Why do unions oppose merit pay?

Bryan Caplan asks:

I don't doubt that unions tend to oppose merit pay, but the reasons are unclear.  Profit-maximizing monopolists still suffer financially if they cut quality; the same should hold for unionized workers.  Why not simply jack average wages 15% above the competitive level, and leave relative wages unchanged?

Or to put the puzzle another way: Once you've secured a raise for all the workers in your union, why prevent employers from offering additional compensation for exceptionally good workers?

Earlier, Megan McArdle considered the topic.  One simple model is to invoke the median voter as either ruling the union or constraining it.  The implication is that most union members fear they will lose from greater accountability, even if the total size of the pie goes up.  As Megan noted, unions are set up to favor the bottom 55 percent of the workers; furthermore productivity can be very unevenly distributed.


Perhaps it's like this post, http://www.marginalrevolution.com/marginalrevolut... where you discuss BS in corporate culture. Merit pay is a threat to morale within a union because it threatens member cohesion. Unions, like corporations, prefer order.

I'm not really sure why everyone fails to understand this. They're rent-seeking organizations whose explicit primary objective is to secure more compensation for less actual work. The question you should be asking is why in God's name would they want to introduce competition into the equation?

You might also note that unions typically do not oppose bonuses based objective measures of performance outside of the control of the employer, such as piece work rates which reward employees based on their output. (Pay is based on x cents per unit; faster employees get paid more). It is when the measure is subjective and in the hands of the employer who can discriminate for other reasons that there is friction and distrust.

One could also note another observation of the median voter in the union: relative pay is often what matters. Multiple behavioral studies have concluded that many happiness measures are only responsive to relative pay rather than absolute pay. Given this, if the median voter in the union fears that he would "lose out" in relative pay with merit-based incentives, then he would not want such a system in order to preserve his/her own happiness.

Union members do not like merit pay for the same reason faculty members do not like it when someone whom they do not like is appointed Dean or Department Chair.

DoctorWes, you make several good points, but I would not treat anyone's arguments as solid evidence for why they do one thing or another. This is especially true when we are interested in the behavior of a large class of institutions, where the arguments of their representatives (even if sincerely believed) are separate from the institutional dynamics that actually structure society.

For instance, the argument that you proposed is only applicable to teachers, so it would not explain any broad tendency to avoid merit pay (if such a tendency exists). However, ti would be possible to generalize that argument into something like "we know how to do our jobs, and attempts at oversight will only distract us from doing them".

On an institutional level, we need to be aware of the fact that ideas are only part of the social dynamic -- they are not the cause of social dynamics. Therefore, we could have a situation where unions that adopt particular policies (justified by particular arguments) tend to be stable, whereas unions with alternative structures (and supporting ideologies) tend to fail.

@Travis Ormsby: Re 1 - artists are among the most unequally compensated people.

Re 2 - this definition of "fairness" is begging the question.

Isn't 3 opposed to 4+5 re assumptions about what good teachers value?

As a former teacher (in a private, non-unionized, school, where the vast majority of my colleagues opposed merit pay):

1) There is a cultural discomfort with the idea that transcends any kind of logical argument; that has to be addressed before the idea can even be debated on its merits.

2) There is enormous, and in many (not all) cases well-founded, concern about the school administration's ability to fairly assess teacher quality. Some of these could be addressed, perhaps, by careful use (and explanation) of statistics and procedural safeguards. But some are genuine, unsolved problems in teacher evaluation.

3) There's widespread distrust of school administrators, in some cases backed up by actual negative experiences with, e.g., administrator favoritism or cluelessness. People are skeptical that merit-pay powers would be used to award the truly meritorious rather than the principal's pet, and they're skeptical that administrators unfamiliar with their subject or grade level can properly evaluate them.

4) Many teachers have a collectivist orientation, related to their drive to do human service in the first place. They don't see classroom outcomes as necessarily indicative of one teacher's success; the work teachers in a school do is interrelated, the things students learn in one class apply in another, and how can you untangle it all and say a student's success reflects on one teacher alone? (I expect teachers are far more amenable to group than to individual bonuses.; I seem to recall cases where these have been implemented with much less rancor than the standard performance bonuses.)

5) Many teachers are offended at the idea that they would work harder for more money, both because money isn't a strong motivator for them in the first place (and is maybe seen as sort of grubby or non-idealistic), and because many of them are already working as hard as they can. (Not that teachers don't want raises, as they are often ill-paid; they just don't want them for performance reasons. And, per Daniel Pink, the motivation research might be with them on this anyway.)

6) Many schools simply don't have a strong culture of evaluation in the first place, and that's a prerequisite to having performance pay. There've been all kinds of media reports about districts where teachers almost never get an unsatisfactory rating. I was evaluated twice in five years, neither of them anything with teeth. Teachers are accustomed to having their classrooms as their own personal terrain, and the visits which would be necessary for evaluation are often perceived as intrusive. There's a lot of work to be done building a more open and collaborative culture before you can build an evaluative culture, and you have to build that evaluative culture before you can have effective performance pay anyway.

With all that said, the lack of a performance pay option was one of the things that ultimately turned me off of teaching. But I believe I am very much an outlier in this regard, and even gung-ho as I am about the idea I respect the significant practical and cultural hurdles. I think this is an area where the union does broadly represent the membership.

Unions had much greater strength in the 1940's->1960's, a period of tremendous productivity growth, than they did in the 70's and 80's. They have greater strength in contemporary France and Germany, which generally meet us for per-hour productivity, than they do here. Simple thought exercise which repeat libertarian talking points shed little light on those two direct comparisons.

If I may make a practical suggestion: Devil take the hindmost. Let say 3% per year be subject to firing, criteria to be negotiated between the union and the employer. Surely, a union can't object to its worst 3% being liable to firing, or to put it the other way, can't justify forcing the employer to retain the worst 3% of their employees.

Since nobody will want to be in the lowest 3%, this will create incentive to a certain minimum performance, an incentive that will increase with time, as the very worst, those incapable or unwilling of improvement, are progressively culled.

"why prevent employers from offering additional compensation for exceptionally good workers?"

It could also be a unity issue, arising from the tendency of people to become envious of those that make more than them. As unions evolved, merit pay became antithetical because it promoted disunity by increasing aggregate jealousy in the group.

Have you ever listened to a group of teachers trash their co-workers? Some of the nasty office squabbles you'll ever hear. I don't imagine adding quantified merit to that equation would help.

The problem with arguments against merit pay by teachers unions is that those arguments are always based on "feelings and idealism". What they don't realize is that outside of the teacher bubble people are starting to question these statements and they are witnessing merit pay work in other areas.

The reason that merit pay works is that if some teaching skills are working somewhere, then they spread quickly.

Merit pay for teachers will eventually end up like non-union administrative pay:

Boards of trustees routinely give over-priced contracts to friends and friends-of-friends, with the result being disparate pay among persons with like jobs.

While this would be absolutely kosher in the world of private-sector business, it makes for a non-productive mix in schools -- where cooperation produces better results than individual competition.

>Why do unions oppose merit pay?

Maybe because it flies in the face of everything unions stand for? That could be it.

[1] The Union wants to be the sole determiner of how much money everyone makes. They are not going to give the employer any power in this decision outside of the bargaining room.

[2] The Union wants to protect its leaders. They are not going to allow pay increases and greater member standing based on anything other than seniority and certain Union-approved tracks.

[3] The Union wants to portray the employer and corrupt and incompetent at all times, and foster an "Us Vs. Them" atmosphere. They are not going to publicly state that the employer is capable of making big decisions like this on its own.

Incentive pay doesn't work very well in a lot of situations. Check out starting on page 114 in this book:

the unions representing actors, screenwriters, directors and athletes are exceptionally powerful, but compensation in these cases is dominated by "merit pay".

none of these people are paid merit pay in the same way it is proposed for teachers. Do they pay the best acting actor on a movie a million dollar bonus? Would that result in an increase in good acting? or would it result in counter productive showboating?

Athletes do get incentive pay, but even then this can result in counter-productive behavior if the incentive are poorly designed. Should Dwyane Wade get a bonus for scoring alot? Not with LeBron James on the same team he shouldn't.

@Sam Penrose: Part of why the unions were so powerful immediately following WWII is that all the other industrialized nations in the world had just finished being bombed into ruins. The US was the last man standing at the time and its workers could command a premium due to scarcity.

@Invisible Finger: the separation of compulsed customers vs. non-compulsed customers is already there. People choose to live in particular neighborhoods for the quality of the public schools or choose to send their children to private schools.

Of course, that doesn't prevent those people from still being taxed for the sub-standard public schools, but some (including myself) would argue that most taxation goes to sub-standard causes anyhow.

The producer or the team owner negotiates with the actors (so to speak) based on whatever each can bring to the table.

This isn't really what is suggested for the teacher's though. Teachers are theoretically free to go to other school districts or private schools and cut a better deal as well. That isn't merit pay.

Merit pay is when you identify the good teachers using some metrics and pay them more. This causes problems. It is disruptive to teamwork and results in gaming the system.

As Megan noted, unions are set up to favor the bottom 55 percent of the workers

Huh? I am kind of incredulous when I read this, if the claim is that unions are set up to favor the bottom 55% of their members by seniority and pay. Which, from reading the McArdle source, it is.

Unions are politically-governed beasts (leaders are elected) run to grossly favor senior members who have disporportionate power within it (just as seniors do in politics generally) at the cost of junior members.

* When a union has the choice of keeping a pay raise and having x% of its membership be laid off, or giving up the raise to save everybody's job, 9 times of 10 they keep the raise and members go overboard -- the most junior members.

McArdle herself recently related the story of senior UAW workers voting to close down an "old GM" plant to preserve their benefits, at the cost of younger workers in the plant who were willing to take a reduction to keep their jobs and the plant working.

* In big urban public school districts across the US, senior teachers get to pick their job slots. In cities like NY this has resulted in all the senior teachers moving to work in selective and nice-neighborhood schools (even if the principals didn't want them) and all the novices going to the slums schools. And as a consequence of senior teachers being paid a great deal more than novices, good neighborhood schools receiving much larger spending budgets per student than poor neighborhood schools.

* As to merit pay, as teacher Travis Ormsby said above:

Teachers value fairness, and in this context fairness means teachers of equal experience and equal education get equal pay.

Fairness for some reason does *not* mean teachers of equal merit get equal pay.

But he's right. Senior teachers who see a junior one in the next classroom being paid more than them just because he's "better" will hate it -- and take it out politically on the union leaders who agreed to it. That will start an internal political war -- give an important issue to the opposition! What politically elected leader wants that? That's why teachers union heads hate merit pay. To protect themselves.

In the world of real professionals (lawyers, doctors, etc.) not to mention competitive business, if somebody comes into the office next to yours having half your experience but getting twice your pay because he's better than you, you get the socially beneficial signal that it can pay off for you to improve your skills. But you don't have the political option of blocking that young guy from appearing by organizing with your inferior-skilled peers to vote your boss out if he does.

Unions are run for their junior members? C'mon. Get real. Unions are run for their senior members just like Congress is run for AARP.

If merit pay for teachers is such a good idea, why don't private schools do it? They don't do it because it would piss the other teachers off and cause dissension. It isn't worth the hassle. There are no unions at the private schools preventing this either.

Companies don't even usually pay those 10x productive superstar programmers that much more than the average programmers. Because it would piss the other programmers off and cause dissension. Good teachers are a lot less than 10x as productive than average teachers.

Why do CEO's oppose direct election of directors, or votes by shareholders on their own compensation or bonus programs?

I have no experience with unions or school politics, but I find all this debate about "what the teachers" union will vote on" silly. Public School's Teachers' pay comes from tax-payers. Tax payers should vote on how the teachers are paid, not teachers.

I think something people who are close to education suspect but do not want to actually say is at play here:

Good teachers do not matter.

Statistically there is very little evidence that 'good teachers' as a whole make much of a difference. From the time I've spent tutoring I've noticed that 'good teaching' does help in some limited areas. For example, when trying to explain complicated ideas I put a lot of energy into coming up with analogies and simple ways to explain these ideas step by step. I think I've gotten quite good at it. Nonetheless, though, this rarely causes a 'spark' of insight by itself in the student. Ultimately what really drives education is time. When the student spends hours doing long division he becomes better at it. When he does ten accounting problems he begins to master it.

So education is kind of like working out. If you got the best physical trainer in the world and followed his program for a week but someone else followed a half-assed program they buy from an infomertial religiously for a whole year they will likely be in better shape. In this context 'merit' only applies in a very limited setting. Not among the typical person looking for a work out program but among elite athelets for whom daily devotion to training is already a given. For the typical person the results of their exercise is less about the 'merit' of their trainer but about the effort they put into it.

So I can think of several reasons a union would oppose merit pay in this context:

1. It's not really 'merit'. At best its random, at worse its a way for management to break union cohesion by rewarding workers more friendly to their way of viewing things and selling them the idea that they are 'elites' held back by the low merit mass of teachers in the union.

2. Since factors other than individual merit would really drive the pay, the result would be competition for those factors. For example, an area where kids have more stable home lives, will produce better results. The result of 'merit pay' then will be less about improving teaching ability and more about scoring those assignments. You can normalize for this by accounting for demographics in the structure of merit pay but only very crudely. For example, not all upper income families are stable and not all lower income ones are unstable. Teachers 'on the ground' can seek to formulate classes of the stable type outside the view of those doing the demographic mapping.

Comments for this post are closed