In a valiant attempt to defuse the ideological conflicts between the reformist and traditionalist wings of the liberal education wonketariat, Matthew Yglesias argues that this disagreement is not not ideological at all. Rather, it is an artifact of past decisions about educational structure:
Take, for example, the hot issue of teacher compensation. The traditionalist view is that teachers should get paid more for having more years of experience and also for having more degrees. The reform view is that teachers should get paid more for having demonstrated efficacy in raising student test scores. This is an important debate, but I think it's really not an ideological debate at all. I think the only reason it's taken on an ideological air is that unions have a view on the matter and people do have ideological opinions about unions in general. But if we found a place where for decades teachers had been paid based on demonstrated efficacy in raising student test scores, then veteran teachers and union leaders would probably be people who liked that system and didn't want to change to a degree-based system. Because unions are controversial, this would take on a certain left-right ideological atmosphere but it's all very contingent.
This is a very interesting thesis, but ultimately I think it's wrong. There is a reason that unions kill merit pay, and it's not because they just happened to solidify in an era when merit pay was out of fashion.
To state the obvious, unions negotiate ironclad contracts to cover dozens, hundreds, or thousands of workers. Once they take effect, those contracts are rarely renegotiated, and they apply to every single worker no matter what the situation. So unions are always going to be looking for the simplest, least subjective metrics by which to measure their members. Furthermore, they will be looking for metrics which are not under the control of the other side. The school board cannot change how many years you have in service, or whether or not you have a degree. But it can change the curriculum, or the tests.
Obviously, people who are not in unions write employment contracts, which are similarly hard to write. But non-union employment contracts operate in an environment where both sides often hope to continue the relationship beyond the initial term. This offers quite a bit of good-faith flexibility, because people who are too rigid about the exact letter of their contracts are apt to find that their contract isn't renewed. Even in contracts with a very definite term, there are reputational considerations. That's just not how unions operate, because the union can't be fired by the employer. When the contract expires, you're going to negotiate another contract. The result is that people in non-union employment contracts can tolerate quite a bit more ambiguity on both sides than people in a collective bargaining situation.
The unhappy corollary of this is that the metrics will not only tend towards simplicity and ease of measurement; they will also tend to reward mediocrity. Again, this is not an accident of history. A collective bargaining unit run by a "majority rules" system is always going to look for a system that rewards the median or modal worker, not the best.
A merit pay system can work in one of two ways. It can benchmark teachers against the average, and reward the people who achieve the most improvement. Or it can set some minimum standard and give a bonus to any teacher who bests that standard. (You could set three tiers, or what have you, but the concept is basically the same).
In my opinion, the first system is probably going to best maximize productivity (though this is an interesting discussion for another blog post). But it would never pass a union vote, because the majority of teachers wouldn't benefit from it, and those who did would have to work harder. The second system might pass. But the union would make heroic efforts to water down the benchmarks until the majority of their members were receiving at least some "bonus" pay.
But compare either system to what now exists in our nation's schools. Every single teacher can stay on for years unless they do something direly wrong. Every single teacher can get a useless education degree, which basically requires a pulse. They have a system that spreads benefits absolutely evenly among all their members.
How would any alternative gather majority support from the union members? I mean, you can add on resistance to change, which I think is significant. But even if they were picking a new system from scratch, the seniority + degrees system is clearly going to satisfy many more members than either of the merit pay alternatives. It would probably be the majority choice no matter what. And of course, over time, teacher's unions select for the sort of people who prefer this arrangement to competitive merit pay for one reason or another
Unions are set up to minimize frictions and maximize benefits for the bottom 55%. That's how they work everywhere--in schools, and out. That's how they have to work. No amount of cajoling, no number of white papers, is going to change that.