Public sector performance: Introduction

The performance of the public sector affects us all. There are at least three reasons why we should be interested in how well it functions: it is big; its outputs are special; and it is getting bigger. The public sector represents a sizeable portion of most industrial economies. In the UK public sector current expenditure represented 38.5 per cent of GDP in 2004/51 (HM Treasury, Public Finances Databank) and it is projected to rise to 39.6 per cent in 2006/7. A malfunctioning public sector affects overall economic performance both through its direct contribution to the goods and services produced by the nation and through its indirect impact on other sectors through the provision of a well educated and healthy workforce, quality infrastructure (both physical and institutional) and so forth.

The public sector has been getting bigger in recent years. Real public sector current expenditure in 2004Ð5 was over one-quarter higher than in 1998Ð9, after several years of negative or zero growth. The country's citizens are entitled to ask their government whether this extraordinary increase was worth it.

Finally, the outputs of the public sector are special. Health, education and policing affect the whole of society and its more vulnerable members in particular. Often the private sector cannot be relied upon to produce these outputs efficiently and/or equitably. But the very reason why these services are provided by the public sector makes their management and the assessment of their performance problematic.

Three of the papers in this issue of the Review come out of the first phase of the Economic and Social Research Council's Public Services: Quality, Performance and Delivery' Programme and all four were presented at NIESR's fourth Public Sector Performance conference. They cover a number of topics, from the general issues relating to specifying targets in the public sector to the usefulness of specific techniques to public sector performance assessment, and illustrate them with examples from three rather different areas: the health sector, local government and tax offices.

Gwyn Bevan's article considers the problems faced and created by a central agency laying down a set of targets and levels of performance that are identical across delivery units. He considers this in the light of the recently deceased (or at least transformed) ‘Star ratings' system for primary care trusts. As well as creating a framework for considering such systems of central control, Bevan provides useful insights into the way the system evolved that come from being involved in the process itself Ð having lead responsibility at the Commission for Health Improvement for agreeing the basis of star ratings for 2003 with Secretary of State Alan Millburn. He concludes that, whatever faults the system had, it did at least deliver reductions in waiting times. Nevertheless, it is perhaps a system that has had its day and he sets out a ‘mixed' alternative to the central-command model and considers the issues this raises for governance of health care performance.

Philip Stevens, Lucy Stokes and Mary O'Mahony stay in the health sector with star ratings, but shift their application to acute hospital trusts in England. They argue that the many attempts at performance measurement and assessment in the public sector have often been created on an ad hoc basis, with little consistency. Some targets are set without a clear idea of how to measure them, and others are designed for one purpose (for which they may or may not be fit), but are pressed into alternative uses. In order to create a framework in which to set the various measures, Stevens et al. outline a typology of performance indicators and a set of desiderata for them to be useful. In this context, they compare the outcome of a performance management system with an alternative measure Ð a productivity measure analogous to those used in the analysis of the private sector Ð and find that the two are almost totally unrelated to each other. Whilst the narrow focus of the star ratings system might explain this, it does raise questions as to the appropriateness of such indicators of performance.

Dirk Haubrich and Iain McLean shift the empirical focus to another composite indicator of performance: the comprehensive performance assessments (CPAs) for local authorities in England. In a study that echoes the call in Gwyn Bevan's paper to consider the specific circumstances of organisations, they appraise three different studies that have investigated the link between an important influence on local authorities' provision Ð deprivation Ð and their performance under this measure. Haubrich and McLean use the natural experiment created by Scottish and Welsh authorities to contrast their more flexible approach to the assessment of local authority performance to the rather more prescriptive English system. Interestingly, early evidence suggests that far from relishing the flexibility of their systems, stakeholders in Scotland and Wales hanker for the certainty and clarity of the CPA.

In the final paper, Finn Forsund, Sverre Kittelsen, Frode Lindseth and Dag Fjeld Edvardsen provide a very concrete example of the analysis of public sector performance using an important development of a common assessment technique. Data Envelopment Analysis has now become a common technique for the assessment of public sector organisations which is particularly appealing to analysis because it places so little structure on the technology of provision. However, a crucial assumption is that the data are measured with certainty. Forsund et al. show that once one accounts for potential random variation in the data, the ability with which one can make practical comparisons of organisations is considerably reduced, although broader categorisations into ‘best practice' and low performers are still possible, which may be enough for the technique to be useful for policymakers.