Cost of the Market in our NHS

Caution … this is a dated opinion piece – there is ongoing work in this area.

We appear to be edging to a consensus that even with heroic “savings” our care system needs significantly more funding.  In an era of austerity and tax resistance that is hard but has to be discussed.  There are no magic solutions.

But sadly we have had yet another rounds of claims that £10bn pa could be saved if the market was removed from the NHS.  Yet again the only basis for the claim is one line in a Health Committee report – in fact misquoting what was actually said. The claim is that the introduction of the internal market increased NHS administration costs from 5% to 14% of total expenditure; so removing the market will save 9% or around £10b.

The source for the claims about 5% and 14% is the 4th Report of the Health Committee, Session 09/10 on Commissioning.  The following extracts are from the report.

According to the official historian of the NHS, Dr Charles Webster, the service has traditionally scored highly on account of its low cost of administration, which until the 1980s amounted to about 5% of health-service expenditure.


An estimate of administrative costs made by a team at York University concluded that management and administration salary costs represent, as a very crude approximation, around 23% of NHS staff costs, and around 13.5% of overall NHS expenditure.

Two things should immediately be pointed out.  The quote about 5% is from an excellent book, A Political History of the NHS by Charles Webster, but the passage in the book does not itself have any references to where the 5% came from or what it actually contained.  It is almost certain that in the era pre 1980 many tasks which might now be characterised as “administration” or “management” were only done as part of a wider job and so would not have been recorded in any way.

And the York University Report – NHS Management and Administration Staffing and Expenditure in a National and International Context, from March 2005, time and again sets out that comparison of costs between countries and between periods in our own NHS are beset with many issues around classification.  So for example the 14% did not include any “estimate” of consultants and others time which might be classed as administration or management.

In fact the report actually put its estimate of administration and management costs on an internationally comparable basis at between 17% and 21% (not the 14% as is often used).

As the York report sets out:-

There are no agreed definitions of ‘administration’ and ‘management’ in health care between (and sometimes even within) countries’ health care systems. Substantial ambiguity exists around any comparisons, particularly as definitions shift as groups of workers are recategorised. Consequently, all cross-national and cross-sectoral figures must be viewed with extreme caution.

Even if we had reliable and comparable figures (and we don’t) then arguing the whole of any increase between the cost base in the 80’s and the costs base in 2003 (the base year for the York study) was due to the internal market and that this was wholly without any compensating gains is not justified.

In a paper which uses the same base information (the 5% and 14%)  Colin Paton[1] suggested that only half of the increase was due to the internal market – although he gave no rationale for the 50% figure.

A recent, thorough and rigorous analysis of the actual evidence about the impact of the setting up and running of the internal market is set out in Reforming Healthcare – What’s the evidence? – by Ian Greener and others in 2014[2]. The change was costly to implement and added significantly to management and administration costs, but the various papers referenced in What’s the evidence? do not suggest additional costs to anything like the scale of the £10bn claims.  No proper costs vs benefits study has ever been carried out.

Even if the additional costs introduced by the changes 25 years ago could be quantified none of the claims about £10bn savings actually set out what costs could now be reduced.  Reducing management and administration costs by £10bn would be a reduction of over 50% and these costs account for around 25% of NHS jobs (depending again on what is included).  A 50% reduction equates to many tens of thousands of jobs.

So far as can be ascertained given all the definitional issues we can see that the NHS since the 70’s has changed dramatically in terms of the role and significance of “management”.  The rise in the use of IT for collection and analysis of cost data occurred not because there was a market but because it was agreed that the health system needed to have some elements of management.  We did not need to know how long the queues were when they were not managed and we did not collect outcomes or activity data and nobody would have used the results.  Managererialism, not marketization, probably accounts for most of the rise in administration costs.

There is one recent study which might give us a better idea about what might actually be saved from removing the internal market. Information about the costs of various types of health systems has recently helpfully been provided by a study by Himmelstein et al[3].  This looks only at the costs of management and administration within larger hospitals across various countries with varying degrees of “market”.  Of particular interest is that it treats Scotland and Wales separately from England in its analysis.

Like the York report it needs to be read with some care but to summarise; the highest costs per capita (after numerous adjustments) are in those systems with complex payment systems.  Systems which have single payer and block funding have the lowest costs.  Systems like ours in England are in between.  For hospitals the best estimate for England is for administration and management to account for between 17% and 21% of total expenditure.  (Spookily but coincidentally similar to the York finding of a decade ago.)  The study suggest that a base level of around 12% is necessary for any system.

In comparisons of relative expenditure England comes out slightly higher than Wales but Scotland is significantly lower.  The explanation though appears to be a technical one in that management of capital in Scotland is more centralised.  In fact there is not a lot of difference between the three nations.  What differences there are appear to be due to costs of administration not of management and to be due to numbers of people not wage levels.

In terms of quality the most recent comparison across the UK from the Nuffield Trust[4] showed no great outcome differences between the various systems even though they had very different structures.

But there are cost differences between systems which can be seen mostly to do with the costs of information used to drive the payments systems, which in England is DRG and activity based plus a bit of block funding. If we went back to very simple single payer funding with no competition for funding, no “commissioning” and a centralised management structure looking after major issues, shared services and capital it appears that there could be savings of the order of 1% to 2% in hospital costs.  (That assumes that much of the information used to drive payments systems will have to be used anyway.)

In England there are also commissioning costs and system management and regulation costs.  Opinions vary about what we would have to do in terms of planning if there is no market and about how much system management and regulation would still be needed.  But we could envisage savings from the £1.2bn spent by CCGs plus spending on CSUs, and much of the regulatory infrastructure.  Maybe £2bn in all.  But much of what is actually done within these bits of the NHS still has to be done somewhere. We do need to know how much various things cost and how variable outcomes are and we do need someone designing pathways and we do need some kind of oversight. That can be done in many ways.

Anyway nothing even at the wildest extremes of what might be possible gets to savings of even one quarter of the claimed £10bn.  Still the mythical £10bn is so firmly entrenched the facts are unlikely to be of much use.

When the Health & Social Care Act is repealed and the competitive market is removed then there will be scope for savings but not of the order of £10bn.  Further savings could come from reducing the number of NHS organisations through consolidation, but merges and other transactions have a bad track record.  We can only “guesstimate” but this might over time be of the order of 1 – 2% of total English NHS expenditure, but there would be considerable transition costs to be met and neither change would be easy and consolidation would be contested!

[1] Centre for Health and the Public Interest – At what cost? Paying the price for the market in the English NHS; Calum Paton

[2] Reforming Healthcare – What’s the evidence? – Greener, Harrison, Hunter, Mannion and Powell – Policy Press Bristol – 2014

[3] Health Affairs, 33, no.9 (2014):1586-1594 A Comparison Of Hospital Administrative Costs In Eight Nations: US Costs Exceed All Others By Far; David U. Himmelstein, Miraya Jun, Reinhard Busse, Karine Chevreul, Alexander Geissler, Patrick Jeurissen, Sarah Thomson, Marie-Amelie Vinet and Steffie Woolhandler

[4] The four health systems of the UK: How do they compare? Professor Gwyn Bevan, Marina Karanikolos, Jo Exley, Ellen Nolte, Sheelah Connolly and Professor Nicholas Mays. The team was led by Professor Nicholas Mays of the London School of Hygiene and Tropical Medicine. Nuffield Trust, London, 2014.