Mark Kleiman (hat tip Brad DeLong) says more clearly what I tried to say a while back: cost-benefit analysis of regulations has a curious way of nailng the costs and underestimating the benefits. He focuses on three points:
- Traditional CBA counts all dollar benefits equally, despite the fact that the marginal utility of a dollar depends a lot on who is getting it; a dollar more for a poor person provides a lot more utility than a dollar more for a rich person.
- Long-term or uncertain benefits, no matter how large (like preventing the inundation of every coastal city) are typically discounted down to zero.
- Benefits that are difficult to quantify because there is no market for them (like feeling better because you are healthy) never get counted. (This is the one I know best because it’s one of the things my wife specializes in.)
Matt Yglesias also comments.
By James Kwak
The Baseline Scenario 
Also, the extent to which CBA is allegedly so necessary seems in direct proportion to how rational the policy is.
Thus environmental policies are expected to measure up in the face of the most meticulous CBA, and it’s allegely of paramount importance that we specify how we’re going to pay for health care reform.
But bailouts, wars, weapons spending, and most of all tax cuts are never supposed to be subject to even the most rudimentary analysis.
This is precisely because no one seriously argues that those have any reality-based rationale. On the contrary the very fact that they are brazenly irrational and ideological is what exempts them from cost-benefit oversight or from anyone ever having to explain how we’re supposed to pay for them.
So whenever I hear anyone say “how are we supposed to pay for” e.g. single-payer, my inclination is to reply “You tell me how we’re going to pay for bailouts, wars, weapons, and tax cuts for the rich, and I’ll tell you how we’re paying for anything else”.
In addition to those ‘costs’ such as war-making, tax cuts,’ etc mentioned above that are never mentioned as costs -in part because economists don’t like to acknowledge the existence of the state and raison d’etat (ideology), there are a broad range of phenomena where neither costs nor potential benefits get counted at all. Nancy Folbre noted that caring work is never properly acknowledged (in who pays for the kids) while everyone bemoans the lack of parental involvement in childrens’ lives that costs societies a lot (i.e. education, values, proper nutrition, etc).
Instead you get stuff like Becker’s economic analysis of the family!
More broadly, CBA like conceptions of efficiency, is inherently ideological but the values are hidden away in the a priori assumptions.
‘Inherently ideological’ like everything else about conventional economics, which is why it is time we began exposing it rather than awarding Nobel prizes for it.
On number 1, yes it sure sounds like a splendid idea, and I have often dwelled on it myself, that is until I face the other side of the equations meaning that the marginal cost of the resources used are not the same either because a tax dollar collected from a millionaire cannot be equivalent to a tax dollar collected from a poor person (like through sales taxes) and so if I use both adjustments we might not really be evaluating anything but just ordering a direct distribution of income. And, if that is what is wanted, there are more transparent goabouts.
On number 2 and 3 that is why we have politicians, academician, civil society and others to try to pressure for the considerations that might not be so easily to quantify and, if you want to give the long term a better chance you might instead go the route of trying to empower democracy and future by the One-Child-One-Vote reform.
“so…. I’m thinking that they need to hold off a while on this regulatory stuff, just until the economy gets on a little more solid footing, what are they, socialists or something? Obama, ha!”
Jim Cramer CNBC 08/25/2009
It is interesting to hear this commentary after just being involved in a recent project to reevaluate BCA guidelines for federal agencies (flood plain management agencies in particular). I can say that there is a significant amount of buzz about BCA in Washington, as well as what seems to be a well-intentioned effort to improve current guidelines.
However, bureaucratic inertia is the primary barrier to implementing more modern BCA considerations like those outlined in point 1. BCA has struggled with the distributional effects of costs and benefits for a long time, yet appointed officials still aren’t ready to consider these. Gotta keep up the pressure.
A similar question arises in the context of computer security. Having done a threat analysis about what could possibly go wrong, you can then rank the risks that might cost you the most harm. You’re probably wasting time trying to put a dollar value on the harm.
So rather than trying to compute costs, you want to generate a relative measure of the costs of a solution that achieve the same benefit, even if those benefits cannot themselves be quantified.
You can then pick the most effective solution, without ever identifying the monetary value of the security issue.