Matt Yglesias dispenses with the notion that liberals aren't allowed to favor merit pay for teachers because this is "teacher bashing".

. . . the implication that the idea that pay should be differentiated based on effectiveness constitutes "teacher-bashing" is bizarre. When it comes to compensation, it seems to me that there's an easy way to distinguish between people who have a favorable attitude toward teachers and people who have a negative attitude toward teachers. If I were interested in "teacher-bashing" I would think our society should dedicate a smaller quantity of aggregate resources toward paying teachers. In fact, I think we should dedicate a larger quantity of resources toward paying teachers. That's because I think education is important and evidence suggests that teacher quality is among the biggest non-demographic factors in determining student achievement. Under the circumstances, it makes sense to invest a lot of money in hiring and retaining teachers.

That said, once we've hit upon a given pot of money to spend on teacher compensation, a question arises of how it should be divided up. One way to divide it up would be evenly--each teacher could make the same salary. That would, however, be a bit weird and we don't do it that way. Instead, we pay teachers more the more experience they have, and we also pay them more when the acquire master's degrees. As I said yesterday, I think the only way to make sense of these forms of differentiated pay is that they're already a system of "merit pay." The point of paying higher salaries to people with advanced degrees has to be the belief that teachers with advanced degrees are more effective than teachers without advanced degree. It turns out to be the case, however, that research says this is wrong. I don't think it's "pro-teacher" to be giving teachers financial incentives to essentially waste their time acquiring advanced degrees that don't help them. This is simply an irrational way of divvying up the compensation pot.

It makes a lot more sense to take that money and try to spread it around in ways that better track actual teacher effectiveness. One objection to this is that it's hard to do really well. And, indeed, it's not easy to do perfectly. At the same time, the fact that creating a better system would be difficult isn't a very good reason to stick with a system that definitely doesn't work. Either way, I don't really like the term "merit pay" which I think is silly. Among other things, as I've said nobody I'm aware of actually believes in paying teachers on a flat salary schedule so the whole idea is a red herring. Paying more for more experienced teachers makes sense, but currently we seem to be giving more weight to seniority than it deserves. Paying more for extra degrees makes no sense. Paying more for people with in-demand technical skills makes sense. Paying more for people who take on more challenging assignments in high-poverty classrooms makes sense. And trying harder to directly measure and reward effectiveness also makes sense. But if I'm "bashing" anyone it's purveyors of useless M. Ed. degrees.

This is one of those odd areas where Matt and I are in total agreement.  We should pay teachers much more than we do.  Right now, they take a substantial portion of their "pay" in the form of near-total job security.  People like this benefit.  But in most cases, they shouldn't have it, because it has predictible effects on performance--particularly when it is coupled with a pay scale that relies on measurable but not very useful traits like advanced degrees (totally useless) and seniority (the benefits of experience eventually level off).  The only thing teachers have a financial incentive to do under this system is keep their butts in the teacher's chair, and acquire useless degrees from programs that mostly teach students how to sit through long and pointless classes.

The obvious thing to do is to strip the protections and up the pay, while using merit metrics to determine how that pay is allocated.  But the union has very good reasons to resist this.  For one thing, depending how you implement it, you'll substantially reduce the role that the union has in setting salaries, and thus its value to the membership.  For another, more than 50% of their membership are, definitionally, average or below-average.  Merit pay is probably not a good deal for them.  Especially if they've spent valuable years of their lives acquiring useless M. Ed. degrees.

On a life-cycle basis, merit pay is only good for the minority of teachers who can produce outstanding results early and often.  The rest used to have the comfort of knowing that they would eventually get to the top if they just ground away long enough.  Hopefully, we can overcome this if we throw in enough money to sweeten the deal--as we should, anyway, if we want to attract great teachers.  But it's a grinding battle everywhere it's been fought.