While economists often lament that governments don’t use cost-benefit analysis enough, Harvard Law Professor and former US regulatory czar Cass Sunstein offers a much more optimistic perspective in his excellent 2018 book The Cost-Benefit Revolution, which I’ve only just got around to reading. Sunstein’s first chapter is titled “The Triumph of the Technocrats”. Hurrah!
Sunstein notes that cost-benefit analysis has been increasingly applied to the assessment of regulations in the US since 1981 when the Reagan administration issued Executive Order 12291, making cost-benefit analysis of regulations compulsory. According to Sunstein, the requirement for cost-benefit analysis has been a great success. In chapter 1 (p. 10) he notes:
From 1981 to the present, cost-benefit analysis has often been a decisive decision rule in significant cases.
Unfortunately, as far as I can tell, the book doesn’t elaborate on what those significant cases were. It is very good on theoretical and philosophical issues regarding cost-benefit analysis, but it would have benefited from some detailed case studies, in my view. One of the practical examples the book does provide demonstrates the adverse impacts of failing to undertake a cost-benefit analysis. Sunstein argues that a cost-benefit analysis would have demonstrated that asthma puffers should have been exempt from the Montreal Protocol to ban CFCs. Sunstein is very good at demonstrating situations where cost-benefit analysis can be used, but the book would have been much better with case studies of its successful application.
As the co-author with Richard Thaler of Nudge, you’d expect Sunstein to include a lot of behavioural economics insights in The Cost-Benefit Revolution, and indeed he does. He notes people have a bias toward the present and can be overly optimistic and under-estimate risks. A cost-benefit analysis quantifying the impacts of a regulation, in terms of its costs, level of risk reduction, and the dollar benefits of that risk reduction, can help us avoid those biases.
Sunstein is no over-zealous advocate of cost-benefit analysis, though. He argues equity considerations may be relevant in policy decisions and, in some cases, you can’t adequately forecast what the outcomes of a policy measure may be. In some cases, rather than developing a detailed policy and deciding whether to proceed with it based on an ex ante cost-benefit analysis, policy experiments may be desirable instead. In chapter 5, “The Knowledge Problem”, Sunstein relates how he learned about the measure-and-react strategy from a Silicon Valley member of the US Defense Innovation Board, after he suggested Defense needed to apply cost-benefit analysis more. Sunstein gives clothing manufacturer Zara as an example of a company applying a measure-and-react strategy. As readers may know, Zara is the exemplar of “fast fashion” and has developed a highly efficient production and distribution system which allows it to quickly produce clothes according to what’s on-trend. Sunstein notes (on p. 98):
The measure-and-react strategy is not a randomized controlled trial, but it serves the same functions. It is increasingly used by private-sector actors, who know what they do not know and try to adjust to what people are doing on the fly. Can governments do the same thing? In many contexts, they certainly can.
There may well be benefits from greater policy experimentation by government. The measure-and-react strategy could usefully be applied to what Nicholas Gruen calls policy hacks (e.g. see his Mandarin article What is a ‘policy hack’?). Incidentally, Nicholas will be appearing soon in an upcoming episode of my Economics Explained podcast, which I expect to release this Thursday.
I should note The Reagan Executive Order mentioned above was hugely influential. Indeed, the perceived necessity for cost-benefit analysis spread throughout the world, including to Australia where the Commonwealth and state and territory governments all now typically require cost-benefit analysis of proposed regulations. Cost-benefit analysis studies are also usually required for infrastructure projects. In my first Economics Explained podcast interview, I talked about the importance of cost-benefit analysis in infrastructure decision making:
Economics Explained podcast interview with Craig Lawrence
Regrettably, unfavourable cost-benefit analysis studies are sometimes ignored for politically-driven infrastructure projects, such as the Rookwood Weir, a project which ought to feature in the next season of Utopia (see my post Rookwood Weir business case should be re-worked).
Finally, I’ve also recently read The Education of an Idealist by Sunstein’s wife Samantha Power, former UN ambassador to the UN in the Obama administration. Her book’s worth reading, too, particularly for its frank and revealing account of foreign policy decision making during the Obama years.
Queensland Economy Watch 
Governments requiring CBA for regulatory proposals is one thing, government agencies doing them is another thing, government agencies doing them well is another thing again….in reality there is little demand for good quality CBA by Australian governments. Governments are only interested in CBA that justifies their decisions (usually after they have made them!). This weakness in policy development is well known and understood by regulatory oversight bodies across Australia. Sadly, regulatory impact analysis has little positive impact on the quality of regulation in Australia, it is simply used as a marketing tool by Australian governments and allows them to say “we have robust regulatory processes in place”! It’s a con job.
Thanks for the comment RIS Meister. I’d like to think the RIS process imposes some constraint on regulatory agencies, and at least stops really silly and costly regulations getting through. I agree there are a lot of problems with it, though.
I’d like to think that too Gene. But when Qld regulations such as the container refund scheme get implemented which have large costs on consumers and marginal environmental benefits, it is hard to fathom that having a regulatory impact analysis process in place has made things better than they otherwise would be. Such processes are pretty impotent in the face of governments that want “to do something”.
The other critical thing is that the analysis be genuinely independent – which is often a difficult thing to do in government. It’s hard enough in any hierarchical organisation because those lower down doing the analysis will usually be doing it in a context in which they know that those higher up are well disposed to it. I think Buffett and some others in business basically ignore a lot of analysis on cost/benefit business cases for this reason – there’s just way too much wiggle room to get the answer the boss wants (or even worse that the underlings thinkthey want. In government a lot of the CBA requirements we have are not at arms length.
The next problem in government is that agencies acquire a culture which predisposes them to certain kinds of answers.
Still, not nay-saying just commenting. Perhaps my bottom line is that I wish people took some of these ideas more seriously by going to the trouble of really trying to think through the kind of institutional context necessary to get what we’re after. And then we’d have to try to evolve the institutions. Even if you concede the importance of arms length CBA, you’re still not out of the woods as heavy adversarialism between public agencies usually ends up in some dysfunction.
Fair point RIS Meister re. container refund scheme, and a similar point applies to the plastic bag ban I suspect. Perhaps if we had an upper house in Qld and a stronger local media then possibly more attention would be paid to the need for CBA and CBA results. Thanks for the comments.
Thanks again Nicholas. Re. heavy adversarialism among public agencies, I’d say that supports a case for stronger central agencies, especially a better resourced Treasury which could effectively analyse all the proposals being advanced by agencies. One of the problems I found was that, due to the huge volume of policy proposals in the budget process, it was challenging to do more than a superficial analysis of proposals, reiterating the standard Treasury lines on different issues. We’d also need better enforcement of Cabinet rules to ensure central agencies get sufficient time to analyse and brief on proposals.
Thanks for the review Gene
I’m more jaundiced about this – and find the absence of examples telling
Of course, if it can be done to shed insight cost-benefit analysis is a sensible part of policy. But far too often it’s done because some economist who’s not particularly close to how things operate says it would be a good idea and there’s an edict from on high – like Ronald Reagans – mandating it.
We mandate it in regulation review and it’s a farce – a box-ticking exercise. It’s nice to see Sunstein talk about measure and react, but that’s a partial insight which takes one towards the conclusion that it’s measuring things in situ not from on high that makes the difference. That and developing independent expertise in situ – the expertise of a teacher or nurse. Now it’s true that this ‘expertise’ is a hard thing to judge – it certainly can’t be vouchsafed by the fact that someone has a piece of paper from a uni. But the more I’ve thought about this, the more it’s jumped out at me that economists think of knowledge as knowledge that is amenable to them at the top of the system. A lot of knowledge isn’t like that.
Toyota didn’t do cost benefit analysis to identify the way in which they got workers on the line to understand their own productivity and endlessly optimise it. My proposal for an Evaluator General is an attempt to try to get the incentives right so that that kind of thing can happen and we can get information flowing from out in the field and counting for something higher up the system – which it doesn’t these days.
There was a range of critiques of CBA in the 70s and 80s by people like Charles Lindblom. The basic message wasn’t that they were a bad idea in principle – obviously, they’re not – but that they tend to be invitations to hubris as those doing the studies get all kinds of things wrong. The solution isn’t that they shouldn’t try, but rather that there are lots of other means of trying to get good value for money and protect against boondoggles which don’t rely on what an economist learned at uni.
Thanks Nicholas. I’m not familiar with Lindblom’s work but will have a look at it. Re. CBA being a box ticking exercise, unfortunately too often it is, but I’d like to think it provides some constraint on over-regulation and boondoggles. That’s difficult to prove, though.