A recent lecture at the ESRI by Canice Prendergast, who is now one of the world’s foremost researchers on workplace incentives and their impact on productivity, addressed this question.
Below is a summary of the main ideas from the lecture.
What is Pay-for-Performance?
Pay-for-performance would include wage raises, the likelihood of getting promoted, the likelihood of a better job, subjective bonuses, profit shares, stocks, commissions, and so on.
Does pay-for-performance work?
There are a series of academic studies over the last decade showing large effects of pay-for-performance on productivity and profitability.
It is not unusual to see productivity increases of 25-35%, with increased compensation accounting for perhaps half of this.
Sales workers, sports players, teachers, and many other occupations all show gains when commission, piece rates, bonuses, etc. are used rather than salaries.
A key study was in a windscreen replacement firm, Safelite, where productivity rose 35%, of which a third was due to selection of better employees.
On the other hand, there have also been a number of studies suggesting that aligning pay to performance is likely to backfire so badly that it should not be used at all.
Examples include the case of police officers in Los Angeles some years ago. The police were seen as slow to respond to complaints about public policy matters. So a performance measure regarding ‘complaints’ was instituted. However, it backfired, as the police stopped arresting criminal suspects (who when arrested made spurious complaints) and so arrests went down and the police response to complaints improved!
Another problematice situation was in the US under the ‘No Child Left Behind’ scheme whereby teachers were paid for the number of children in schools they brought over the line i.e. above the 40% pass mark.
The result was that teachers concentrated their efforts on children they could bring ‘over the line’ and ignored the rest (those who would pass anyway and those who had no hope of making it).
Another bonus that backfired in the US involved paying police 20% of the proceeds from drugs raids. The result was the police stopped arresting people with drugs, and only arrested those with money and drugs, as that is what they were paid for.
These studies identify two issues 1) measurement issues, and 2) psychological responses to performance pay.
The central questions become whether the assumed increase in productivity is (i) worth the increased compensation costs, or (ii) beneficial to the organisation, even ignoring compensation costs.
Pay Crowding Out Intrinsic Motivation
Studies have shown that Extrinsic Motivation (e.g. pay-for-performance) crowds out Intrinsic Motivation where employees do work that they inherently enjoy.
If you pay one group to do that task for some period of time, but do not pay the other to do it, and then see what happens after the payment period finishes.
Who continues to do the activity? Typically, it is the group that is not paid.
Why pay-for-performance should not be used?
A: Performance pay may indeed change behavior, but not in the way that the employer would like.
B: Performance may not improve enough to warrant the additional compensation costs.
Performance Appraisals – Can we trust performance evaluations done by supervisors?
Another problem with bonuses is that supervisors can end up giving everyone the same bonus in order to keep them all happy, even though performance can vary greatly from one team member to another.
Performance Measures – Alternatives to pay-for-performance
Performance measures sometimes only cover a subset of the activities of the employee.
Do you know what the employee should be doing?
Can you monitor inputs rather than outputs?
Can you observe the employees’s actions?
Pay-for-performance incentives have to be evaluated in the context of answers to the questions above.
For example, one of the main reasons sales staff are paid on a performance basis is because they are away from the office and it is used as a way of ensuring they will perform.
McDonalds, for example, pays its franchisees several times more than in the outlets it runs directly as the latter are easy to monitor.
What else can be used when pay for performance fails?
Some things you can do include, job design and better hiring.
So, overall there are pros and cons for pay-for-performance systems.
Be careful about what you measure as ‘what gets measured gets done’, but is it what you intended?
And remember that employees are intrinsically motivated as well as by external factors such as performance related pay.