The latest issue of the New Republic has a semi-defense of No Child Left Behind that I more or less agree with, but let's leave that aside now. In the article, Phillip Gordon argues that Democrats should start touting merit-based pay for teachers, something Kerry threw his weight behind during the campaign last year, and something I've endorsed in the past without thinking much about it. But there comes a time for actually thinking about things, I guess, so let's go. The reasons for doubting the magic and sparkly effects of merit-based pay might be familiar to many people, but not to me, so bear with a bit of blundering here.

The most interesting study I've read on the subject was an old one, done on 20 Social Security Administration offices that instituted merit-based pay, and found essentially no improvement in office performance—and this was despite the fact that in this case the measures of worker productivity were easily measured (i.e. accuracy of claims processing, time it took to settle claims). Other research seems to be a bit more ambivalent. Some more digging around, though, unearthed a pretty persuasive 1998 article by Jeffrey Pfeffer for Harvard Business Review (not online, sorry), which argues against merit-based pay, citing a study by consulting firm William M. Mercer. The results: 73 percent of surveyed companies in the preceding two years had tried to find way to tie pay to performance. 47 percent reported their employees found the system unfair and not at all sensible, and 51 percent said it added little value to the company. The Mercer study concluded that most merit-based pay did two things: "absorb vast amounts of management time and resources, and they make everybody unhappy." Hmmm.

Now I don't really mind if merit-based pay is "arbitrary" or "unfair" if it has a positive effect. But it doesn't always seem to. And the experience of businesses would very likely apply to merit-based pay in school, no? Not to mention the fear that teachers whose pay depends on boosting test scores, for instance, might well try to "cheat" on their students' exams. (Steven Levitt has showed that cheating of this sort does in fact happen.) It doesn't make sense to run schools more like businesses if the business model in question is, in fact, flawed.

The alternative, as detailed for instance here, is some sort of group-oriented incentive system where all employees share in the rewards of success. That sounds touchy-feely, but it might well work. Pfeffer cites a study of a manufacturer that put in this sort of compensation system: "grievances decreased, product quality increased tenfold, and perceptions of teamwork and concern for performance all improved." Nice! I don't think this is the sort of thing that would be impossible to put in place for schools, although I'm not sure what the specifics would be: perhaps reward all teachers in a given grade if their students improved collectively based on some metric, like value-added tests? Maybe. Meanwhile, the economists among us might worry about free riders, but my understanding of behavioral research is that free riders in this sort of situation are fairly rare; as mom always warned, peer pressure can be a hell of a thing.

So merit-based pay: not all it's cracked up to be? Seems that way. One crucial caveat though: In places where such systems have actually been put in place, merit-based pay for teachers ends up effectively being nothing more than an across-the-board pay hike. In one Denver pilot program, 85 percent of teachers met their objectives and received a raise. This appears to be the case in Britain as well. So it might be that, in practice, merit-based pay isn't actually merit-based pay so much as a much-needed raise for teachers that is wrapped in the sort of packaging that helps parents feel good about spending the money. Well, okay. I'm all for raising teacher salaries—by a lot—and if this "performance-based" talk is what people need to swallow the accompanying property tax hikes, well then bring it on. But actual merit-based pay? Probably not. Now giving schools more flexibility to fire ineffective teachers on the other hand...