Commentary by Allyson Pollock et al misrepresents the findings of economic analyses of quasi-markets says Julian Le Grand. Looking at the evidence (and recognizing the defects of state agencies’ administration of healthcare) shows that quasi-markets with fixed prices perform better. Competitive mechanisms in the NHS were also supported by previous Labour reforms.
In their recent blog on this website, Allyson Pollock, Alison Macfarlane and Ian Greener are highly critical of the research of my LSE colleagues, Zack Cooper, Alistair McGuire and Stephen Gibbons on the impact of competition in the NHS quasi-market. As Henry Overman has noted, most of these criticisms are either misdirected, or are already answered in the papers themselves, or have been carefully rebutted already. I was not an author of the papers concerned; but elsewhere in the Pollock and colleagues blog, there are specific references to me that Overman does not deal with, and that do require attention.
Of these the most significant is that I (with Cooper) ‘sweep aside decades of careful economic theory and evidence that which shows why markets do not work in health services and distract the reader from the facts that their work is ungrounded and far from empirical’.
The theoretical and empirical literature on market failure in health care is indeed large. However, most of it relates to what we might think of as ‘full’ markets with private finance and private provision. There is much less on ‘quasi’-markets: these are competitive markets, but where the state in some form provides the finance and there is competition between a diversity of providers, private, public, non-profit, etc. Such quasi-markets are common in health care, including many social insurance schemes in continental Europe, Medicare in the United States – and, since 1991, the English NHS.
Now there are theoretical problems with quasi-markets – mostly involving patient information. But there are theoretical problems with state systems as well –mostly involving the absence of market incentives. The question then becomes an empirical one: which performs best – or least worst – in practice. And here, contrary to Pollock et al’s assertions, most of the evidence tends to support the quasi-market, especially when prices are fixed. Simon Burgess and Carol Propper from the University of Bristol’s Centre for Markets and Public Organisation surveyed the evidence for the United States and other countries on hospital competition, finding that competition with fixed prices both reduced costs and increased quality. There were other, more specific, findings, including one that the effect of competition was to give more appropriate treatments, with sick patients in less competitive markets receiving less intensive treatment and having worse health outcomes than those in more competitive ones.
Closer to home, these results tie in, not only with the results of Cooper and colleagues, but with those from the comprehensive set of evaluations of the Blair Government’s market-oriented health reforms reported in the Kings Fund publication Understanding New Labour’s Reforms to the English NHS. They concluded (on p. 131) that ‘the market-related changes introduced by New Labour tended to have the effects predicted by proponents and that most of the feared undesirable impacts had not materialised to any extent’.
I was an adviser to that New Labour government and helped develop and implement those reforms. The reforms were not as pervasive as I would have liked and, perhaps in consequence, the gains from the reforms were not as large as I would have liked. But, as the evaluations showed, there were gains and these were positive. Even more significantly, the apocalyptic consequences predicted by the reforms’ critics – critics who were almost as hysterical as those currently criticising Andrew Lansley’s reforms – were not realised.
So both theory and evidence, both past and present, indicate that a state-funded healthcare system that incorporates market-type incentives will save more lives and reduce more suffering than one that does not. And that is why I support competition within the NHS.
Please read our comments policy before posting.
There are two issues that are being mixed up in this whole debate. The first issue is whether the recent research on choice and competition is any good. All research can of course be picked apart. I can pick apart any article, including all of my own. The likes of Cooper, Propper etc. undertake careful analysis. They probably do just about as well as they can with the methods and data available. The second issue is whether academics should actively engage to shape policy on the basis of only limited evidence. They shouldn’t. They should seek to inform, in a non-advocative way, but not to make more claims for their research than their research deserves. People on all sides in this debate are guilty of this. The debate is ideological, consensus is impossible, and no-one really knows whether or not they are right. My take is that there is not enough evidence to suggest that changing the NHS in the way that is planned will improve it, and there are significant risks that it will do much harm, in all sorts of ways. It is also unwise to change it significantly in a period of significant resource constraint, particularly when the NHS has, on the whole, improved substantially over the last ten years.
My understanding is that a theory of itself cannot show anything: it can only make predictions which can be tested by experiment and observation.
Economic theory is non-specific in the summarising title of this post, but Professor Le Grand posits that whatever this theory is, it does make two predictions: “state-funded healthcare which incorporates market-type incentives will save more lives” and “state-funded healthcare which incorporates market-type incentives will reduce more suffering.”
The inclusion of a comparative in both predictions is unhelpful. (More than what?), but it seems safe to assume more than “state-funded healthcare which does not incorporates market-type incentives.” That, however, remains unclear because of the ill-defined “market-type” and “incentives” (what incentives, and for whom?).
To a non-specialist like myself, we seem to be in a debate where a change was implemented from theory (not evidence) and there is inevitably a resulting difference of opinion as to whether that change caused the predicted consequences (or even whether they happened at all).
Of the two predicted consequences “reducing more suffering” would appear a difficult measure to quantify by observation against the introduction of “market-type” incentives. I know when I am suffering but have only a subjective sense of how much so, and it would be exceptionally difficult to measure a reduction of suffering across a population.
Reduction in deaths does seem a grandiose claim. Does the evidence really show that when all other factors are removed, the introduction of “market-type incentives” really achieved statistically significant reductions in the number of deaths? How much “market-type incentive” was required?
Again, I am by no means a specialist in this area, but as a sceptical thinker I tend to suspect the post-title unproven. It looks like what we expect from governments, unfortunately: an ideological approach derived from another discipline (in this case, economics, which it seems is notoriously unreliable) imposed without robust evidence from scientific trials
Ian makes some very valid points.
Indeed, I was recently reading an article by Professor Propper in the Winter 2011 edition of centre for market and public organisation research in public policy
http://www.bris.ac.uk/cmpo/publications/bulletin/winter11/propper.pdf
Professor Propper writes:
“Quality in hospital care is very difficult to measure, so to date researchers
have only examined crude proxies measured at the hospital level.”
She then concludes at the end of the article by saying:
“The evidence base for choice is small and the extent to which choice has been used is quite limited. But this is not an argument for a return to ‘command and control’: rather it is a call to try to understand where exactly choice may work and where it may not. The White Paper may offer that opportunity.”
The Jury therefore is still very much out. Perhaps this is a good example of confirmation bias – that befalls us all, (academics too) – which Daniel Kahneman, the Nobel Prize Winning Economist engagingly explains in his latest book, Thinking, Fast and Slow.
I have liked and admired Julian Le Grand’s work for years, and he is on many my teaching reading lists. But I do have to wonder what is going on here.
There is a basic category error in the piece above – competition doesn’t save lives. It never can. Clinical workers save lives, hopefully with the support of managers. To simply assert (as economists are inclined to do) that competition changes incentives is to not engage with the difficult, messy, empirical work of exploring exactly what has changed and how (if anything, and the moral of reorganisation in the NHS is how little changes). So the key question here is what exactly changed after 2006, and how did it change both managerial and clinical behaviour. The answer, I’m afraid, is that we really don’t know.
Even reviews of research that consider the econometric work that Cooper and Propper favour on their own terms find there are big gaps (for example, Bevan and Skellern in the BMJ (http://www.bmj.com/content/343/bmj.d6470). Most particularly, their research is based on a ‘black box’ where incentives and changes are assumed rather than being empirically demonstrated.
It is also interesting that Le Grand cites the King Fund’s work as supporting his claims. Again, that is very contestable, as David Hunter’s review of the book, again from the BMJ (http://www.bmj.com/content/343/bmj.d7786) suggests. My reading of the book was that Labour’s market-based improvements were extremely modest when compared to what was achieved elsewhere. It therefore seems odd that Le Grand presents this work as supporting him, when I’m really not sure that’s what the book is saying.
Above all, it seems to me that it’s time we started being a bit more humble about our work. Le Grand and Cooper, in their FT piece, seemed to be suggesting that there is only one way to do research, and people who do other kinds of work are mere ‘intuitionists’. This reads to many of us as an intolerant and rather blinkered view of social research. In order to understand what is going on in the NHS we need a range of different methods and different types of work. To imagine we can provide some kind of definitive answer, as they appear to be suggesting, from work that doesn’t empirically examine whether the changes they believe are going on are actually happening, without actually asking anyone involved in their implementation, seems a little odd.