Be pragmatic: ask not ‘whether’ but ‘how’ to adopt public sector performance-related pay
So should the public sector adopt PRP? This is the wrong question. A better question is: can we design performance related pay to work well in the public sector? If we can’t, then we shouldn’t adopt it because it is too risky. But, if we can, then we should not be confined by ideology. We should take a pragmatic approach and adopt PRP where it delivers better public services.
There seems no reason why the public sector cannot in principle benefit from some PRP. It works in the private sector. And there are lots of similar jobs in the public sector, such as electricians and porters, even though there are also many other roles, such as teaching and fire-fighting, that are largely particular to the public sector.
That said, there are differences between the two sectors. Many public sector employees work in teams, often across department boundaries, where cooperation, rather than rivalry, is essential. They tend to have many bosses, including professional associations and managers. Their organisations also tend to have rather fuzzier goals involving values such as equity and fairness. In comparison, life is simpler in the private sector, where value maximisation is king.
The case for introducing PRP in any particular instance in the public sector should be down to practical considerations. Is there evidence that it will result in increased productivity? What’s the best design – what would be the most effective incentives?
As far as evidence is concerned, there is a problem: it is difficult to demonstrate causality in this field. Additionally, there is little robust evidence on the impact of PRP in the public sector, even though it employs as many people as does manufacturing.
For example, in education, there is mixed evidence – in some countries, such as Israel, research shows that PRP for teachers delivers better academic results with few side effects, whereas US evidence is more mixed. A scheme in England, albeit poorly designed, showed substantial improvements of about half a GCSE grade per pupil.
In healthcare, there is no consensus in favour of PRP. However, some recent studies show very promising results. Again, among public servants, there is only a little robust causal evidence. A Jobcentre Plus scheme based on team performance had a substantial positive effect in small offices, but a negative effect in larger centres. At HMCE (now incorporated into HMRC), a team-based approach did raise individual tax yields; the biggest factor explaining the improvement was reallocation of the most efficient workers to those teams by the (incentivised) managers.
Where does this leave us? We know incentives work in many sectors and with many different types of workers. But we should remember that people will do what you pay them to do. They will follow the scheme – they think that is what you want. So getting the design right is vital – matching the scheme to production, picking the right incentives. It may be necessary to rethink a public sector organisation when introducing a scheme, so that people have fewer tasks and it is easier to incentivise them than if they are multi-tasking.
An important development for the public sector, if it is to emulate the successes achieved in the private sector, is to ensure that its goals are clearer. If public sector organisations were less ‘fuzzy’, it might be easier to incentivise their staff with PRP than is often the case at the moment.
Simon Burgess is Professor of Economics & Director, Centre for Market and Public Organisation, University of Bristol. He was a keynote speaker at the PIRU event, ‘Payment-for-performance: incentivising quality in public services,’ on 19 February 2013 at LSHTM.