Cost benefit analysis – don’t get over-excited

Cass Sunstein’s new book, The Cost Benefit Revolution, almost managed to achieve the opposite of its aim, which is to persuade me of the merits of ubiquitous use of this tool for assessing social welfare so beloved by economists. I’m an economist so I think CBA is pretty useful but the book’s just so Panglossian about its ability to apply objective reason to regulatory decision making. “In a nutshell, quantitative cost benefit analysis is the best available method for assessing the effects of regulation on social welfare,” he writes (his italics). Except it doesn’t. He isn’t alone in thinking this. I once said to a Very Distinguished British economist that economics needed to take social welfare far more seriously. “Oh but it already does,” he replied. “We use CBA all the time.”

But the point is – as Sunstein does sort of concede in accepting that there are what he calls Hard Cases – it ignores (and the Pareto criterion ignores too) one of the key points about social welfare, namely distribution. Whose costs and whose benefits? It also – as has always been known in the literature, reiterated recently, and always ignored in practice – is a linear approximation suitable for use to assess the net effects of small changes. It furthermore has no validity in assessing any intervention that might change the path of growth or productivity or have any non-marginal effect.

Economists do think of CBA as one of the practical contributions of the field to policy, and I’m in favour of doing the CBA exercise as a source of information about decisions. In practice, for big decisions, that’s exactly what happens. Politics takes over, not least for the distributional reasons. After all, if CBA had been in use in the 19th century, Bazalgette would never have built London 150 years worth of sewerage capacity, and we would lack the magnificent town halls of cities like Manchester and Leeds.

Elsewhere in the book, in the excellent historical sections describing the (non-partisan) spread of CBA in US government, Sunstein gives some persuasive examples of how to use CBA well. (One of the poignant sections describes how an analysis changed Ronald Reagan’s mind – a Republican president open to reason.) There are nice, classroom friendly, examples of using CBA in different domains of policy and is actually more nuanced than the general enthusiasm would suggest about how to use CBA in the context of environmental science, especially when risk attitudes and discount rates differ greatly between groups of voters. All in all, it’s worth the read for anyone interested in the role of reason in policy making. Just don’t drink the same Kool Aid.

PS Tim Harford reviewed the book today and is a bit more enthusiastic.

Share

2 thoughts on “Cost benefit analysis – don’t get over-excited

  1. In principle it is a good idea. But it needs reliable data and clear headed analysis. Also, it should not be encumbered by any value judgement that bears directly on either the data or analysis. If you do not and also if things change too much and too rapidly in the course of the work then you could either end up with the costs being badly wrong and the benefits not achieved. It is not easy and needs great care and the ability to withstand fake figures, news or special pleading and pressures.

  2. Cost Benefit Analysis is just one denomination within The Church of Cardinal Utility. Like all religions it has not logical foundation:

    https://www.macrobusiness.com.au/2013/05/the-insufferable-conceit/#comment-241591

    There IS a logically consistent ways of aggregating ordinal preferences: the Democratic Eigensolution to the choice of aggregation device which is then used for subsequent aggregation.

    The religion of Cardinalism is not the solution.

Comments are closed.