Why Merit Pay Will Improve Teaching

One of the bitterest controversies in today’s school-reform debate is merit pay—rewarding teachers not for seniority and the number of ed-school credits they’ve piled up, as public schools have done since the early 1920s, but for what they actually achieve in the classroom. Education reformers argue that merit pay will give encouragement to good teachers and drive away bad ones, and thus improve under-performing public schools. But most teachers’ unions adamantly oppose the idea. We don’t have reliable means to measure a teacher’s classroom performance, the unions charge, so merit plans will inevitably result in supervisor bias and favoritism: "Just too many cliques in the system," one teacher typically complains in a recent survey.

Nowhere is the incentives debate raging more fiercely than in New York City, where teachers’ contract negotiations have been at an impasse for months. Mayor Rudolph Giuliani has demanded that merit pay for individual teachers be part of any deal; the teachers’ union response (at least so far): fuhgeddaboudit.

Missing from the argument, though, are lessons from the private sector, where sophisticated, effective performance-based compensation has been de rigueur since the 1980s—part and parcel, experts believe, of corporate America’s hugely successful restructuring. Also ignored are experiments with comprehensive merit-pay plans that are under way in a few innovative school districts across the country—districts burdened with much less political resistance than Gotham.

To get a sense of what merit pay could do for the public schools, consider the benefits it has showered on American industry over the last two decades. Before the eighties, merit pay in U.S. firms—if it existed at all—was pretty simple: the boss gave you a fat bonus if you (or your unit) met sales or production goals. But as international economic competition pummeled them in the early 1980s, U.S. corporations, desperate to regain their competitiveness, began to experiment with measuring individual worker performance. They establishing pay incentives to improve it in formerly hard-to-measure categories of output and in previously intangible areas like customer service or product quality. Of course, the bottom line was still the bottom line, but these intangibles, companies now reasoned, mattered to the long-term economic health of the firm, even if they didn’t show up right away in the quarter-by-quarter numbers.

Familiar today, the new performance criteria—and the multi-faceted compensation plans built on their foundation—were strikingly original at the time. Retailers hired "mystery" shoppers to check out how employees treated customers, and based salaries, in part, on what they found. Businesses built into sales contracts "integrity" clauses that gauged not just how many widgets an employee sold but how long clients stuck with him. Banks remunerated loan officers not just for the lending they brought in but for the long-term quality of their loan portfolios. Some auto dealers tied part of salesmen’s pay to how customers rated them in follow-up surveys. Companies combined these kinds of individualized incentives with rewards for everybody if the whole firm did well. Airlines, for example, gave the entire crew bonuses if the fleet’s on-time performance improved.

Predictably, when U.S. businesses first introduced these innovations, workers grumbled, especially in heavily unionized industries like auto manufacturing, where any change threatened cushy labor arrangements. "They said that you couldn’t measure some things, that the pay systems were too subjective, that supervisors were too subjective—in short, everything that teachers today are saying," observes Alan Johnson, a New York-based compensation consultant. Many efforts stumbled at first, too, and companies had to discard or overhaul them. Creating effective programs, it became clear, would not be an overnight fix. "Even today, with all we know, it takes three years to start up an effective incentive-pay program," cautions Martha Glantz, a compensation expert with Buck Consultants in Manhattan. "You can spend the first year just deciding what the company’s goals and missions are, and collecting the data."

But American firms, needing to change or perish, forged ahead, winning over employees who liked the challenge of incentive pay and, through trial and error, developing pay plans that worked. By the mid-1990s, half of all major American corporations used such incentives. "It’s no longer credible to say you can’t measure something or that the only thing you can measure is a simple output," says Johnson. Merit pay played a crucial role, most observers believe, in generating the zooming productivity gains and superior product quality that American firms began recording in the late 1980s and that have been central to the nation’s economic prosperity ever since.

The public education monopoly has long resisted merit pay with the same ferocity with which private-sector workers at first greeted it. Opponents have constantly invoked previous attempts that failed—though their only examples have been two experiments that are over 100 years old, and another from the 1960s, before modern notions of performance pay emerged. The conventional wisdom among educators had long been that any attempt to pin down exactly what makes for good teaching, let alone measure and reward it fairly, was doomed to fail. "There was a general feeling that you were either gifted as a teacher or you weren’t, and that good teaching wasn’t something you could define," says Charlotte Danielson, an expert on teaching at the Educational Testing Service in New Jersey. As a last-ditch defense, some educators even argued that factors outside school, especially a student’s socioeconomic and family situation, had a much greater impact on student performance than teachers did, so that using merit pay based on student performance to promote good teaching, even if it could be defined, wasn’t fair. The unasked question was why teachers should ever receive salary hikes if what they do doesn’t matter.

Over the last decade, these views have utterly collapsed, undermining the intellectual case—weak as it was—against merit pay. A key figure has been University of Tennessee statistician William Sanders, who discovered how to measure a teacher’s effect on student performance. Rather than try to filter out the myriad sociological influences on pupils, a nearly impossible task, Sanders used complex statistical methods to chart the progress of students against themselves over the course of a school year and measure how much "value" different teachers added. Now called the Tennessee Value-Added Assessment System, Sanders’s approach proved what every parent already knew: not only did teachers matter, but some were lots better than others. Other education experts, including Danielson, author of several popular books on pedagogy, developed widely accepted criteria to judge good teaching, which put paid to the absurd notion that it was too elusive to define.

If the 1990s helped re-establish the centrality of teaching in the education debate, however, that victory hasn’t melted away the teachers’ unions’ political opposition to merit pay. Even so, a small number of school systems across the country, under intense pressure from parents, politicians, and administrators to improve student performance, have turned to merit pay to promote better teaching. And, after some give-and-take, they’ve managed to get the teachers’ unions on board.

Cincinnati’s public school system, the first to experiment with performance incentives, persuaded its teachers’ union in 1997 to do a test run of merit pay. Two years later, a ten-school pilot program, designed by administrators and teachers, got under way. Essential to union support was the pilot’s proposed use of peers to evaluate teachers. "The peer evaluators, who have no stake in how teachers are judged, are important to the perception of the fairness of the system," observes Kathleen Ware, associate superintendent of Cincinnati schools. Using Danielson’s criteria of good teaching—they include class preparation and clarity of presentation—the principals and peer evaluators devoted 20 to 30 hours to assessing every teacher in the ten chosen schools. Based on how they scored, teachers then wound up in one of five salary categories, with "novices" making the least money and "accomplished" teachers the most.

The pilot proved successful. A majority of teachers involved found it fair and judged the standards used as appropriate for the whole school district. The city’s board of education adopted it in the spring of 2000, and, in a subsequent election, union members signed on. Teachers will go through evaluations every five years, though those looking to move up quickly can request an appraisal after just two years. New teachers and one-fifth of all experienced teachers are having evaluations done this year, but no one will start getting paid under the new system until 2002.

Unfortunately, Cincinnati’s new program doesn’t directly use student test scores in its evaluations. Bringing in scores would have generated too much union hostility for the plan to gain acceptance, reports school superintendent Steven Adamowski. And in all likelihood, tying pay solely to test scores is a bad idea; nobody would call meritorious a teacher who boosted scores but left his students psychological wrecks because of his bullying. Yet leaving tests out altogether also makes no sense. After all, what better way to determine how well students are doing—the only reason for the concern over teaching quality in the first place—than test scores? Cincinnati’s program recognizes this implicitly. The district will monitor test results of students whose teachers score the highest ratings. If those students don’t show substantial improvement, school officials will toughen the program’s standards. In addition, the district will rely on student tests to award bonuses to all teachers in a school whose kids take big strides.

One major benefit of merit systems is that they enable schools to pay teachers—especially young and ambitious teachers—fatter salaries. That’s the overriding rationale behind Iowa’s new incentives program, enacted by the state legislature to keep better-paying nearby states from spiriting away top teachers. The state has anted up $40 million for salary increases, but, in a program similar to Cincinnati’s, Iowa will now evaluate teachers thoroughly to make sure the extra dough goes only to the good classroom performers, not the duds. "We believe this system will help us retain the best people by re-professionalizing teaching," enthuses John Forsyth, chief executive of the Des Moines-based Wellmark Blue Cross and Blue Shield. Forsyth helped the state dream up the new system, looking to the market for inspiration. "We know good teachers make a difference," he continues, "and we’re going to pay those who do make a difference." How much more reasonable this is than the New York union’s claim that, because a few city teachers jump to higher-paying suburban schools each year, all teachers—the good and the bad alike—need equal raises to make them stay.

In Iowa’s new system, good teachers will now be able to reach higher salary levels much earlier in their careers than before. "The private pay consultants who looked at our old seniority system thought it must have been designed specifically to keep teacher pay low and save school districts money," says Ted Stilwill, director of Iowa’s Department of Education. "The only way for teachers to get paid more under that system was for them to stick it out for years." As in Cincinnati, union opposition means that Iowa won’t rely on test scores to evaluate teachers—at least not directly. But in addition to paying teachers based on their performance evaluations, the state will also offer modest yearly bonuses to all teachers in a school whose students do well on standardized tests, with the biggest bonuses going to the school’s best instructors, rather than all teachers getting equal rewards.

Helping to develop Iowa’s plan has been an eye-opener for the state’s education chief. "Private-sector compensation experts taught me that businesses use pay as a way of getting everyone to follow common goals," Stilwill says. "Being in the public sector most of my life, I never understood that."

Though Cincinnati and Iowa have skirted the controversy of using student test scores, Denver is confronting it head-on. The city has launched two pilot plans that link pay directly to scores. One pilot program uses student scores in standardized tests of basic skills; another relies on scores in specific subjects. Principals and teachers agree at the beginning of the school year on what kinds of improvements in test scores they’ll shoot for and then face evaluation at the end of the year to see if they’ve met their goals. Denver is also instituting a third merit-pay pilot program that instructs teachers in the principles of good teaching and then evaluates their teaching skills and rewards them accordingly. Denver will later measure how their students perform on tests to see if its criteria for good teaching really produce results. Through these experiments, Denver hopes to figure out what motivates teachers best and what works best for students.

Denver’s teachers’ union, surprisingly, has contributed mightily to developing the pilots; the head of the project’s design team, Brad Jupp, is a union negotiator. He says the union is participating because the demand to make schools more accountable is so intense that teachers would rather help create performance systems than have them imposed from above. "If you believe that your union members are doing a good job—and we do—then you want a system that accurately measures that," says Jupp.

Experiences from the private sector suggest that it will take several years before these imaginative programs work out all their kinks. In the meantime, controversy will dog the new programs. Critics will pounce on their every mistake as evidence that paying teachers for performance is a bad idea. And unions are likely to push for watered-down plans in order to deflect criticism without giving up too much.

This last is exactly what’s happening in New York. The teachers’ union has offered the mayor a compromise: pay every teacher in a school or district a bonus if the school’s or district’s test scores rise. A trial run of such a system, supported by Gotham’s business community, is under way in two urban districts; other states, including California and Georgia, already have school-based bonuses.

But group bonuses will never substitute adequately for true performance pay, compensation experts believe, since they don’t single out good and bad teachers. All they’re likely to do is to frustrate first-rate teachers working in schools with mediocre staffs. "If you have four workers doing well in a unit that is not otherwise performing, over time those four will leave the company and go somewhere that they can be rewarded for their superior work," says consultant Glantz. And since schoolwide bonuses don’t put any pay at risk—poor performance doesn’t mean less money—they won’t help rid schools of lousy teachers either. By contrast, teachers who score poorly in, say, Cincinnati, now find themselves shunted into the lowest salary level, discouraging them from sticking around. Mayor Giuliani, rightly, has rejected the union’s offer.

Of course, smaller cities like Cincinnati are far removed from the we-don’t-do-windows union obstructionism of a New York or a Los Angeles. For 15 years, Cincinnati’s teachers’ union has accepted some kind of peer evaluation of teachers; it’s easily one of the nation’s most flexible teachers’ unions. Unions in New York and Los Angeles, conversely, have fought almost every education reform tooth and nail. Last year in Los Angeles, thousands of teachers ferociously protested against a proposed merit-pay plan, eventually killing it. Moreover, in the current public school monopoly, there’s nothing really comparable to the outside economic pressure that forced American industry to develop its merit programs—one more argument for school choice.

Without individualized merit pay, teacher evaluations will remain perfunctory at best. Today, New York principals fail less than 1 percent of all teachers in annual evaluations. New York hopes to get principals to crack down on bad teaching by rewarding them financially when their schools do well. That will eliminate one of the main objections to teacher merit pay—that it leads to supervisor favoritism: even if a principal hates an effective teacher’s guts, he’s not going to want to lose someone who’s helping him sweeten his own salary. But until teachers are part of any performance-pay system, the impact of such innovations will be severely limited—and students will continue to get shortchanged, regardless of how much their teachers take to the bank.