principle into the UN Framework Convention on Climate Change represents a mismatch between the problem and the proposed solution; this is described at length in Rayner and Prin’s essay Wrong Trousers that was discussed in my post IPCC diagnosis: permanent paradigm paralysis.
The ideas of resilience and anti-fragility (e.g. adaptation) are much better suited to the climate change challenge, since benefits can be obtained now in the most vulnerable regions and vulnerability can be reduced to a broad range of threats. Further, the adaptation solution avoids the hubris of thinking we know what the future climate holds.
So does this imply we should do nothing about energy policy? Well there are many other drivers for energy policy, such as human and ecosystem health, environmental quality, energy security, etc. If climate change is NOT a ruin problem, we need to stop pretending that urgent action to reduce CO2 emissions is justified by anything other than naive reasoning and the role of personal and political preferences to address climate risk, in a way that increases global government control.
So is my reasoning correct? The bottom line is that I think much more attention needs to be given to this kind of meta-analysis and assessments of the risk from climate change, including speculation on the plausible worst case scenarios.
I will ping @nntaleb on twitter, to see if I can engage him in this discussion.