Thursday, 7 November 2013

Stuff people want

Economists would probably agree that public decisions should be guided by open, contestable and rigorous analysis. Most would also sign up the view that we should weigh up the costs and benefits of decisions and act accordingly.
But what counts as a benefit? As a first cut, we might talk about an attempt to count everything that makes people happy. Cost-benefit analysis is normally based on the utilitarian view that we should try to add up and attach values to all of the things that people like or dislike about a potential change.

That works pretty well in market settings where we have a reasonable shot at estimating demand curves and supply costs. There is also a big literature on non-market valuation, which can be used for where there are no trading data - to estimate things like the value of open space or clean air, or the cost of road congestion.

But there are some big holes in the utilitarian approach. One is that it has pretty weak microfoundations because pleasure and pain are polymorphous. As MacIntrye put it
"the pleasure-of-drinking-Guinness is not the pleasure-of-swimming-at-Crane's-Beach and the drinking and swimming are not two different means of achieving the same end-state....consequently, appeal to the criterion of pleasure will not tell me whether to drink or swim"
Another is that it could deny basic rights that most people agree on, like the freedom to choose your life partner. There was an emotive campaign in New Zealand this year opposing gay marriage. On a strict utilitarian view, the distress suffered by those who hated the idea would count against permitting gay marriage. But if you consider that life-partner choice is a basic right, you'd classify the opponents as having what Sen (pdf) called "meddlesome preferences" and could argue that they should not get any weight at all.

I haven't yet had to deal with these issues in a cost-benefit analysis, but they're definitely out there, lying in wait for unsuspecting economists.

1 comment:

  1. Brad Taylor and I had a go at some of this a few years back... need to get back to that paper sometime and send it out.

    There's a libertarian anarchist literature around whether it's possible to have a functioning property-rights respecting market order without the state. Assume it's possible for now (David Friedman; Caplan & Stringham, others). Would that state be appreciably more liberal? Harder question.

    Suppose that democracy's efficient in providing what the median voter wants and that market-chosen law would instead approximate mean voter preferences (intensity-weighted by willingness to pay). In that case, widely but weakly held meddlesome preferences impose illiberalism under democracy (relative to market-chosen law) while narrowly but intensely held meddlesome preferences impose illiberalism under market-chosen law (relative to democracy).