29 October 2010

Experts, compassion, innovation: "They build castles of words, and call it knowledge"

Megan McArdle linked to what was indeed an outstanding post by Jim Manzi on technocracy and elites. Go read what he had to say. Here's his conclusion:
The essential Progressive belief that Klein expresses in undiluted form is that crafting public policy through legislation is a topic for which, in simplified terms, the benefits of expertise outweigh the benefits of popular contention. Stated more cautiously, this would be the belief that the institutional rules of the game should be more heavily tilted toward expert opinion on many important topics than they are in the U.S. today.

This would be a lot more compelling if the elites didn’t have such a terrible track record of producing social interventions that work.

An aeronautical engineer can predict reliably that “If you design a wing like this, then this plane will be airworthy, but if you design it like that, then it will never get in the air.” If you were to build a bunch of airplanes according to each set of specifications, you would discover that he or she is almost always right. This is actual expertise. I’ve tried to point out many times that the vast majority of program interventions fail when subjected to replicated, randomized testing.

Our so-called experts in public policy talk a good game, but in the end are no experts at all. They build castles of words, and call it knowledge.
I broadly agree with McArdle's take on Manzi, but think she misses the mark a couple of times.  Here's one:
Moreover, like all elites, the current meritocratic class is self-interested in numerous ways. It is easy for us* to recommend free trade, carbon taxes, and so forth; most of us live in cities where we don't have to drive that far, and/or command incomes that make the price of gas rather incidental to our budget. And think tanks, policy magazines, and congressional staffs--however threatened they may be by other forces--are not yet likely to be outsourced.

I'm not saying that these ideas are wrong; I myself support both of them. But I am also aware that I do not really emotionally comprehend what it is like to be trying to support a family of four on $38,000 a year in rural West Virginia.
I disagree that the problem is insufficient compassion or emotional comprehension of other people's circumstances. I think it's purely a knowledge problem.

Even if you're overflowing with compassion, even if your emotional comprehension of a situation is perfectly tuned, there's no way for people to understand the wildly complicated system of interactions that has led to that situation, and which you would need to understand to effectively modify it.

In fact I think a surplus of emotional attachment can be as damaging as a lack of it in elites. Surely we can't be too compassionate for people who have had adverse reactions to drugs, right? Well we can be if that compassion causes the FDA to delay drugs longer than necessary, harming people who might have benefitted. Where's the compassion for those people?  Even the very caring can't be compassionate about the ne voit pas, unknown suffering. Or how about all the compassion for people whose rent is too damn high? If that compassion leads us to counter-productive rent control policies we end up hurting the people whose plight we were so compassionate about originally.
The other reason I don't necessarily trust elites is that they really like thinking big. You don't get hundreds or thousands of people into a vociferous debate over making some modest improvement to Medicaid reimbursements; you get them animated by proposing a radical overhaul of the health care system. Yet most innovation isn't big; it's continuous, incremental improvement. Companies are forced to this by market discipline, but we don't draw that many policy people from business; they're viewed as tainted by the commercial association.

So we get what most interests wordsmiths: a succession of enormous plans (health care exchanges! privatize social security!), most of which fail. We get very few mechanisms to improve them. I'm not even sure the government can work on that sort of improvement--we've stripped so much autonomy from officials and civil servants in the name of equity that few people have any authority or will to push through reforms.
Quite right, innovation is almost always incremental. Revolutionary advances are shockingly rare.

The problem isn't that the chattering class' penchant for Big Ideas is displacing the little shuffles forward that we need, it's that there's little reason to innovate in any way, big or small.  There's no competition, or other incentive to innovate. Incremental innovation isn't just something business do. You need to right incentives for it to happen. There's whole shelves of journal articles about setting up the proper environments to make this happen: rewards and bonuses and tournaments and voting systems and so on. (Something I'm sure McArdle is aware of.)  Almost all of these systems are about harnessing and amplifying "popular contention" — competition — exactly the thing that Manzi identifies as the thing progressives would like to minimize in favor of expertise.

Furthermore, incremental innovation isn't just like revolutionary innovation, except more gradual.  It's also a lot messier.  It requires trying a lot of things and getting most of them wrong.  Governments aren't good at that.  (Monolithic enterprises, even private ones, stagnate for this reason.)  The states are no longer, if indeed they ever were, the "laboratories of democracy." Washington can force states, and municipalities, and citizens, to do any damn thing they want at this point.  They're all about picking the One Right Choice, and that's the antithesis of what you want for innovation.  But that's all right for the elite, because they're the ones that get to make that one choice.

~ ~ ~
The fact is that up to now a free society has not been good for the intellectual. It has neither accorded him a superior status to sustain his confidence nor made it easy for him to acquire an unquestioned sense of social usefulness. For he derives his sense of usefulness mainly from directing, instructing, and planning — from minding other people's business — and is bound to feel superfluous and neglected where people believe themselves competent to manage individual and communal affairs, and are impatient of supervision and regulation. A free society is as much a threat to the intellectual's sense of worth as an automated economy is to the workingman's sense of worth. Any social order that can function with a minimum of leadership will be anathema to the intellectual.

— Eric Hoffer


  1. I'd say it isn't a knowledge problem at all, it's an arrogance problem. The supposedly "unintended" consequences of legislation are generally obvious to everybody long before it is signed into law. No, for those who push that legislation anyway said consequences were either intended to occur, or were deemed acceptable collateral damage in service to their vision of utopia.

  2. I think there's a fair amount of arrogance on display, especially with the simple and easy to predict "unintended" consequences of the things I mentioned like rent control. But even if you were to take arrogance out of the equation, you still have a knowledge problem. I'm interested in not just the things that elites don't know, but things they can't know, no matter how wise and benevolent and humble they might be.