Thursday, July 31, 2008

Teacher Merit Pay

Across the nation, as the public becomes aware that 90% of all school expenditures goes for the compensation and benefits of the teachers, staff members and administrators, and that it is the increases in the costs of the compensation and benefits that drives need for ever-increasing funding, the notion of performance-based pay – for teachers in particular – is getting more and more attention.

One reason is that most of us in the private sector live in such a compensation regime every day. Our bosses evaluate our performance and make decisions about our compensation all the time. Business owners have their value judged every day by their customers. Competition, innovation and being rewarded for results are seen by many to be the foundation of the American economic system. Even one of the largest of American government institutions – the United States Postal Service – has learned how to live in, and even embrace the competitive world (note the FedEx and UPS pick up boxes in front of the local post office).

So it is with some dismay that we come to realize that the largest component of our property taxes goes to pay for the compensation and benefits of folks who work in a system where only length of service and education level determine compensation – not actual performance. Surely, many say, one thing we need to address in this school funding crisis is to connect teacher pay to teacher performance.

From a theoretical perspective, I agree. But as is usually the case, the devil is in the details.

The first problem is determining how to measure the performance of a particular teacher. The most common answer is to tie teacher compensation to standardized test scores. Just to simplify things, let's say we're talking about a 5th grade teacher who has the same classroom of kids for the whole day, all year. How much of the performance of those kids on standardized tests can you attribute to that one 5th grade teacher, versus the kindergarten, 1st, 2nd, 3rd and 4th grade teachers that had the same kids in earlier years? And hasn't research shown that one of the best predictors of a kid's performance in school, especially in the early years, is the degree in which that kid's parents get involved – reinforcing and tutoring at home? Should teachers be rewarded for teaching in neighborhoods where there are many such parents? Should teachers in different settings be penalized?

Some school districts have experimented with performance pay for teachers. One of the most-observed examples is Denver Schools, which instituted a system they call ProComp. The system was developed as a joint effort by the School Board and the teachers' union. Teachers on board when the program was instituted were given the option to join or stay on the old system (50% switched), while all newly hired teachers are automatically enrolled in the program. Denver voters approved $25 million/year in new taxes to fund the system.

So how is it working?

Everyone seems to agree that it needs some tweaks, and leaders of both the district and the union say they knew this would be the case. School officials want to make some changes now, including increasing starting salaries and giving additional incentives for teachers willing to work at "hard to staff" schools, and to teach science and math. The principals in the district seem to think this is working.

The teachers' union wants to wait for an external evaluation of ProComp, scheduled for next year. What would motivate the teachers' union to wait?

One is that the money to increase starting salaries would be freed up by putting a cap on the upper end of the salary scale. School officials point out that a teacher with less than 12 years experience is twenty times more likely to quit than a teacher with more than 12 years experience. Therefore, it made sense to them to shift salary dollars from the top of the payscale to the bottom. I think the leaders of the teachers' union, themselves likely to be senior teachers, want to put off any such radical changes until the external review is completed – hoping it won't support this approach.

The teachers' union preferred an across-the-board 3.5% increase for everyone – business as normal.

The real issue is a big, irresistible plum sitting out there for everyone to see. It turns out that very little of the incentive money has been paid out. For the school year just ended, only $7 million of the $25 million was awarded. The ProComp fund now holds an $86 million surplus. One evaluation of the system, which looked back at 2 years worth of data, stated that there was only a slight difference in the test scores of the students of teachers who participated in the program versus those who did not.

I suspect that there are some, notably the teachers' union, who say the system is flawed because it failed to pay out that $86 million that they may view as rightfully theirs. On the other hand, taxpayers might be very happy that, since there was little difference in the test scores between the participants and non-participants, not much money was paid out, and therefore demonstrating a faithful adherence to the plan they voted for. Presumably, the taxpayers would like it even more if test scores rose dramatically and there was cause to pay out the bonuses.

Who can say whether a merit pay system will ever be seriously discussed in Hilliard Schools? I've been involved in the design and administration of incentive compensation plans for many years. Even in a dollars-and-sense business environment, it's not simple to design a plan that actually achieves the results one desires. There are lots of bad things which can happen: a) the plan fails to connect incentive goals to organization results, raising expense (the reward payout) and lowering profits (because the overall business results, e.g. higher revenue or lower costs were not achieved); b) the plan fails to recognize paths to success outside the incentive goals, which can lead to organizational profit, but no reward to the team; and, c) the incentive goals unintentionally pits one part of the team against another, allowing one team to be rewarded while the others – and the organization – lose. There are many more. Designing an effective plan has to be at least as difficult in a school district.

Lastly, one thing I learned from the compensation experts is that not everyone is motivated by money. Some are – they often go into a sales role where compensation is very directly tied to measurable performance. Others are motivated by freedom, or by working with inspiring leaders and colleagues. It would be a mistake for those of us who are not teachers to believe we understand what motivates someone who chooses teaching as a profession, and it would also be a mistake to treat all teachers as a homogenous group, motivated by exactly the same thing.

Our goal has to be a common understanding. People of the community – respect the difficult task teachers have today. It's not the same profession it was when we were in school. And teachers, please realize that the rest of us don't understand why you expect us to make sacrifices to increase your compensation just because you put in another year of service.

This kind of common understanding takes a long time to develop. Unfortunately, it is only 61 days until early/absentee voting begins on our operating levy.


  1. One of the best and most objective blogs I've seen for awhile. You certainly lay both sides out there and it's easy to see both sides.

    One question that has been on my mind and now seems like a good time to ask it....

    I realize you are in favor of merit pay, as am I. Like you, I realize it is a difficult and risky proposition (i.e., the potential risks you cited in your blog entry). I have the following question.

    1) Would merit pay actually save the district money or would it simply be a method to reward good teachers and "motivate" bad ones?

    It's true that some teachers get a 7% raise and I think we hold on to that as the actual increase in expenditures for teacher payroll. While some get 7%, others get 3%, so there is a weighted average here. So the actual increase in teacher's payroll is closer to 5%. Still a very good raise. And it is not my point to "justify" such a raise, I'm just normalizing everything for this question.

    If a merit system is used, what would the average payroll increase be? 3%, 4%, 5%? I really don't know. But let's say one-third gets zero raise, one-third gets a COL increase of 2.5%, and the top third get a 4.5% increase. That takes the average payroll increase down to about 2.3%. That would certainly be a savings. Is that enough to reward merit? I don't know one way or the other, but it's an interesting point of discussion.

    If today we increase teacher payroll by 5% and under this new system we raised it ~2.5% to 3% it would certainly reduce the rate of expenditure increase. But it wouldn't solve the problem entirely. It fixes a lot, but it only delays the need for more money from the taxpayers. Instead of every 3 years, it might be every 5 years. Better, but I still need a better school funding system than that. I can't continue to pay an additoinal $1000 every 5 years anymore than I can every 3 years. I'm lucky, I think I can swing another levy or two. Many aren't, and I understand that. But let's go out 15 or 20 years. None of us can continue to support that kind of burdern.... even if a reasonable wage increase is instituted.

    Anyway, that's my 2 cents.

  2. KJ:


    One of the characteristics of a good merit pay system is that there is a win-win nature to it: the organization wins and the employee wins.

    The way the organization wins is through productivity. While the employees may be getting paid more individually, as a team they produce more for less money. For example, if workers in a steel mill get a 20% bonus for producing 40% more steel in a year than they did the previous year (while maintaining quality and safety), then presumably the mill and workers both make more money.

    In a team goal environment like this, an interesting phenomenon happens - the high performing workers want the management to get rid of the low performers who jeopardize their bonus.

    There have been experiments in school systems, I think NYC is one, where incentives are group based - evaluated in terms of the performance of the team. In the context of Hilliard Schools, it might mean paying rewards based on the State Report Card results.

    It's not quite that easy of course. The demographics of our schools, particularly at the elementary level, is not uniform. As we'll see with this year's report card, some elementary schools scored "Excellent" while others were "Continuous Improvement." Is that because there are teachers of different capabilities in those schools, or is it because of the different demographics? Should the teachers in the lower performing school be denied merit pay when in truth they may be doing a great job and getting extraordinary results?

    Or maybe a merit pay system could be designed so that no elementary team gets a merit bonus if ANY elementary school scores below "Effective." Such a system might lead to more inter-school collaboration for example.

    Where's the savings in all that? Perhaps we add an element to the merit pay that gives rewards for keeping the comp & benefits expenditure growth under a goal amount - say 5%. By involving the teachers in devising the expense control solution and changing the game from the current adversarial system, some pretty inventive stuff might happen.

    And note that the rewards don't have always be bonus checks to the teachers. Could be that for achieving the performance/expense goals, a team might be rewarded with new equipment, or a special training event, or ??

    Anyway, the point is that these things don't have to cost more money.

  3. Teacher Magazine just published an interesting article about a teacher merit pay program in North Carolina.

    I realize that you may not be able to read this article unless your sign up for their free e-newsletter, so here's the gist of it:

    The NC bonus program is set up so that all teachers at a school building receive a bonus if students meet certain expectations on state exams. Apparently those expectations were met in many schools, and something on the order of $107 million in total bonus money was to be paid out.

    Yet the state General Assy ordered the State Dept of Ed to give out only $94 million. At the individual teacher level, this means that instead of the $1,500 bonus they expected, it will be hundreds less.

    The teachers say they're getting screwed (my word, not theirs).

    Education critics say "the bonuses were too easy to get" and "they've become an entitlement in the minds of the teachers."

    I don't have enough facts to judge either way.

    I worked in a business where our sales folks were paid a salary plus commission. The commissions were based on a quota - the amount of new business each salesperson was expected to bring in. The sales rep would receive one commission rate below quota, and a greatly increased rate above quota. Some reps in some years made many hundreds of thousands of dollars. That's a good thing because the company made $millions, and all of us got nice bonuses.

    As an executive, I spent a good time going around the world to help in the sales effort when I could, and to learn about what was going on at the front lines. On one such trip, a highly compensated sales rep complained to me that his quota had been raised mid-year, and it wasn't fair.

    So I asked my boss - the President of the company - about it. He said that the sales rep had a unexpected increase in business from this customer that cause his commission to skyrocket even though the rep had made no real effort. If this rep got paid this huge commission, he would sit on his butt the rest of the year and not try to build up a more sustainable revenue stream.

    I told my boss that it seemed that a deal was a deal - and if the comp plan said the rep should be paid, we should pay him.

    Then my boss told me the rest of the story. A few years before, this same customer had a huge downturn in business, and the rep ended up way below his quota. So the rep made the argument that it wasn't his fault that the customer cut back, and he shouldn't be penalized. His boss wanted to keep the rep, and appealed to the President for some relief. The President agreed to lower the rep's quota so that he would get a decent paycheck that year even with the downturn.

    All merit pay systems have one thing in common - they can be gamed by the participants. I saw reps delay the signing of additional business at the end of the year because they had already hit their quota, and wanted to use the new business as a basis for meeting their next year's quota. By the way, in some cases they were told to do so by top level executives who were under their own performance plans and wanted to hide the revenue pipeline from the Board of Directors.

    Comp plans are like laws and sausage - you don't really want to see how they're made or administered.