What do the sub-prime crisis, the controversies about climate change and the acrimonious disputes about BT Brinjal have in common?  All three are a product of methodological weaknesses in the way in which risk and uncertainty have been handled in the relevant debates and decisions.

The sub-prime crisis illustrates the hazards of an “other things remaining constant” approach to risk assessment.  A portfolio of mortgages was evaluated by aggregating the default risks of each individual borrower on the basis of that borrower’s economic characteristics and the historical default data of each borrower class.  Mortgages from different risk classes were packaged and the aggregate was evaluated on the assumption that these risks of default are independent.  Some insurers provided cover against default and, on the strength of this, rating agencies gave these securities an investment grade.

But the default risks were dependent on house prices continuing their upward march. When house prices started falling all of the borrowers were affected and this systemic risk was not factored into the rating of the portmanteau securities that included “good” and bad” mortgages. 

If the primary source of risk is individual, the risk assessment would have worked the way it works for the risks that life insurance actuaries calculate.  But if some external event alters the risk profile of every borrower this “other things remaining constant” procedure does not work.  We have known this in India in the sphere of crop insurance where an adverse event like a drought affects all borrowers. That is why we do not have a viable model for insuring against crop failures.

Droughts are predictable, at least in a statistical sense.  But economic forecasts have been notoriously poor in predicting turning points. An event like the huge drop in house prices that generated the systemic risk for mortgage securities was not predictable in the sense that it fell outside the usual confidence limits of economic forecasts..  It was what Taleb would call a Black Swan event.  This is a risk at the tail end of a probability distribution with low probability but huge costs.  How many risk managers had worked out the implications of this possibility in evaluating the mortgage securities?

The current controversies about climate change illustrate another problem in the way in which we handle uncertainty.  Climate science involves a large degree of uncertainty and the IPCC, to its credit, has been punctilious in giving confidence limits to its assertions.  But the public debate is being conducted by replacing this scientific uncertainty with adversarial certainties.

The discipline of heuristics has something to say on this.  It talks of a prior attitude effect which evaluates supportive evidence and arguments more favorably than contrary indications and also picks and chooses these to suit its case.  It also suggests that knowledgeable subjects, because they possess greater ammunition with which to counter-argue incongruent facts and arguments, will be more prone to these biases.  One sees these biases both among the climate alarmists and the climate skeptics.  When arguments are weak, the proponents also resort to mud slinging thus diverting attention from the core argument.

The mainstream of scientists in the IPCC, however, have been professional and cautious and came out with a categorical statement about the human impact on global warming only in their third report.  The fact that the warming trend has not been observed for some years does not invalidate the science on which the IPCC’s projections rest. Much of the recent skepticism is nit picking at the periphery of the case. My confidence in the IPCC rests on the wide spread of it scientific network and the fact that its basic argument is supported by virtually every National Science Academy that has spoken on the matter and in most peer-reviewed academic articles on the subject.

The real case for action on climate is as insurance.  The “pathological reports” on the state of health of the climate system have been getting progressively worse.  The worst case could involve catastrophe and horrendous costs.  The immediate actions that we need to take are not just low cost but often have large collateral benefits.  Later, if the alarm proves unwarranted low carbon policies can be relaxed.  But the reverse option of undoing past neglect if the prognosis proves to be right would be hugely expensive.

The BT Brinjal case also illustrates some of these problems when the science is arcane and uncertain so that the lay public is at the mercy of the adversarial experts.  The problem also arose because of the lack of confidence in the judgment of the experts in the GEAC. The minister concerned, Jairam Ramesh, did well to listen to all sides.  But he, like the rest of us, was confronted by competing certainties of experts who refused to see the other side of the argument.  He did well to come out with a statement that is an object lesson in how such decisions should be presented to the public.

When confronted by risks that cannot be readily calculated by looking at the frequency of past occurrences, it is best to treat probability as a measure of the degree of belief and link it to the availability of evidence.  When evidence is scarce, the dispersion of belief can be justifiably wide.  As evidence improves the dispersion of belief should come down.  Such an approach would avoid the clash of competing certainties and stresses the need to mobilize evidence with the clear aim of reducing the spread of belief.

An over-simplified view about the predictability of uncertain events, treating linked risks as independent of one another, the peddling of certainties by prejudiced parties and doubts about the objectivity of the experts on whom lay decision makers have to depend are the common hazards of dialogue and decision when outcomes are uncertain.  The answer must lie in investing in more and better evidence to reduce the range of uncertainty, looking not just at individual but also at systemic risks, paying particular attention to potentially catastrophic risks, working hard to maintain an open dialogue, being courageous enough to change one’s views and ensuring that official expert groups that advise on the underlying science include specialists of all persuasions. 

Coping with uncertainty requires the courage to confront the evidence, even when it clashes with ones preconceptions, and the caution to avoid catastrophe.