Risk and Reason

Risk and Reason: Safety, Law, and the Environment
Cass R Sunstein

Peter Rossi

If the idea of a nanny state or a police state does not appeal, how about a cost-benefit state?

This conception of a state is precisely what Sunstein argues for in his new book, Risk and Reason. The supporting argument is essentially as follows: cost-benefit analysis (CBA) ensures 'that when government acts, it does so in a way that is informed by a full account of its consequences' (p. 190). In other words, if government is going to produce regulation, it seems desirable that it should do so in the most efficient way possible. Surprisingly, Sunstein does not provide a methodology for CBA, instead treating the analytical tool as a black-box that produces answers from the data fed into it. This lack of detail is disappointing, especially as CBA is by no means an uncontroversial method.

Superficially, there seems to be little to object to in this notion of efficiency. After all, resources are scarce, and the corollary is that actions should be economic; that is to say that they should not be wasteful. However, the assumption that Sunstein's conception is predicated on lies at the crux of the matter. To enter into a discussion of CBA, we first have to assume that government should regulate. This qualification is not discussed in Risk and Reason. For instance, Sunstein might ask what is the most efficient way to prohibit drugs, rather than enquiring about the moral validity of prohibition itself. To be fair, Sunstein confines the discussion to arguably less clear-cut cases, such as air pollution, but he leaves this problem unsolved. It is for this reason that CBA is described, on more than one occasion, as a 'pragmatic' approach. For example, if CBA leads to 'paralysis by analysis' (when government uses too many resources in the CBA to legislate properly), it should be abandoned (p.120).

Another pragmatic suggestion is that if the CBA shows that the costs of proposed legislation outweigh the benefits, the legislation should not automatically be scrapped. For example, in discussion of sulphur dioxide emissions, '[i]f the costs mean lower profits, and the benefits mean longer lives, it seems clear that the government should act, even if the costs are higher than the benefits' (p.124, also p.255). This raises numerous questions, but the most pertinent here is: if the CBA is not decisive, what is the point of using it? The answer Sunstein gives throughout the book is that CBA brings the consequences of a regulation 'on-screen', by making them apparent to all concerned, hence facilitating more efficient policy-making and also acting as a heuristic device. But this answer raises more serious questions, such as can all the consequences of a regulation be known, and if so, can their associated risks be calculated and then monetized? There is far too little space to discuss this here (see Risk, by John Adams, for an excellent discussion of this very problem), but I believe the answer to be no, no and no!

We can never know all the consequences of government regulation or any complex human action, we cannot measure risk in any meaningful sense, and subsequently, we cannot cost it. If this statement is true, then CBA is virtually useless, except as a very primitive heuristic. Socialist planning provides us with ample evidence that government command and control policies do not work as intended. They raise the problem of what Hayek called unintended consequences. Sunstein in fact welcomes a 'Hayekian turn' towards a free-market approach to risk (p.131, and Ch.10), but only on pragmatic efficiency grounds.

More controversially perhaps, we cannot measure risk in a meaningful sense, because of its highly complex (even chaotic) and individualistic nature. For example, if we say that the risk of a road accident is 1 in 10,000, of what use is this figure? Does it apply to an intoxicated, male, teenager driving home from a party or a well-trained policeman driving to work? Also, is the risk actually higher, but not recorded in the traffic accident data as such, because drivers perceive the risk and act to reduce it by driving more carefully? Furthermore, this risk might be acceptable to one person, and not to the next. Consequently, aggregating risk is a process of compromise which results in a measure of risk to a fictitious average individual, and is of little use to you or me.

Therefore, the problem of monetization, fundamental to CBA, is seemingly intractable. Sunstein provides numerous data purporting to cost government policy, but they are not convincing for the reasons discussed above. Costing regulation also ends up valuing lives saved or lost, as a measure of policy efficacy, which is a distasteful business to say the least. It raises the issue of a Bangladeshi being worth less than an American! Also, in the CBA black-box, a method called willingness-to-accept is used, whereby an individual is asked how much s/he would be willing to accept as compensation for a certain effect. If the effect is death, and the individual not unreasonably answers 'infinity', how is this measured? One approach is to ignore it, treating the answer as an 'outlier'.

Not withstanding the substantial problems with CBA, the first four chapters of Risk and Reason are very interesting discussions of the psychology of risk. Sunstein describes how the unashamedly technocratic approach of CBA is an attempt to reduce the influence of interest groups on policy making, and also to minimise the spread of scare stories when accurate information is in short supply. This section includes a particularly useful discussion of the Love Canal incident in the USA in the late 1970s which some claim as the genesis of modern environmentalism. The Love Canal story consists largely of lies and exaggerations, and has lead to a movement based largely on lies and exaggerations.

If you believe that CBA is a valid technique, Risk and Reason will provide you with plenty of supporting evidence, particularly related to US law. Overall however, Risk and Reason is an interesting book built upon the questionable foundation of CBA. The penultimate chapter on free-market approaches to risk overshadows the preceding discussions of CBA, for it becomes apparent that a free-market in risk is far superior to crude governmental substitutions of the price mechanism. In some of the accounts given by Sunstein, it seems that CBA is little more than glossy socialist central planning, and as a result, a highly dubious technique.