The Folly of Merit Pay

There’s no end to the possible uses for that nifty little Latin phrase Cui bono?, which means: Who benefits? Whose interests are served? It’s the right question to ask about a testing regimen guaranteed to make most public schools look as though they’re failing. Or about the assumption that people with less power than you have (students, if you’re a teacher; teachers, if you’re an administrator) are unable to participate in making decisions about what they’re going to do every day.

And here’s another application: Cui bono when we’re assured that money is the main reason it’s so hard to find good teachers? If only we paid them more, we’d have no trouble attracting and retaining the finest educators that—well, that money can buy. Just accept that premise, and you’ll never have to consider the way teachers are treated. In fact, you could continue disrespecting and de-skilling them, forcing them to use scripted curricula and turning them into glorified test-prep technicians. If they seem unhappy, it must be just because they want a bigger paycheck.

In 2000, Public Agenda questioned more than 900 new teachers and almost as many college graduates who didn’t choose a career in education. The report concluded that, while “teachers do believe that they are underpaid,” higher salaries would probably be of limited effectiveness in alleviating teacher shortages because considerations other than money are “significantly more important to most teachers and would-be teachers.” Two years later, 44 percent of administrators reported, in another Public Agenda poll, that talented colleagues were being driven out of the field because of “unreasonable standards and accountability.”

Meanwhile, a small California survey, published last year in Phi Delta Kappan, found that the main reason newly credentialed teachers were leaving the profession was not low salaries or difficult children. Rather, those who threw in the towel were most likely to cite what was being done to their schools in the name of “accountability.” And the same lesson seems to hold cross-culturally. Mike Baker, a correspondent for BBC News, discovered that an educational “recruitment crisis” exists almost exclusively in those nations “where accountability measures have undermined teachers’ autonomy.”

That unhappy educators have a lot more on their minds than money shouldn’t be surprising in light of half a century of research conducted in other kinds of workplaces. When people are asked what’s most important to them, financial concerns show up well behind such factors as interesting work or good people to work with. For example, in a large survey conducted by the Families and Work Institute, “salary/wage” ranked 16th on a list of 20 reasons for taking a job. (Interestingly, managers asked what they believe matters most to their employees tend to mention money—and then proceed to manage on the basis of that error.)

Educational policymakers might be forgiven their shortsightedness if they were just proposing to raise teachers’ salaries across the board—or, perhaps, to compensate them appropriately for more responsibilities or for additional training. Instead, though, many are turning to some version of “pay for performance.” Here, myopia is complicated by amnesia: For more than a century, such plans have been implemented, then abandoned, then implemented in a different form, then abandoned again. The idea never seems to work, but proponents of merit pay never seem to learn.

Here are the educational historians David Tyack and Larry Cuban: “The history of performance-based salary plans has been a merry-go-round. In the main, districts that initially embraced merit pay dropped it after a brief trial.” But even “repeated experiences” of failure haven’t prevented officials “from proposing merit pay again and again.”

“Son of Merit Pay: The Sequel” is now playing in Cincinnati, Denver, Minneapolis, New York City, and elsewhere. The leading advocates of this approach—conservatives, economists, and conservative economists—insist that we need only adopt their current incentive schemes and, this time, teaching really will improve. Honest.

Wade Nelson, a professor at Winona State University, dug up a government commission’s evaluation of England’s mid-19th- century “payment by results” plan. His summary of that evaluation: Schools became “impoverished learning environments in which nearly total emphasis on performance on the examination left little opportunity for learning.” The plan was abandoned.

In The Public Interest, a right-wing policy journal, two researchers concluded with apparent disappointment in 1985 that no evidence supported the idea that merit pay “had an appreciable or consistent positive effect on teachers’ classroom work.” Moreover, they reported that few administrators expected such an effect “even though they had the strongest reason to make such claims.”

To this day, enthusiasm for pay-for-performance runs far ahead of any data supporting its effectiveness—even as measured by standardized-test scores, much less by meaningful indicators of learning. But then that, too, echoes the results in other workplaces. To the best of my knowledge, no controlled scientific study has ever found a long-term enhancement of the quality of work as a result of any incentive system. In fact, numerous studies have confirmed that performance on tasks, particularly complex tasks, is generally lower when people are promised a reward for doing them, or for doing them well. As a rule, the more prominent or enticing the reward, the more destructive its effects.

So why are pay-for-performance plans so reliably unsuccessful, if not counterproductive?

1. Control. People with more power usually set the goals, establish the criteria, and generally set about trying to change the behavior of those down below. If merit pay feels manipulative and patronizing, that’s probably because it is. Moreover, the fact that these programs usually operate at the level of school personnel means, as Maurice Holt has pointed out, that the whole enterprise “conveniently moves accountability away from politicians and administrators, who invent and control the system, to those who actually do the work.”

2. Strained relationships. In its most destructive form, merit pay is set up as a competition, where the point is to best one’s colleagues. No wonder just such a proposal, in Norristown, Pa., was unanimously opposed by teachers and ultimately abandoned. Even those teachers likely to receive a bonus realized that everyone loses—especially the students—when educators are set against one another in a race for artificially scarce rewards.

But pay- for-performance programs don’t have to be explicitly competitive in order to undermine collegial relationships. If I end up getting a bonus and you don’t, our interactions are likely to be adversely affected, particularly if you think of yourself as a pretty darned good teacher.

Some argue that monetary rewards are less harmful if they’re offered to, and made contingent on the performance of, an entire school. But if a school misses out on a bonus, what often ensues is an ugly search for individuals on whom to pin the blame. Also, you can count on seeing less useful collaboration among schools, especially if an incentive program is based on their relative standing. Why would one faculty share ideas with another when the goal is to make sure that students in other schools don’t do as well as yours? Merit pay based on rankings is about victory, not about excellence. In any case, bribing groups doesn’t make any more sense than bribing individuals.

3. Reasons and motives. The premise of merit pay, and indeed of all rewards, is that people could be doing a better job but for some reason have decided to wait until it’s bribed out of them. This is as insulting as it is inaccurate. Dangling a reward in front of teachers or principals—"Here’s what you’ll get if things somehow improve"— does nothing to address the complex, systemic factors that are actually responsible for educational deficiencies. Pay-for-performance is an outgrowth of behaviorism, which is focused on individual organisms, not systems—and, true to its name, looks only at behaviors, not at reasons and motives and the people who have them.

Even if they wouldn’t mind larger paychecks, teachers are typically not all that money-driven. They keep telling us in surveys that the magical moment when a student suddenly understands is more important to them than another few bucks. And, as noted above, they’re becoming disenchanted these days less because of salary issues than because they don’t enjoy being controlled by accountability systems. Equally controlling pay-for-performance plans are based more on neoclassical economic dogma than on an understanding of how things look from a teacher’s perspective.

Most of all, merit pay fails to recognize that there are different kinds of motivation. Doing something because you enjoy it for its own sake is utterly unlike doing something to get money or recognition. In fact, researchers have demonstrated repeatedly that the use of such extrinsic inducements often reduces intrinsic motivation. The more that people are rewarded, the more they tend to lose interest in whatever they had to do to get the reward. If bonuses and the like can “motivate” some educators, it’s only in an extrinsic sense, and often at the cost of undermining their passion for teaching.

For example, a recent study of a merit-pay plan that covered all employees at a Northeastern college found that intrinsic motivation declined as a direct result of the plan’s adoption, particularly for some of the school’s “most valued employees—those who were highly motivated intrinsically before the program was implemented.” The more the plan did what it was intended to do—raise people’s extrinsic motivation by getting them to see how their performance would affect their salaries—the less pleasure they came to take in their work. The plan was abandoned after one year.

That study didn’t even take account of how resentful and demoralized people may become when they don’t get the bonus they’re expecting. For all these reasons, I tell Fortune 500 executives (or at least those foolish enough to ask me) that the best formula for compensation is this: Pay people well, pay them fairly, and then do everything possible to help them forget about money. All pay-for-performance plans, of course, violate that last precept.

Measurement issues. Despite what is widely assumed by economists and behaviorists, some things are more than the sum of their parts, and some things can’t be reduced to numbers. It’s an illusion to think we can specify and quantify all the components of good teaching and learning, much less establish criteria for receiving a bonus that will eliminate the perception of arbitrariness. No less an authority than the statistician-cum-quality-guru W. Edwards Deming reminded us that “the most important things we need to manage can’t be measured.”

It’s possible to evaluate the quality of teaching, but it’s not possible to reach consensus on a valid and reliable way to pin down the meaning of success, particularly when dollars hang in the balance. What’s more, evaluation may eclipse other goals. After merit-pay plans take effect, administrators often visit classrooms more to judge teachers than to offer them feedback for the purpose of improvement.

All these concerns apply even when technicians struggle to find good criteria for allocating merit pay. But the problems are multiplied when the criteria are dubious, such as raising student test scores. These tests, as I and others have argued elsewhere, tend to measure what matters least. They reflect children’s backgrounds more than the quality of a given teacher or school. Moreover, merit pay based on those scores is not only unfair but damaging, if it accelerates the exodus of teachers from troubled schools where they’re most needed.

Schoolwide merit pay, again, is no less destructive than the individual version. High stakes induce cheating, gaming, teaching to the test, and other ways of snagging the bonus (or dodging the penalty) without actually improving student learning. In fact, some teachers who might resist these temptations, preferring to do what’s best for kids rather than for their own wallets, feel compelled to do more test prep when their colleagues’ paychecks are affected by the school’s overall scores.

It may be vanity or, again, myopia that persuades technicians, even after the umpteenth failure, that merit pay need only be returned to the shop for another tuneup. Perhaps some of the issues mentioned here can be addressed, but most are inherent in the very idea of paying educators on the basis of how close they’ve come to someone’s definition of successful performance. It’s time we acknowledged not only that such programs don’t work, but that they can’t work.

Furthermore, efforts to solve one problem often trigger new ones. Late-model merit-pay plans often include such lengthy lists of criteria and complex statistical controls that no one except their designers understand how the damn things work.

So how should we reward teachers? We shouldn’t. They’re not pets. Rather, teachers should be paid well, freed from misguided mandates, treated with respect, and provided with the support they need to help their students become increasingly proficient and enthusiastic learners.