Sunday, September 18, 2011

the worst thing that can happen is actually pretty bad

This is wrong, not just for the methodological reasons given. The problem is more serious. What's so great about optimal decisions after all? Absolute optimality has costs.

Specifically, even if consultation doesn't help you achieve an optimal decision, it may help avoid a decision that is dramatically pessimal for some particular person or group of people. That's worth something. Economists, especially, ought to remember this as the whole idea of Pareto efficiency comes with Pareto's further point that a step towards optimal efficiency in any given market is not necessarily a step towards perfection overall.

Further, in practice, people who refuse to consult in the hope of making an optimal decision are quite likely to be rationalising the avoidance of information they don't like or just their own arrogance. In that case, they're very likely to make a really bad decision. It's better to miss the cost-index speed by a couple of knots than stall the plane into the sea. Programmers are taught to remember that premature optimisation is the root of all evil.

A couple of points in the upshot. Do we lose more from not attaining perfection than we do from horrible bungling? For a practical example, are all those superbly conceptualised deadweight losses and x-inefficiencies that are meant to be in the economy in the event of any state intervention whatsoever anywhere near enough to match the lasting losses of financial crisis? Enormous efforts are made to squeeze the long-term unemployed back into the job market. But the numbers are clear - people lose work in recessions and then don't find it again.

Wouldn't it be better to hold the line, putting up with one parmesan shaving less on my martini in the good times, rather than waste time and money and bother a lot of poor bastards trying to fix it later? Kurzarbeit beats the piss out of A4E.

More generally, loss-aversion is often considered to be a cognitive bias. One of the things the recurring drama of the rogue trader tells us is that someone who tells you it's only a little loss and it's worth it for the potential upside is very often lying or deluded. Isn't loss aversion actually quite a sensible heuristic for most purposes? After all, many of the things it would warn you against are actually sure losers, like gambling, smoking, and voting Conservative.

That gets me to the final point. How dare anyone have the intellectual dishonesty to argue against social democracy on the grounds that stuff happens and we need safety first and to respect individual variety and old institutions and no plan survives contact with the enemy? Avoiding disasters and silly experiments is our thing, and if that's conservatism, well, vote Labour.

Of course, my own experiment with voting Lib Dem is exhibit A.

No comments:

kostenloser Counter