How long can an ecosystem endure pressures without collapsing? How high is the climate sensitivity? What are the long-term consequences of growing and consuming GMO? How much renewable energy is feasible? How many species can get extinct without destabilizing an ecosystem? What is the (future expected) value of biodiversity? How long will oil, coal and uranium last? How large are the dangers from nuclear power? These are a few important questions that are highly uncertain. We probably cannot answer them exactly – at least not ex ante. So, how to deal with uncertainty when it comes to decision-making?
In conventional economics, the common approach to almost every problem – from an investment in a new desk computer to climate change mitigation – is the cost-benefit analysis. You calculate the costs and the benefits of an action (when needed, you discount them), then compare them with each other – if the benefits are higher, the project is worth doing. Simple as it sounds, this approach, while making sense in normal business processes, bears many problems when applied to more complex issues. For instance when applied to environmental problems. One critical aspect is the mixture of irreversibility and non-substitutability, two inherent characteristics of environmental challenges. Another is uncertainty.
While it is common practice in environmental economics to largely ignore uncertainty – “best estimates” and the like are treated as if they were exact parameters -, some conventional economists have tried to account for uncertainty in their CBAs. An interesting example is the Stern Report on the economics of climate change. The authors didn’t simply assume that the best estimates from climate science are “certain”, but they applied the so-called Monte Carlo analysis. In this case, the uncertain variables are imputed as probability distribution functions and the modeling software is randomly “picking out” numbers from the distribution, producing hundreds of different “scenarios”. The range of these scenarios gives an idea about the underlying uncertainty and its implications.
Nevertheless, in some cases this seems not to be enough. As shown by Martin Weitzman, when the considered probability distribution functions are “fat-tailed” (i.e., the probability of extreme outcomes is not approaching zero) and there is much at stake (e.g., a stable climatic system), every cost-benefit analysis will be overrun by what he calls the Dismal Theorem. In short: whatever the concrete numbers imputed, if there is some probability of extreme (negative) outcomes, and if these outcomes would be catastrophic, then it will always be worthwhile to undertake much to prevent them.
Weitzman’s Dismal Theorem can be applied to many large, complex environmental problems (I named some examples in the opening paragraph). It is a mathematical expression of what is commonly called the precautionary principle.
The precautionary principle as a policy recommendation would make sense if people tended to risk-aversion. Indeed, as empirical studies have repeatedly shown, we are strongly risk-averse. Of course, there are exceptions – gambling addicts, stuntmen, bankers are a few of them. But on average people behave highly risk-averse. (Otherwise the insurance business wouldn’t be so profitable.) So, if the society is risk-averse, it would be appropriate to choose policies that reflect that. Indeed, it is what we see in our everyday life: public health insurance for instance. Tax-financed unemployment schemes. And so on. All these are some forms of the precautionary principle.
Surprisingly, in the area of environmental protection, this rule is not applied that often. Cost-benefit analyses are often called for despite the uncertainties involved (as well as non-substitutability, irreversibility and valuation issues). As shown above, CBA is not a very good tool in this field. The question is: what do alternatives look like?
I have no ready-made proposal. But it appears reasonable to rely more on public discussion or, more pathetically expressed, on democracy. Cost-benefit analyses are allegedly “objective” and “scientific”. Indeed, they would be – if there were no uncertainty, if changes to the environment were reversible, if environmental “goods and services” were easily valuable, if they were substitutable… But they are not. Therefore policy makers should turn their attention to the second-best – and this seems to be the precautionary principle in general, and more participation by the civil society in particular. This may be less scientific and less objective than the ideal (a sound CBA). But it certainly would be less subjective than an analysis in which arbitrary judgements by some technocrats and/or bureaucrats are necessarily involved.