The Merits of Merit Pay
School systems want to compensate on "merit". But are government bureaucracies actually capable of measuring such a thing?
Last week, I defended bad merit pay schemes as perhaps the best that hyper-bureaucratized government school systems could do.
Jay Greene, however, argues that what he calls "phony merit pay" really doesn't do much good: Let me explain the difference between true and phony merit pay. True merit pay—the kind of compensation for job performance found in most industries—provides effective employees with continued employment and regular raises while ineffective workers lose their jobs. If you do a good job you get to keep getting a pay check and if you don’t you have to look for work somewhere else. That’s true payment for merit because un-meritorious workers stop getting paid altogether.
In phony merit pay — the kind that hardly exists in any industry — there is a mechanistic calculation of performance that determines the size of a small bonus that is provided in addition to a base salary that is essentially guaranteed regardless of performance. You can stink and still keep your job and pay. The worst that can happen is you miss out on some or all of a modest bonus. To make it even more phony, in the few cases where this kind of phony merit pay has been tried, the game is often rigged so that virtually all employees are deemed meritorious and get at least some of the bonus.
It isn't that these schemes aren't tried in private firms--they frequently are. Complicated performance metrics for sales staff are a perennial favorite. The problem is that, as Jim Manzi once pointed out to me, when you implement one of these things, suddenly, your sales staff is "like Aristotle". They've got sophisticated reasons why your regression is all wrong.
Jay Greene makes the same forceful argument in favor of decentralization that Noah Millman did—in terms of more vouchers and charters. That's a great idea. And hopefully, the recent strike won't slow progress towards that goal.
On a somewhat related note, I see that Freddie DeBoer is complaining that we don't pay teachers well because we don't value them:
During the arguments about teacher unions this past week, I made a simple point: we say we value education, but we don't pay teachers nearly as well as jobs that do nothing of benefit to society. A rejoinder I heard constantly was "see, the difference between a public school teacher and Well-Compensated Useless Job X is that taxes pay for the former. That's why people want to fight to reduce teacher compensation." And I really think that's indicative of a certain shortsightedness that comes from constantly reading and thinking the types of arguments that are featured on blogs. Yes: in a deracinated kind of way, that argument makes sense. But in the larger perspective, it doesn't matter why we don't want to pay teachers. In capitalism, what is valued is what is paid for. As long as teachers make less than those in their same educational cohort, teaching will be correctly perceived as a low paying occupation. The genesis of opposition is irrelevant. And that's what really matters for what we were talking about, what the consequences are for our society.
We do, in fact, pay teachers pretty well relative to their level of education and hours worked. But that's a side point. The real point I wanted to make is that you hear this sort of thing all the time, and it misses a fundamental point about markets: it treats them as if the demand-side were all that mattered. In this mental model, the supply of workers is fixed, so the only thing that matters is how much we want them; the more we want them, the higher we bid their wages.
But of course, there is another side to the market: supply. And the reason that teachers are not paid as well, as, say, chartered financial analysts is that it is considerably easier to meet the requirements for being a teacher than to pass the CFA exam. It's no accident that the professions which require a lot of math pay more than the professions that are based more around reading and writing. Nor that math and science are the hardest positions to staff at most high schools.
But the majority of positions in high schools, or primary schools, are not staffed by math and science majors. They're staffed by humanities or education majors, who are relatively plentiful, and also, in less demand elsewhere in the market. With a glut of candidates, schools don't have to pay that well, and they don't.
This is not a measure of their value as people, nor of their value to society, nor even of whether society recognizes that value. It's just a measure of the fact that you can find a lot of people who are willing to do that job.
This is, of course, precisely the feature of the labor market against which the left chafes: that the market often sets wages below what we feel the value of the work is, or below what we feel the intrisic value of human labor demands. And I certainly understand why the left resists this--indeed, I too think we should resist it, when it takes the odious form of thinking that someone's earnings are a measure of their worth. On the other hand, I think that assigning pay based on supply and demand, rather than "the essential dignity of human labor" is simply the way that the market has to work. You can tinker at the edges--I don't think our current minimum wages are destroying the economy or anything--but you cannot simply wade in and assign by fiat wages that are more in line with what you think people should earn, in some idealized world where salaries are handed out by God as a reward for exemplary moral performance.
In some sense, this is the problem with the numerical compensation schemes: the assumption that there are a couple of key variables which one can precisely measure and then turn into an accurate assessment of the "value added" by a teacher. Teacher evaluations will never be perfectly accurate, and since there are many dimensions to good (or bad) teaching, they will always be somewhat subjective. The question is whether the government as we've currently constituted it is capable of designing a system which takes those subtleties into account.