Greenhouse policies in a complex world – some analytics

How do you design an appropriate response to anthropogenic climatic change when the costs and benefits of dealing with the problem are uncertain? The physical and environmental effects of climate change are uncertain, there is controversy over the discount rate to be used and damage functions are highly nonlinear, involve important irreversibilities and policy makers must address very long time horizons.

Robert Pindyck’s ‘Uncertainty in Environmental Economics’ discusses these issues for environmental problems with emphasis on climate change issues. I specialise his discussion to the design of adaptation policies for preventing species extinctions in the face of global warming – a topic that holds my current interest.

Meteorological science shows that the relation between greenhouse gas emissions (GHGs), temperatures and climate patterns is highly uncertain. We don’t know the exact effects on climate of measures to control emissions. Nor do we know the effects of climate changes on ecosystems or of the effects of policy measures that seek to improve the survival of species. We are therefore uncertain about the adaptability and resilience of different species both now, and in the longer-term, in the face of climate change and an adaptation policy response.

These uncertainties have implications for the design of adaptive policies. Hybrid policies make more sense than dedicated uniform policies in settings where policy makers choose between ‘price’ and ‘quantity’ instruments and the same will probably be true if the policy task is to select among various ways of helping biodiversity to adapt to climate change. Strengthening existing conservation programs might make sense as might providing translocation sites and, perhaps, captive breeding programs. The optimal policy hybrid will depend on the extent of the uncertainties.

The effects of uncertainty on the intensity of policy response are complex. Without irreversibilities uncertainty often leads to lower policy intensities – here less conservation effort. This would be the case if conservation benefits are an increasing concave function of investment in reserves and so on and if conservation costs are convex in the extent of conservation. But irreversibilities are fundamental so this special case is not robust or particularly useful.

Pindyck also suggests there is substantial discount rate uncertainty in GHG problems. In my view this can be exaggerated since, if discount rates are anything other than very low, the policy answer is simple – do nothing – the discounted benefits then won’t justify the costs. Moreover, intergenerational arguments suggest we should discount at a low rate. Pindyck’s response would be that being uncertain about the discount rate means itself that the rate that is used should be set at less than the expected future discount rate with the difference here increasing as the time-horizon for planning increases. Indeed, for long-term planning, a rate close to zero is not implausible.

Climate change issues do indeed involve very long time horizons. This exacerbates the uncertainties associated with estimating future costs and benefits particularly if a low discount rate is used.

It is also plausible to suppose damage functions defined on adaptation conservation effort are highly non-linear with highly uncertain ‘tipping’ or ‘threshold’ points. An insufficiently large effort might tip the environment towards a catastrophic state where many species become extinct.

In addition, there are significant irreversibilities that pull in two opposite directions. These substantially complicate policy planning.

(i) Species loss effects – in particular species extinctions - are irreversible. This creates incentives to adopt active management policies early even if benefits from doing so fall short of costs. The existence of the possibility of a future benefit from not allowing a species to go extinct provides an ‘option value’ rationale for species protection that reflects the irreversible loss of options those extinctions imply. This provides a bias towards early adoption of policy.

(ii) Sunk cost effects however also arise because policies to deal with GHG problems impose sunk costs for society. For example, investing in captive-breeding programs or in translocating species are discrete investments that might prove unwise should uncertain policy costs and benefits turn against such provisions. For example if long-term conservation costs should turn out to be high and long term benefits low. This provides a motivation to wait for better information before investment in species protection is undertaken. Cost benefit analysis which accounts for the value of waiting for better information, will be biased toward postponing adoption of such policies.

Thus irreversibilities combined with uncertainty can sometimes increase and sometimes decrease desired conservation effort. What a pity! There is no neat analytical result! Without good empirical evidence – and this seems lacking to this point – there is no way of making a simple qualitative judgement. One needs to assess the relative scale of species loss and sunk cost effects.

On this issue some of the literature suggests a reduced level of control and waiting as opposed to early action: see Pindyck (2000), (2002). One can ask if this story is changed when allowance is made for the prospect of a catastrophic risk of widespread species extinctions. In a paper I wrote with William Reed in 1994 we show that if the degree of risk of catastrophic collapse is strongly related to the extent of policy intervention – so for example a critical level of lack of conservation effort will bring about a wave of very costly extinctions - that greater stringency in control is sought.

My intuition too is that the case for early action is motivated by standard arguments that emphasise the increasing value of natural capital relative to person-made capital as society advances. With increasing wealth person-made capital can be replaced but natural capital cannot be. Also with increasing wealth and supposing that demands for biodiversity conservation are a luxury good – more is demanded with increased affluence – there is again a ‘preference’ argument favouring early adoption of adaptation policies for improving the resilience of biodiversity.

The AEI-Brookings Joint Centre for Regulatory Studies has a range of readable economic papers at its website – including this excellent paper by Robert Pindyck. I’ve attached the site to my blogroll.

Reference: Harry Clarke & William Reed, ‘Long-Run Consumption Pollution Tradeoffs in An Environment Subject to the Possibility of Pollution-Related Catastrophic Collapse’, Journal of Economic Dynamics and Control, 18, 1994, 991-1010.