Skip navigation.

The least worst way to improve public services

Julian Le Grand


First published in Renewal Vol. 11, No. 2 (2003)


If the defenders of the public realm are to convince us, then they must provide us with evidence.


It is now commonplace to observe that the improvement of public services is the Government’s greatest domestic challenge. It has to succeed in levering up quality in areas such as education and health care; otherwise it will lose the support of the broad coalition that elected it, from working-class party activists to middle-class professionals. It is also commonplace to assert that it has so far failed to do so – at least in the perception of many pundits and, probably, also of many readers of Renewal. But is this perception accurate? And to the extent that it is, why? And how can the Government do better?

To some extent, the perception of failure is correct. Despite the sharp increases in resources going to the NHS, there has been a dramatic fall in productivity, with the result that activity is stagnating. Hospitals and other NHS trusts are struggling to meet waiting time targets, fiddling the numbers if all else fails. Schools seem to be having a hard job meeting exam performance targets. The professionals and others working in the services make regular claims of demoralisation. Perhaps most crucially, the wider public do not seem to feel that services have got better.

This perception of public service failure is leading to widespread calls from the left for the Government to deflect from its chosen courses in these areas. The argument is increasingly couched in terms defending the so-called ‘public realm’. The Government, ostensibly the guardian of this realm, has apparently violated it through an obsession with targets, numbers, and inspections. Worse, some parts of government appear to be captivated by the private sector, including the infamous private finance initiative. And even in areas from which the private sector is currently barred, such as running clinical services in NHS hospitals, there are moves afoot to promote patient choice and competitive independent institutions like foundation hospitals.

All of these, the critics argue, should be swept away. Boundaries should be set around public services so as to prevent their being infiltrated by alien methods and values. Teachers, doctors, nurses, as professionals, should be allowed to get on with their jobs. There should be no targets, no league tables, no apparatus of monitoring and inspection. Private sector providers should not be permitted to provide public services (although an exception might be made for some non-profit organisations or mutuals). Competition and quasi-markets should be replaced by collaboration and networks.

At its most extreme, this is a plea for a return to the status quo ante – the situation that existed before the quasi-market reforms in public services of the 1990s. In the pre-1990s NHS, doctors had complete clinical freedom. They decided who they wished to treat, what treatment to give them and how much each patient should get. In education, while local government officials determined school admissions policies, teachers had widespread freedom over curricula and teaching methods. The 1960s and 70s were described as the ‘golden age of teacher control’.

But the reality was far from golden, in either education or health care. In the National Health Service, patients had to live up to their appellation and be patient. They had to wait patiently in queues at GPs’ surgeries or at outpatient clinics; if they needed further treatment, they had to be prepared to wait their turn on lengthy hospital waiting lists. When the time arrived for them actually to go to hospital, they were supposed cheerfully to accept being on a public ward, being served horrible food and being treated by doctors too busy, or too elevated, to have time to explain what was happening to them. Similarly, the parents of children in state schools were expected to trust the professionals, and to accept that teachers knew what was best for their children. Moreover, as with the NHS, especially following the comprehensive reforms of the mid-1960s, parents were supposed to concur that the overriding objective was greater equality and hence to accept whatever degree of uniformity of educational provision that attaining this objective required.

So the defended public realm did not serve us very well in the pre-1990s welfare state. Moreover, the critics of the types of policies currently being pursued by the Government have ignored some real success stories. In particular there is evidence from both health care and education that competition in public services can work to drive up standards and increase efficiency.


Improvements in schools

Take primary schools, for instance. My LSE colleague Howard Glennerster (2002) has examined the performance of pupils in national attainment tests over time and reached some striking conclusions. First, the percentage of pupils reaching a given level of achievement in England steadily increased from 1995 (the first full year of tests) to 2001, especially at the end of primary school. Some of the improvements were remarkable; for instance, the percentage of pupils gaining the expected level of competence in maths at the end of primary school moved from 45 per cent in 1995 to 70 per cent in 2001. This should be set against the fact that the best available evidence suggests that there was no improvement in the maths skills of children in the early years of secondary school for thirty years prior to 1995.

Even more importantly, these improvements were not confined to good schools. In fact, the lowest-performing schools in 1995 were the ones to show the greatest improvement by 2001. The same is true of schools ranked according to the wealth of the area; over the period schools in poor areas were catching up with schools from rich ones. The polarisation that many critics aver is the inevitable consequence of competitive pressures did not materialise.

Of course, we must be cautious in interpreting these figures. There are anecdotal stories of ‘teaching to the test’ and indeed of outright fraud. Unpublished results from the University of Durham (2002) found little improvement in reading skills for year six pupils from 1999 to 2002 and in vocabulary between 1997 and 2002. However, they assess only reading, whereas the national tests include assessments of writing, spelling, and handwriting as well as reading. Moreover, the same team did find a 5 per cent improvement in science and a 9 per cent improvement in maths between 1997 and 2002. In international comparisons, too, England has moved up the league tables, outperforming most other European countries in science and maths.

So what was the source of these improvements? It was not increased resources: public expenditure per pupil in England increased only a little over the period at primary school level, and was virtually static at secondary school; public education spending as a whole actually fell as a percentage of GDP. This itself leads to another significant fact: not only were there improvements in outcomes over the period but, since these were achieved with little if any increase in resources, there was also an increase in the efficiency with which those resources were used.

In fact, there are a variety of possible policy instruments that could have contributed these improvements: the education inspection agency Ofsted, the publication of league tables, the numeracy and literacy hour introduced by the Government, parental choice of school. There is not much evidence to enable us to disentangle the separate effects of these. However, there is evidence that they all in different ways lead to increasing competitive pressures on schools: and that in turn helps to drive up education standards.

That competition can work to raise standards in education has been confirmed by research directly on that issue. The economist Stephen Bradley and colleagues at the University of Lancaster (2001) investigated the performance of all secondary schools in England over the period 1993–7, measuring the performance of each school on two dimensions of ‘output’: school exam performance and attendance rates. They then related these outputs to various input measures, including the proportion of pupils ineligible for free school meals and the proportion of qualified staff, to calculate the relative efficiency of the schools concerned. These were related to the extent of competition in the area, as measured by the number of schools of different types (including selective) within a two-kilometre radius.

The conclusions of the study were quite unambiguous. The relative efficiency of the schools was directly related to the amount of competition they faced. The greater the degree of competition, the more efficient they were. Moreover, the strength of the effect increased over time, corresponding to the evolution of competitive pressures. Competition also encouraged changes in efficiency over time: the greater the competition, the faster the rate of change.

It is worth noting that there is international evidence concerning the importance of competition in affecting school outcomes. Harvard economist Caroline Hoxby (2002, 2003) has reviewed the evidence on the impact on the performance of public schools in Milwaukee, Michigan, and Arizona of competition from ‘choice’ schools. All three of those areas have experimented with allowing parents to choose schools other than their local public schools through the mechanisms of either vouchers (Milwaukee) or charter schools (Michigan, Arizona). It had been widely predicted that public schools in the areas concerned would suffer an overall drop in performance as the choice schools ‘cream-skimmed’: sucking in better students and staff. However, Hoxby found evidence of strongly improved performance by the public schools, from which she concluded that the efficiency-inducing effects of competition were more than enough to offset any potential effects of cream-skimming. She also examined the effects of competition with private schools on public schools and of competition between public schools through parents choosing place of residence. Again she found that competition had a positive impact on performance.


Health care

The evidence concerning the impact of competition in health care is more varied. But again it suggests that the impact can be significant. One indicator of the efficiency in the way that the NHS uses hospital resources shows a rise in productivity during the early and middle years of the Conservative’s internal market, when competitive pressures were greatest (Le Grand, 2002). This petered out as those pressures weakened in the last years of Conservative administration. And when the internal market was abolished in 1997, efficiency began to fall – at an ever-increasing rate.

On a more micro-level, Carol Propper and colleagues at the University of Bristol have shown that competition between hospitals can reduce costs, and, more disturbingly, may also reduce quality (2003). However, they argue that the reduction in quality arose because purchasers ignored quality in their contracts, concentrating only on costs. The lesson they draw is that competition is indeed a powerful force for changing behaviour, but that it has to be properly harnessed to ensure desirable outcomes.



All methods of improving public services have their problems. The search is not for some theoretical error-free method of producing high quality, responsive services, allocated with perfect equity and produced at maximum efficiency. Rather, it is to find the least worst way of doing things: to produce as much by the way of quality, responsiveness, efficiency and equity given the limitations imposed by resources and human nature. And finding the least worst way cannot rely upon theoretical constructions or even romantic visions of a past golden age; rather, it requires the marshalling of evidence concerning the advantages and disadvantages of different ways of doing things. If the defenders of the public realm are to convince us that the government is on the wrong track, especially with respect to the usefulness of competitive pressures, then they must provide us with some evidence that their chosen path will do better (it would also be useful to have a clearer idea as to what that chosen path actually is). Until they do, their criticisms of current government policies cannot be taken seriously.



Bradley, Stephen, Johnes, Geraint, and Millington, Jim (2001) ‘School Choice, Competition and the Efficiency of Secondary Schools in England’, European Journal of Operational Research, 135: 545–68.

Glennerster, Howard (2002) ‘United Kingdom Education 1997–2001’, Oxford Review of Economic Policy, 18: 120–36.

Hoxby, Caroline (2002) ‘How School Choice Affects the Achievement of Public School Students’, in Paul Hill (ed.), Choice with Equity, Stanford: Hoover Institution.

Hoxby, Caroline (2003) ‘School Choice and School Productivity (Or, Is School Choice a Rising Tide that Lifts All Boats?)’, in Caroline M. Hoxby (ed.), The Economic Analysis of School Choice, Chicago: University of Chicago Press.

Le Grand, Julian (2002) ‘The Labour Government and the National Health Service’, Oxford Review of Economic Policy, 18: 137-153.

Propper, Carol (2003) ‘Death and the Market: the Questionable Impact of Competition on Health Care Quality’, University of Bristol: Leverhulme Centre for Market and Public Organisation Bulletin – Market and Public Organisation, Issue 8: 7-9.

University of Durham (2002)

Privacy policy