Thursday, 6 November 2008

The Wrong Side of the Public Good Trap

"I want you to imagine that it's about 500 years ago, and you are one of a line of men with spears, and the spears are pointed that way. And the reason they're pointed that way is that coming at you is a line of men on horseback, and they've got spears too. You do a very rapid cost-benefit calculation... Given what everybody else is doing, what should I do? If everybody else stands, my best chance is to run; there are 3000 of us in this line, and whether I run has a tiny effect on whether the charge will succeed. On the other hand, if they're going to run, my only hope is to run faster... And so, all of us in the line rapidly do this calculation, we all run and most of us die."

"Welcome to the dark side of rationality. That's market failure."

So began a delightfully eclectic lecture on 'Market Failure: the case for and against government,' by David Friedman, libertarian heavyweight and Professor of Law at Santa Clara University.

[Update 2008-11-12: The talk is now available as a video.]

But back down to earth. Starting with a definition of market failure - in which either some goods are underproduced, since their full benefit cannot be privatised (the public goods problem - eg street lighting), or overproduced, since their full cost is not borne by the producer (the negative externality problem - eg pollution) - Professor Friedman illustrated how many such problems, theoretically posed as insoluble, had in fact been resolved by private agreement and clear definitions of property rights. The underproduction of television and radio programs, due to the inability to restrict who receives the signals, had been mitigated by the combination of advertising and content. Meade's famous case of beekeepers, who profit from the proximity to agricultural land but who cannot be made to bear the cost thereof, is found to be resolved by contractual agreements between farmers and beekeepers. Even more famously, the case of the lighthouse, which had been universally thought an archetypical public good requiring government intervention from Mill to Samuelson, was shown by Ronald Coase to have historically funded in Britain by 'light dues' paid by ship owners, not out of general taxation.

These imperfect solutions certainly pale into comparison with the possibilities of a "world run by an all-powerful, benevolent and all-knowing regulator" - but that isn't the alternative. Indeed, the alternatives to laissez-faire are beset by exactly the same type of problems as free markets are, but suffer from them to an even greater extent. On the political market, the cost of investigating all the issues in an election, discovering the various candidates' policies, weighing up the relative merits and flaws in each, and then queueing up to vote has a substantial cost. Its concomitant reward? "To increase by one the number of people voting for the good guy." There's an excellent video (which I thought was a parody, but turns out is in fact a genuine attempt to rally the vote for Obama - though the customization is pretty cool) which propagates the ridiculous notion that one vote can be the deciding factor in an election.

Rather, "it's much more fun to say, 'this guy is handsome, I like him,' or 'this person is saying things that make me feel good' and vote on that basis." How apposite. This problem of Rational Ignorance, where the cost of acquiring information exceeds the expected benefit from it, stands in the way of an efficient outcome on the political markets. Indeed - it exacerbates an inefficient outcome.

Another problem in the political solution to market failure comes from the distorted payoffs for government employees, and the corresponding sub-optimal levels of risk taking in government regulated fields. Professor Friedman illustrates this by considering the incentives structure of a government regulator:
"Suppose you are working for the FDA, the regulatory organisation that decides whether it's legal to introduce a new medical drug onto the market. You say to yourself, 'There is a new medical drug. We don't yet know for sure if it's safe. If it's safe, it'll safe quite a lot of lives. If it's unsafe it'll do bad things. Should I approve it or should I either reject it or insist on further testing?' And you say to yourself, 'If I reject it - and it was safe - nobody will ever know.' Because, after all, if it's rejected, we won't discover if it's dangerous, and while it's true that rejecting it will kill 1,000 people per year, that's a statistical 1,000 people. There won't be any person I can point at and say - 'he died because of that.' Typically, the new drug means that instead of curing 52% we now cure 55%. Nobody knows who the 3% would have been cured. On the other hand, what happens if I approve the drug and it turns out to have serious bad effects. The answer is, my career is over. Can you say thalidomide?
The rational individual decision to 'play safe' by regulators therefore causes them to persistently err, without the proper feedback mechanism - the opportunity to profit - to incentivise making correct decisions.

A third instance of government failure can be seen by considering the parallel of a popular concern about the free market - monopoly and market dominance. Free trade, it is acknowledged near-universally amongst economists, is almost always beneficial to the country that practises it. Despite this, there are only two periods of substantially free trade in history: post-Corn-Laws 19th century Britain and post-war Hong Kong. This is explained by the concentration of special interests - agricultural lobbyists, automotive groups, trade unions &c - who can exploit the political machinery to use coercion against their competitors, from which they reap the majority of the benefit. In contrast, the cost of these policies is dispersed amongst a general public, each member of which pays fractionally more for goods than would be the case under free trade. Each additional cost being small, it isn't worth the individual effort to oppose it. But the combined cost of a tariff across the entire population is substantial, raising prices domestically and weakening exporting sectors of the economy, which have fewer customers due to less trade between domestic citizens and foreigners in the protected goods. So,
"when people say, 'an auto tariff is needed to protect American workers,' I say, 'Yes, it's needed to protect American workers from American farmers.' So the effect of an auto tariffs is that in situations where it's cheaper to grow cars [by exporting agricultural products and importing cars with the income] than to build them, we still build them."
It is the dispersion of costs that the political process permits - far wider than most externalities of private exchanges - that explains a remarkable paradox observed by Gary Becker: in countries with small agricultural populations, farmers are greatly subsidised; in countries with large agricultural populations, they are heavily taxed. That the farmers are a concentrated interest in the former case, and that the volatile urban masses, who benefit from redistribution from dispersed farmers, are concentrated in the latter case is the most compelling explanation of the extraordinary negative correlation between the number of farmers and the degree of agricultural subsidy.

Professor Friedman concludes with a clear and compelling statement:
"Market failure indeed exists on ordinary private markets, and prevents them from reaching the ideal result; but it not only exists, but is endemic, on political markets. Market failure is really coming because I'm taking an action in which a sizeable part of the cost or benefit is going to other people. If you think about the markets on which political decisions are made - the market where we elect politicians, or the market where politicians trade favours and such - almost every case, the people making the decisions are bearing a negligible fraction of the cost or receiving a negligible fraction of the benefit... If what you're considering is, 'Should the production of food, automobiles, schooling or anything else, be handled by a private market or by the government?,' the logic of this argument is that market failure will result in both doing a less than perfect job, but it is likely to have much more serious problems for the political option."

Following the main talk, there is a wide-ranging questions session, taking in topics as diverse as pirate democracy, life extension, nanotechnology, selling citizenships, private law enforcement, Saga-period Iceland, direct democracy, the psychology of voting, Sarah Palin, prayer and Usenet (to the latter of which Professor Friedman is a prolific contributor).

One particular remark from the questions is memorable for its sense of proportion:
"As catastrophes go, global warming is a pretty wimpy catastrophe. Global warming, on current estimates, in 100 years, is going to raise sea levels by a foot or two, and raise the temperature of the earth by about 3 degrees centigrade... That could be bad for some people - it might even be bad on net, we don't really know. It could be good for some people, certainly bad for some. [gestures to his new book, Future Imperfect] I've got three different technologies that could wipe out our species in less than 100 years."


_ said...

Thanks for posting this!

Seienchan said...

Wow, I'd forgotten how enjoyable this talk was! Should have got our conversation on Objectivism beforehand though; that was fun.