With budgets under pressure, governments around the world are trying to get more from fewer public sector workers – the perennial productivity issue. A number have resorted to introducing ‘high-powered’ incentives into the public sector.
- In some cases this has been achieved by simply privatising service provision;
- In other cases government has introduced higher-powered incentives within the public sector.
Examples include pay for performance for teachers and doctors (Atkinson et al. 2009; Lavy 2009; Gravelle et al. 2010).
Perhaps oddly, rather less attention has been focused on the use of higher-powered incentives for the delivery of services that are closer to private sector administration – for example, the payment of welfare benefits or the collection of taxes.
There are reasons to expect such schemes not to work. These include poor measurement of output and public sector motivation being driven out by the use of extrinsic incentives (Dixit 2002; Besley and Ghatak 2005; Burgess and Ratto 2003). Moreover robust evidence on the use of high-powered incentives within the public sector remains pretty scant (Bloom and Van Reenen 2010). There are exceptions: a bonus scheme for tax collectors in Brazil increased tax receipts. But the Brazil scheme (Kahn et al. 2001) is an exception and the bonuses paid under that scheme were very large, amounting to over 50% of annual salary. Most governments either will not contemplate such generosity or are limited by union pressure not to reward so handsomely. It is thus an important question to ask whether schemes that pay relatively small bonuses to individuals in the public sector who are undertaking private-sector-type tasks actually work.
One government that did embrace such a scheme was the UK government between 1997 and 2010. Despite being a left-of-centre party, Tony Blair’s government embraced a strong belief in the use of market-like incentives within government. The administration sought to introduce schemes as incentives to several agencies, including those that collected taxes and paid cash benefits. Not all schemes were undertaken, owing primarily to union resistance, but some schemes did operate. These included two incentive schemes aimed at indirect tax collectors and those responsible for helping unemployed people find work.
The two schemes shared a number of key features.
- They were directed at frontline employees doing clerical tasks.
- Employees were rewarded for achieving a number of targets. Several targets were selected to avoid employees focusing on a narrow set of tasks at the cost of others.
- The targets tended to vary in their measurement precision: some were measured precisely using routine management data (e.g. the number of unemployed people finding work), while others were less well measured and often survey-based (e.g. customer satisfaction).
- The bonuses paid for hitting the targets were not huge but neither were they trivial. In total these could be as much as 5%-7.5% of annual salary.
- Both schemes were team-based, so that employees were rewarded for team, rather than individual, performance. Any bonus was only paid if the whole team hit the target. Teams tended to be defined by grouping together all workers within an office or a group of offices. So instead of being a small production unit, in some cases teams were quite large.
- Finally, both schemes were introduced as trials, allowing performance comparison between those with an incentive and those without. This allows researchers to capture the effects of the incentive schemes more robustly than is often the case in government schemes, where political necessity to show results or reduce equity challenges leads schemes to be introduced in all government offices without a control group.
Good news and bad
Our evaluations of these schemes bring some good news and some bad. Under both schemes some teams hit their targets and were awarded bonuses. So relatively small bonuses did lead team members to increase their output. And the costs of achieving the output gains were smaller than a general pay rise to achieve the same gains.
But the news was not all good. In an evaluation of the scheme for the collection of indirect tax (VAT), we use high frequency management data from the agency to look at how employees hit their targets (Burgess et al. 2010). We find that teams that hit their targets and got their bonuses did so primarily by reallocating the more productive workers to incentivised tasks and away from the tasks that gave no rewards, rather than by greater effort of the members of the team. This might be good for short-run outputs but in the long run those tasks from which these more productive employees were shifted will suffer.
Our evaluation of the scheme in the agency placing unemployed people into jobs looked at free-riding within teams. In that scheme the rewards were defined at district level (Burgess et al. 2012). Districts were composed of several offices and varied in terms of number of offices within a district and number of individuals within an office. Analysis of the success of these schemes showed substantial evidence of free-riding. Only incentives in smaller teams were successful, while the introduction of incentives in larger teams in fact reduced output. The same scheme also allowed evaluation of whether how precisely outputs were measured mattered – we find that it did. Outputs that were measured using routine data collected at high frequency improved, while ones that were measured imprecisely at lower frequency did not.
In sum, the news for policymakers is that well-designed schemes can work and may be more cost effective than a more general pay rise. This may be part of a smarter approach to public-sector pay setting (e.g. Propper and Van Reenen 2012). The evidence also suggests that team-based rewards give opportunities for managers to achieve their targets by smarter assignment of their staff, and that the use of team-based pay in the public sector can, just as in the private sector (Holmstrom 1982), induce free riding. So incentives matter, but their design is critical.
Atkinson, A, S Burgess, B Croxson, P Gregg, C Propper, H Slater, and D Wilson (2009), “Evaluating the Impact of Performance-Related Pay for Teachers in the UK”, Labour Economics, 16(3):251-261.
Besley, T and M Ghatak (2005), “Competition and incentives with motivated agents”, American Economic Review, 95(3):616-636.
Bloom, N and J Van Reenen (2010), “Human resource management and productivity”, NBER Working Paper 16019 and forthcoming in O Ashenfelter, and D Card (eds.), Handbook of Labor Economics, Vol. 4.
Burgess, S, C Propper, ML Ratto, S von Hinke Kessler Scholder, and E Tominey (2010), “Smarter task assignment or greater effort: what makes a difference in team performance?”, The Economic Journal, 120(547):968-989.
Burgess, S and ML Ratto (2003), “The role of incentives in the public sector: issues and evidence”, Oxford Review of Economic Policy, 19(2).
Burgess, S, C Propper, ML Ratto, and E Tominey (2012), “Incentives in the Public Sector: Evidence from a Government Agency”, CEPR Discussion Paper 9071.
Dixit, A (2002), “Incentives and organisations in the public sector: an interpretative review”, Journal of Human Resources, 37(4):696-727.
Gravelle, H, M Sutton, and A Ma (2010), “Doctor behaviour under a pay for performance contract: treating, cheating and case finding?”, Economic Journal, 120:129-156.
Holmström, B (1982), “Moral hazard in teams”, Bell Journal of Economics, 13:324-340.
Kahn, CM, ECD Silva, and JP Ziliak (2001), “Performance-based wages in tax collection: The Brazilian tax collection reform and its effects”, Economic Journal, 111(468):188-205.
Lavy, V (2009), “Performance pay and teachers’ effort, productivity, and grading ethics”, American Economic Review, 99(5):1979-2011.
Propper, C and J Van Reenen (2010), “Can pay regulation kill?”, Journal of Political Economy, 118(2):222-273.