The ‘knightly’ assumptions of ‘trust and altruism’ (T&A)

Writing in the 18th century, both David Hume (Reference Hume1785) and Adam Smith (Reference Smith1976) were keenly aware of the need to design models of governance to tackle problems of knavish behaviour. However, in the United Kingdom, following the scandal of the Crimean War, the Northcote-Trevelyan reforms to British government to tackle corruption (Greenaway, Reference Greenaway1985) produced a model of governance that Hood (Reference Hood1995) described as progressive public administration (PPA), in which administrators were assumed to be a ‘Jesuitical corps of ascetic zealots’. Hood sought to outline and explain how the new public management (NPM) differed from PPA, and why the take up of NPM varied by country. Le Grand, who is one the most influential advocates of quasi markets, which is a key policy in the set of approaches known as NPM, highlighted that, in the United Kingdom, before the election of the Thatcher government in 1979 (Le Grand, Reference Le Grand2003: 5), successive governments assumed that all who worked in government and public services had become ‘knights’:

Although in England, as described below, this assumption has been rejected, in the other countries of the United Kingdom, this assumption of ‘knightly’ behaviour seems to have remained as the default model of governance of health services (Bevan, Reference Bevan2014). This model of T&A assumes that as all are knights there is no need for normal incentives in which failure results in losses and success in gains. This model includes what Oliver describes as ‘reporting general performance’. The corollary of the T&A model is that, as any failure cannot be due to want of trying on the behalf of the ‘knights’, it can be due to external causes only, and it therefore follows that failure ought to result in extra resources. There are two fundamental problems with the T&A model: it tolerates wide variations in performance, and, where failure is perceived to be rewarded, results in perverse incentives. These can be argued to have corrosive effects on real ‘knightly’ behaviour: as those who work hard to deliver high performance see their efforts being ignored, and those who make little effort and deliver poor performance being rewarded.

The model of ‘choice and competition’ (C&C)

The initial serious policy response to tackle variations in provider performance that were outcomes of the T&A model was the alternative of C&C. This model is intended to create external incentives through a quasi-market system in which there is patient choice of providers and money follows the patient. Both Barber (Reference Barber2007) and Le Grand (Reference Le Grand2009) saw this model as having the greatest potential to deliver high performance in public services. This model assumes that users choose better performing providers, and that providers respond to the consequences of these choices on their market shares. However, a paradoxical empirical finding from systematic reviews by Marshall et al. (Reference Marshall, Shekelle, Leatherman and Brook2000) and Fung et al. (Reference Fung, Lim, Mattke, Damberg and Shekelle2008) of ‘reporting general performance’ is that patients did not use information on quality of hospital care to switch from poor to good hospitals. Hibbard (Reference Hibbard2008) offers three explanations for this finding: consumers are largely unaware of differences in quality, have different criteria from those used in public reporting and find it difficult to understand the complex information that is presented. Bevan and Fasolo (Reference Bevan and Fasolo2013) argue that in order for this model to work, patients need advice, as shown by the London Patient Choice Pilot (Coulter et al., Reference Coulter, Maistre and Henderson2005).

The models of ‘naming and shaming’ (N&S) and ‘targets and terror’ (T&T)

The third model of N&S is described by Oliver (Reference Oliver2015) as ‘Public ranking with apportioned blame’. Hibbard et al. (Reference Hibbard, Stockard and Tusler2003) showed N&S to be more powerful in a controlled experiment in a hospital in Wisconsin than the models of T&A and C&C. Marshall et al. (Reference Marshall, Shekelle, Leatherman and Brook2000) and Fung et al. (Reference Fung, Lim, Mattke, Damberg and Shekelle2008) did, however, find that systems of public reporting of performance did sometimes have the effect of improving provider performance, but this varied across systems. Hibbard (Reference Hibbard2008) argued that the explanation for this variation is that the outcomes of public reporting depend on its design. They argued that for a reporting system to be an effective means of N&S it needs to be organised as a ranking system, published and widely disseminated, easily understood by the public (so that they can see which providers are performing well and poorly) and followed up by future reports (that show whether performance has improved or not). They argued that the power of N&S comes from its impact on the reputation of those publicly reported to be performing poorly even though they know this will have minimal impact on their market shares. One way of putting this is that those who have enjoyed the esteem of being regarded as ‘knights’ are deeply troubled by being publicly ‘named and shamed’ as ‘knaves’.

The N&S model tends to be used alongside other models: the US Veterans Health Administration (VHA) combined N&S with pay for performance (P4P) for its senior managers (Oliver, Reference Oliver2007); a justification for the publication of school league tables of examination results was that this was important information for parents in the C&C model and the ‘star rating’ regime that applied in the English NHS from 2000 to 2005 combined N&S with T&T (Bevan and Wilson, Reference Bevan and Wilson2013). Although Oliver also describes ‘Public ranking with apportioned blame’ as the fourth model of ‘T&T’, it is helpful to use that model for where those responsible for running organisations are held to account against a limited set of targets that clearly signal priorities with strong threats of sanctions for failure and rewards for success. Oliver (Reference Oliver2015) favours the way the US VHA made strenuous efforts to win support from clinical staff by involving them in the choice of indicators. Although this makes sense where there is enthusiasm for such change, I would argue, however, that this did not apply in the English NHS in 2000. The problem the government faced was that the seemingly intractable problem of long waiting times was deemed as acceptable by those working in the British NHS, indeed it might be argued that it had become one of its defining characteristics. Under these circumstances, the government was right to combine the massive increases in funding with ambitious targets for reducing waiting times that were part on the new models of T&T and N&S. This combination made it clear that the rules of the game had changed for chief executives: failure would result no longer in being rewarded with extra funding of their hospitals, but with the threat of being fired and their hospitals being ‘named and shamed’ as ‘failing’. The emphasis on T&T in ‘star ratings’ was part of a shock therapy and seems to have shifted to N&S after the first two years.