The mainstream acceptance of both the scientific consensus on global warming and the need for a globally binding capon carbon emissions does not by itself indicate that we are close to reaching a satisfactory solution to climate change.   The optimal policy response is still subject to massive uncertainty from three sources: risk around our central forecast of climate change costs, uncertainty about the cost of reducing emissions, and uncertainty about how we should share the cost burden.

The solution to the textbook case of environmental pollution is a tax equivalent to the cost to society of cleaning up (or adapting to) the pollution.   That principle is now being applied to the problem of global warming through the application of a carbon tax (or the theoretically identical cap and trade system).

However there is an important difference between the textbook case and climate change – our estimate of the costs of climate change come with wide error bands.   Even if we accept that the International Panel for Climate Change’s scenarios remain the best assessment of current scientific knowledge, it remains the case that there is substantial uncertainty around that "best" forecast.

A review of the IPCC methodology by two forecasting experts was uniformly scathing about the process used to produce consensus forecasts, and it is well-documented that experts tend to overestimate the precision of their predictions.   (As an analogy, the current global financial crisis stems in part from outcomes that fell far outside the predicted range.)

In simple terms, this means that the impact of climate change could still be much better, or much worse, than we are assuming.   Of course, the costs of alternative outcomes are not necessarily balanced: the good news scenario is that we’ve spent money for nothing; but the worst case is that even drastic measures can no longer prevent catastrophic climate change.

What does this uncertainty mean for our textbook response of a carbon tax?   With an uncertain cost, it is no longer sure that setting the carbon tax equal to our best forecast of cost is optimal.   It is more appropriate to think of a carbon emissions target (and corresponding carbon tax) as like taking out a form of insurance policy.   What we are really buying is protection against a range of outcomes, but not again stall possible outcomes.

Suppose are current target gives us an 80% chance that global warming will be within an acceptable limit.   For an extra cost we could be 90% sure – but the point is that we quickly hit diminishing returns in terms of the extra safety we get for our dollars.

This kind of analysis produces objections from those that believe that the worst-case scenario for global warming, even if only a miniscule probability, is severe enough to justify spending any sum to avoid.   However, this position is difficult to reconcile with the rather small sums we spend to protect ourselves from other extremely nasty but rare eventualities – bird flu spreading to human populations for example – or the relatively small sums spent on improving life expectancy in the third world.

Since global warming is a collective action problem, the world as a whole has to choose how much insurance to buy.   This leads to our second problem: our forecasts for how much greenhouse gas abatement will cost are also subject to tremendous uncertainty (and the uncertainty increases the more ambitious our target becomes).

"Climate change insurance" is like any other good: if it is cheap we will be happy to buy a high level of protection, but if it is expensive we become more prepared to take our chances (and at some level of expense adaption to climate change becomes cheaper than prevention).      

There might be some cause for optimism here: attempts to reduce sulphur dioxide emissions in the US via a market mechanism were ultimately much cheaper than initially predicted.   However, a larger proportion of carbon dioxide emissions are from energy, for which there is already a price incentive for conservation.   And recent increases in oil prices have produced little in terms of reduced oil demand (so far).

Even if we settle on an acceptable cost, the question remains of who will pay?   This means allocating the cost both between rich and poor countries, and between current and future generations.  Equity arguments cut both ways here: it is unfair to expect developing countries to do too much (although they are quickly becoming the major source of emissions growth).   It also seems unfair to place most of cost burden on the current generation to the benefit of future, richer generations.

One of the more heated economic debates on climate change response has actually concerned the appropriate interest rate with which to discount future costs – too high a rate leads us to ignore the distant future costs of global warming, but too low implies a concern for the future out of proportion with the amount of saving and bequeaths actually observed.

No matter what we resolve to do in response to global warming, there will remain reasonable and informed debate as to whether we are doing enough or too much, and whether we are spreading the costs fairly.   And it seems inevitable that the policy response we ultimately settle upon will still leave us with a greater than negligible risk of drastic climate change.