Problems with Performance Pay

Graphic: This diagram shows Maslow's hierarchy of needs, represented as a pyramid with the more primitive needs at the bottom.

Julie Bishop, the Minister for Education in Australia's Howard Government, has announced that the Commonwealth Government intends to force the introduction of performance based pay for teachers in Australia's public schools.

Background

For the benefit of my international readers, education is a state responsibility under Australia's Federal Constitution.

Each state maintains a centrally managed public school system intended to make universal education available to all, but also sets curriculum, standards and final examinations for the whole education system, including the growing non-Government sector.

While the Commonwealth Government does not have constitutional responsibility for education, Commonwealth Governments of all political persuasions have become increasingly active in the education area in recent years through provision of funds. Now Minister Bishop want to use this leverage to force new teacher pay arrangements on the states.

Performance based pay arrangements exist to some degree in the non-Government sector. The proposal to extend the concept has met with strong opposition from State Governments and teachers' unions.

I have written a fair bit about education issues on my personal blog. In this context, I promised to write a piece looking at performance pay from a management perspective. Ninglun, one of my blogging colleagues, has just reminded me of that promise.

I am putting the post on this blog because schools are in fact professional services organisations, while the post also provides me with an opportunity to discuss performance pay issues in a broader context.

I should make my personal position clear. Twenty years ago I supported performance based pay. I no longer do so except in narrowly defined circumstances. My own experience with the practical application of performance based pay has been negative. One of the problems with performance based pay is that it all seems so intuitively sensible. After all, shouldn't we reward people for relative contribution, thus motivating them them to do better?

Thus in the Australian case Julie Bishop argues the need to reward the better teachers, in so doing increasing the maximum amount that teachers can get, attracting more people to the profession. Now I have no necessary brief for the way that teachers in public schools are currently paid, but I do think the the Minister's proposals are likely to have perverse results.

Money as Motivator

Does money motivate? Does higher pay, or the chance of higher pay, lead to improved performance? I have seen little evidence to suggest that it does, although I have to be careful how I phrase this.

Clearly, all firms operate in the marketplace for talent. They need to set pay structures that will attract and retain the people they want. One of the reasons why Australia has had a problem attracting teachers is that pay scales for teachers have been too low relative to other opportunities.

Individual firms also adopt specific recruitment and pay policies targeting the type of person they want (can afford). This includes specific areas like sales or corporate finance where performance pat including commissions are common. This may but need not lead to commercial success.

All this said, there is little hard data that I am aware of to say that higher pay or performance pay will improve performance at sector level beyond the market pay required to attract and hold people in the first place.

In fact, in his study of the transition of firms from good to great, Jim Collins found that there was no link between remuneration and performance. The only discernible difference between the good to great and the rest was that the good to great companies in fact paid their senior executives slightly less!

Pay and the Hygiene Factor

To understand some of the reasons why performance pay may not work and even be counter productive we need to look at the work of Herzberg.

Frederick Herzberg (1923 - 2000) was a noted psychologist who became one of the most influential names in business management. He is possibly most famous for his work in job enrichment and Motivator-Hygiene theory. According to Wikipedia (link above) his 1975 publication "One More Time, How Do You Motivate Workers?" is still the most requested article from the Harvard Business Review.

In 1959, Herzberg proposed the two factor theory of human motivation in the workplace. According to his theory people are influenced by two factors:

• Satisfaction and psychological growth are a result of motivation factors. These include achievement, recognition, work identity, responsibility, promotion, growth.
• Dissatisfaction is a result of a lack of hygiene factors. These include pay, fringe benefits, relationship with co-workers, physical environment, supervisor-employee relations.

So Herzberg puts pay in the hygiene class, something that can demotivate if its wrong, but not a motivator in its own right.

Now link this to the work of Abraham Maslow (1908-1970).

In 1943 Maslow proposed his hierarchy of needs. This is often depicted as a pyramid (see graphic) consisting of five levels: the four lower levels are grouped together as deficiency needs associated with physiological needs, while the top level is termed growth needs associated with psychological needs.

While deficiency needs must be met, growth needs are continually shaping behaviour. The basic concept is that the higher needs in this hierarchy only come into focus once all the needs that are lower down in the pyramid are mainly or entirely satisfied. Growth forces create upward movement in the hierarchy, whereas regressive forces push prepotent needs further down the hierarchy.

I have never been a supporter of all the detail in Maslow, but his work does explain why I think that Herzberg was correct to classify money as a hygiene factor. It all comes back to the impact of pay and changes in pay on Herzberg's motivational factors and the higher levels in the Maslow pyramid including especially esteem.

This is best illustrated by example.

Practical Examples of the Hygiene Factor at work.

Take a worker in the firm and give him/her a one-off bonus for good performance. This increases the employee's self esteem, and is likely to motivate.

Now move to a pay system combining a base salary with a performance bonus. In this, the first year, staff get varying levels of bonus. This has a number of effects.

In doing their family financial planning, staff have to make estimates in advance of likely income. Those who get more than expected are pleased, their esteem increases. Conversely, those who get less than expected are disappointed, their esteem is reduced. If the shortfall is significant, then problems may arise with lower levels on the Maslow hierarchy if, for example, the family is suddenly short of cash and has to cut back.

Humans are social animals and like to know their place in the pecking order. No matter how much people are told to keep bonus details confidential, the office will soon work it out. This introduces a new dynamic as people are happy or disappointed with their relative position.

I have previously suggested that one of the most common mistakes made in designing a performance appraisal system, one of the most common reasons why they fail to contribute to performance improvement, is the linking of performance appraisal and pay.

You can see why just on the simple example given above. Scarce staff and management time is effectively diverted from business not just into running the process, but in handling all the consequent complaints. This is why so many index based pay linked appraisal systems end up with average scoring just above the average, why there are so few highs and lows. It's just easiest.

You get what you measure

One of the common problems with performance measurement systems in general, performance based pay in particular, is that you get what you measure.

This need not matter if there is a clear, direct and discernible link between what you want from staff and pay policy.

One of my friends is a futures trader for a bank. He gets a small pool of money to play with, a relatively small base salary plus a bonus of a percentage of the money he makes. He is very happy with that.

In similar vein, Australia's Macquarie Bank, one of the world's most successful investment banks, has bonus arrangements directly related to fees generated over time. Macquarie attracts a particular type of person interested in the possibility of earning well above average incomes, of becoming a millionaire.

These are reasonably clear cut cases, although even here problems can arise.

To begin with, you build in an incentive to undertake those things that maximise pay. This can and has lead to risky decisions and even fraud. Then you get a second problem if the institution's market position slips or if conditions change, leading to declines in firm fee income and hence performance bonuses. Those key rainmakers that you really wanted to keep may be the first to jump ship.

Cloudy Relationships

Problems become more complex where the relationship between those things being measured and the success of the organisation are more complex or even contradictory.

Take time based charging, a common practice in many parts of the professional sector. If, as is usually the case, performance measurement focuses on billings, this creates an incentive to maximise billings at the expense of the client. Problems here have been well explored on Chris Marston's Inside the Firm of the Future blog as well as on blogs such as Adam Smith Esq or David Maister's. David's latest post actually provides a specific example of assessment problems.

Problems arise too where you actually want your people to perform across a range of dimensions.

Assume that you want staff to take a degree of personal responsibility for marketing, but measure only billings and focus on that in remuneration. You can be sure marketing will be ignored.

Assume that you need your staff to work in teams, but focus instead on individual performance. You can be sure that team work will suffer.

Particular problems can also arise where elements of performance depend on factors outside individual control or are not directly measurable.

Julie Bishop's Position

I now want to tease all this out a little by taking Minister Bishops' arguments as reported in the Australian press.

To improve national standards, Minister Bishop, wants teachers to be assessed on the improvement of their students over the year. The aim is for the best teachers to benefit most, especially at disadvantaged schools.

In coming to her views, Minister Bishop appears to have been influenced by some US studies including a University of Arkansas study showed maths results rising by about 4 percentage points at schools with a generous performance pay system. She also believes that there are disincentive affects built into the current pay system where pay peaks after nine years, reducing incentive for further teacher improvement.

Her preferred model is for performance bonuses based on two things: improvements in student exam scores compared with statewide results; subjective assessments by principals, parents, students and other teachers.

She also argues that this would benefit teachers in disadvantaged schools most, attracting teachers to disadvantaged schools is a problem, because their students have the greatest room for improvement.

The Arguments

I note at the outset that I agree with Minister Bishop about the need to increase the pay range for teachers. I have also long felt that there needs to be more flexibility in pay structures. However, this will not be provided by the uniform pay performance structures that she is proposing that simply substitutes one rigid uniform system for another.

Turning to the evidence, I have not read the overseas studies but would be cautious about them. Here I am grateful to Bruce MacEwen in his recent review of The Halo Effect, by Phil Rosenzweig, a professor of strategy and management at the International Institute for Management Development in Lausanne for making a core point, the difference between correlation and causation.

This is a major problem in education studies where many different variables come into play. The fact that there is an apparently statistically significant relationship, if only at 4 per cent, between maths results in Arkansas schools and schools with performance pay does not of itself say anything about a causal linkage. We don't know about student demographics, we don't know about school resourcing, we don't know about broader school management.

So without actually having read the studies I remain cautious.

The next thing is Minister Bishop's objectives. Here I think that we can put the arguments about the need to increase the pay range aside because this is really a separate objective that can be dealt with in several different ways independent of Minister Bishop's proposal.

Minister Bishop wants to raise the standard of Australia's public schooling. However, what does she mean by this? What are or should be the objectives of schooling? How does this link to public education? This is an area where I have written a fair bit recently simply because I am worried that with our obsession with quantifiable measures we are loosing site of broader educational objectives including especially developing the capacity to think.

Turning now to her score card measures.

The first thing to note is that performance is to be measured in individual terms. It is true that good teachers get better outcomes even in bad schools. But educational performance at student level depends upon a wide range of factors including increasingly popular team teaching, facilities and school management.

If we link this to my earlier discussion we can see three things. The first is that the individual focus may work against team performance, especially for team contributions not directly reflected in individual results. The second is that payment or non payment of the bonus may, in some cases, depend upon factors totally outside the teacher's control.

The second thing to note is that the first leg is totally focused on exam performance. I have daughters who have just gone through/are going through the NSW year 12 Higher School Certificate. In this case assessment results are made up 50:50 of continuous assessment and the final exam.

Again this worries me because I have formed the view at a personal parental level that the results obsession is creating undue pressure and damaging the standard of education. The last thing I want is to teachers give a direct incentive to increase the pressure on student to perform in the HSC or, alternatively, to drop out.

Two examples to illustrate what I mean.

Take a reasonably bright student who is doing okay and simply does not need to get maximum exam results to do what he or she wants to do. That student may prefer to take the opportunity to enjoy other aspects of school life including clubs and sport. But this means that he/she is not going to show maximum improvement in results, thus putting teacher pay under some pressure.

Take a not so bright student who is unlikely to do well but who is enjoying the school experience. There may now be an incentive to force that student to drop out.

I am sure that the second leg in the pay assessment process - subjective assessments by principals, parents, students and other teachers - is intended to offset the exam bias. However, this variant of the now popular 36o degree assessment process has its own problems.

I have only had limited direct experience with 360 degree assessment processes in a small number of larger clients. Introduced by HR people keen to try the latest, they simply did not work. A key problem was lack of clarity in just what was to be achieved.

I think that this presently holds true for Minister Bishop's proposal. Again speaking as a parent, I simply do not want to be involved in an assessment process that will directly affect the pay of individual teachers.

In all, the Minister has some distance to go to convince me that her current proposals make much sense.