Teachers and Incentives, Positive and Negative

I’m flattered that both E.D. Kain and Conor are interested in my opinion on this subject. Honestly, I probably should blog more about education myself. From my experience of six years serving on the board of a charter school, I have learned a great deal about what works – but also where the hype has run well-ahead of the reality. But the problem is, the more you know the more you know how hard it is to make definitive statements of any kind. And that’s not terribly satisfying to people looking for clear direction.

Conor highlights two questions: the importance of getting rid of bad teachers, and the importance of monetary incentives like merit pay. Ed reformers are generally in favor of both; the unions are generally opposed to both. For me, it’s a split decision.

I completely agree on the importance of getting rid of truly bad teachers – not teachers who are sub-par but teachers who are sub-basement awful, the teachers everybody knows have to be gotten rid of somehow, but who instead bounce around the system for years. What many people don’t fully appreciate is the degree to which the continued presence of these teachers undermines the entire school, with effects that radiate through the entire system. A single out-of-control classroom affects the behavior of kids in all the other classes they attend – they don’t just suddenly switch into good behavior mode once they get to their next period, or even to the next year. Administrators who know they can’t do anything about glaring, massive problems in a given class have very little incentive to try to do anything else – what’s the point? Teachers who enter idealistic and eager to make a difference get ground down by having to work alongside abusive or incompetent colleagues.

Giving principals much greater authority to hire and fire isn’t a complete solution. But it’s an absolutely necessary component to any solution. When a manager has no authority to hire or fire staff, he or she has no leverage to change staff behavior. And without leverage to change staff behavior, basically you don’t run the organization in any meaningful sense.

The main objection to the idea of giving more power to principals, though, is a good one, and it’s why that solution is incomplete. The incentives on principals, after all, don’t necessarily point in the direction of running good schools, and some principals are also terrible. Giving them more power will only make it easier for them to abuse their staff. They may fire the best teachers rather than the worst – the ones who challenge their authority rather than the ones who are short-changing students. Subjection to arbitrary authority may also be just as alienating to idealistic potential teachers as the impotence of authority to address incompetent, abusive and disruptive staff.

All of which points to the importance of getting the right people in leadership positions in schools, and evaluating them effectively. But this is where the difficulty with proper assessment really comes to the fore. It’s not that hard to identify really terrible teachers. It’s much harder to design a bureaucratic evaluation mechanism properly tuned to align incentives for leaders to run their schools correctly. A major effect of the kinds of school-evaluation tools that have already been deployed has been to encourage teaching to the test above all. That may actually be an improvement in some schools, and certainly helps identify schools that are completely failing to teach – but in many others, it has incentivized the leadership to destroy much of what made the schools effective in the first place.

I’m skeptical that any such bureaucratic evaluation system can really work. What I’d generally be more optimistic about is the use of consumer choice to evaluate schools. Theoretically, if you had universal public school choice (let’s leave private schools out for the time being; including them raises a whole collection of unrelated issues that I don’t want to tackle here), where public dollars followed the students and schools could not select their students but were given public resources to expand in response to persistent demand in excess of the number of slots available – in such a system, you’d expect poor schools to fail as students fled and strong schools to expand.

And, indeed, I think that is what you’d get if you implemented such a program. Under competitive pressure to attract students, I do believe the overall quality of education would gradually but persistently improve.

But you’d also get a side-effect of more student sorting, sorting that would reinforce existing disparities in similar ways to our current system (where we sort students by the price of their parents’ house) but with somewhat more sorting by cognitive ability and somewhat less sorting purely by class or race (though there’s going to be a lot of overlaps between those two sorts, inevitably). Why? Several reasons. More involved and informed parents will do more to effectively evaluate the choice of schools, while the least-involved and informed parents will apply late, or fail to apply at all and get stuck with the dregs. Schools, meanwhile, will structure their offerings to appeal to the students they want; while they can’t select their students, they can try to shape the applicant pool in various ways to make it more favorable. And they can always kick out students who don’t live up to the school’s standards.

For some, that’s a serious bug; for others, it’s more like a feature – good cognitive sorting will enable schools to more effectively challenge the next generation’s elite, which is probably more important than anything else we do to future national competitiveness. I think it’s both. I think it’s vital to challenge students at very level of natural ability. But I also think it’s vital not to leave any students behind in the sense of leaving them unable to function effectively in a modern society and economy. There are various possible ways to put a bit of a thumb on the scale to incentivize schools not to “fire” tough students – one already exists, in the form of more money for students with IEPs, but others could be devised. But that overall framework – greater parental choice within the public school system, with thumbs on the scale to try to mitigate the pernicious side of sorting – strikes me as the right way to get the incentives lined up, and much more correct that applying some kind of universal bureaucratic evaluation mechanism for schools.

But what about merit pay? Well, honestly, I’m skeptical based on what I’ve seen. Yes, teachers care about money – everybody cares about money. But when I say that nobody goes into teaching for the money, what I mean is that teachers, in general, have chosen a career that undercompensates them in cash terms relative to other professions with similar educational qualifications. So they have chosen to forego income for other rewards. Those might include: the desire to work with children; the desire to “make a difference”; the desire for a job that ends at 3pm and lots of vacation; the desire for a job with great benefits from which you basically can’t be fired; etc. If “more cash in my pocket” isn’t the main driver why people go into education, then cash rewards for great test scores probably isn’t the right way to incentivize teachers to perform.

On the other hand, I think it is vital to build positive incentives into the system. The negative incentives of “you could be fired” and “your school could be closed” are important – but they are not really motivators; they are mechanisms for eliminating the demotivators of terrible colleagues and terrible leaders. For teachers, though, I think the big positive recognition to offer is recognition. You want the system to recognize really excellent teachers and provide them opportunities to advance in the hierarchy if that is their desire. For monetary incentives, though, I think the focus should be elsewhere. I think extremely high performing schools should be rewarded with bonuses for the school itself. That creates a proper positive incentive for collective action to improve the school environment, which is what you want, not an incentive that pits teachers against each other. And I think there’s a case to be made for monetary incentives to students for exceptional performance, though there are very serious risks about going that route as well.

The main change I would make to the monetary incentives on teachers relate to the incentives to enter the profession. Right now, a great many of the monetary rewards to teaching are back-loaded. You get paid much more if you’re a 25-year veteran teacher than if you are a 5-year veteran. And while starting salaries are generally quite low, the benefits – which matter much more to older people than to younger – are generally very good. These incentives select for a teacher population that is looking for a very long-term gig. And I’m not sure that’s what we actually want the teacher population to look like. The learning curve for a teacher is very steep in the first few years, but after that it flattens out. After 3 years or so on the job, you’ve learned most of what you are going to learn about how to be a good teacher. After that, you’ll just keep doing it. Meanwhile, a big chunk of what it means to be a good teacher is to be good at keeping and holding a class’s attention. If you look, in particular, at Direct Instruction systems like Success for All that have gained currency in recent years, the kind of skills a teacher needs to perform well in such a classroom look more like the kinds of skills that make for a good actor or sales rep. You don’t need to be an academic genius; you need to be able to keep a group of easily-distracted people focused on you and on the material you are walking them through. Those skills, in turn, are going to be fresher and sharper in relatively new teachers than they are in teachers who have been doing the same old slog for 25 years.

It is folly to try to build an education reform program around getting only “great” teachers. There aren’t that many “great” teachers out there – there aren’t that many “great” anything out there. But that doesn’t mean you couldn’t incrementally improve the quality of the teaching pool – if you knew what you actually wanted. The current system’s monetary incentives are designed to generate a pool of “lifers.” We’re actually spending quite a lot of money to generate such lifers, because so much of teacher compensation is back-ended in the form of great benefits and salaries keyed to seniority. A much flatter pay scale, with significantly higher incomes for starting teachers, and with a greater share of benefit costs paid by the teachers themselves, would change those incentives. You’d wind up with fewer long-term veteran teachers – but you’d also attract some people, both young people and second-career folks, who don’t consider teaching now because of very low starting salaries. I suspect you’d wind up with a stronger teaching pool overall at a lower overall cost.

That’s not the end of the story, of course. Transition costs for moving from the current system to my proposed compensation scheme would be significant, as promises to existing veteran teachers would need to be kept; for a decade at least, you’d actually increase costs as the costs of new teachers increase while veteran teachers retire on their generous pensions. The unions, meanwhile, would be expected to fight any such transition bitterly – and understandably so. But if we’re going to have a fight with the unions, these bigger structural questions are where to do it, not over merit pay.

More generally, I agree with E.D. Kain that any reform that is going to succeed needs to get a decent amount of support from teachers. The rhetorical strategy of trashing teachers is a disaster for education reform. And all the unions are doing is looking out for the economic interests of teachers. Many of the reforms we’ve seen so far have been heavy-handed bureaucratic interventions that, to many teachers, feel quite unfair, and for good reason make them distrust the reformers. But at the same time, you have teachers lining up to work for high-performing charter schools like the KIPP schools or Democracy Prep, the charter I’ve been affiliated with. And this in spite of the fact that, if they are lucky enough to be hired, they will work harder than they would at a traditional public school, they will be evaluated more strictly (and expected to conform more strictly to a set curriculum and classroom-management style), and they will have no job security (teachers in these schools are generally hired under one-year contracts). Obviously, evaluation as such, or lack of job security as such, isn’t the problem; it’s that the teachers at these high-performing charters have enough confidence in the school leadership, while teachers in the regular system have little reason for confidence in the distant bureaucrats responsible for their fates. And that, in turn, should clue us in to the possibility that the some of the largest-scale reform initiatives – like applying high-stakes testing to the entire school system – may be backfiring, driving out good teachers and weakening good schools even as they do succeed in identifying the worst schools. That doesn’t mean “education reform” is futile. it means education reformers need to critique their own efforts with as much vigor as they have critiqued the current system, and learn based on actual results how to reform their reforms, so that we’re creating the proper incentives, positive and negative, to push the system incrementally in the right direction.