"We are spending record levels on the NHS" says every Conservative minister who is challenged on the performance of the health service under the Conservatives. Fine, but which government has ever actually cut spending on the NHS? Both major parties promised large-sounding increases in spending on the NHS in the 2019 election, with Labour proposing an even larger-sounding amount. I think most people, beyond a dull sense that the NHS probably needs more money, have no sense of what these spending increases might mean, how much is "enough", or why the NHS always seems to need more money. I don't, and I am an economist. This also, I think, leads people to the conclusion that the NHS is an inefficient "black hole", or that immigration is putting too much strain on the health service so that no amount of increased spending can alleviate the stress.
In this post, I propose to explain ways of measuring healthcare "effort": how much, compared to need, the UK spends on healthcare. I will use this to evaluate claims about whether healthcare spending is "enough", and whether it is sustainable.
Start with the headline figure: total spending on the NHS in the UK. This seems like a natural measure of how much resource is being put into healthcare, but there are two basic problems with this measure. Firstly, a pound in 2019 is not the same as a pound in 2015, because of inflation. Secondly, if the number of people in the country goes up and spending on the NHS does not, then the spend per person has gone down, and we would expect services to get worse.
Inflation may well be a particular problem in healthcare, and tends to be somewhat higher than in other areas of the economy. The principle theoretical reason for this is the Baumol effect: because healthcare is (still) labour intensive, productivity gains are slower. However, since other sectors are increasing their productivity, wages must rise or nobody would want to work in healthcare. Therefore, healthcare must contend not just with the general rise in prices, but by a rise in the real unit labour cost. This is one reason why developed countries find themselves having to spend more on healthcare just to deliver the same standard or care. There is no way around it though, until robots get as good as humans at diagnosing illnesses, and patients get used to the idea of robotic nurses.
There is also the problem of population growth to contend with. However, UK population growth is not neutral with respect to healthcare needs: the UK also has an ageing population. People who might have died of cheap illnesses such as heart attacks or accidents end up, thanks to innovations like statins, better road safety, and healthier lifestyles, now live long enough to need care for more expensive illness like coronary artery disease or cancer (Alzheimer's being dealt with predominantly by the social care sector). So again, more spending on healthcare, even in real (inflation-adjusted) terms, is required just to keep the level of healthcare constant.
Naturally any measure of healthcare effort is controversial, but I propose the following as good measures:
Real spending per capita: this takes spending, deflates it by the GDP deflator (the measure of inflation for the economy as a whole) and divides by the number of people. This is a rough-and-ready measure which adjusts spending for the fact that prices go up and the population goes up.
CPI-adjusted spending per capita: a criticism of the GDP deflator is that it overestimates the loss of wellbeing as a result of price changes because it assumes no substitution. As an example, suppose that the price of white bread goes up by 10% but the price of brown bread does not. The GDP deflator effectively assumes that the hit to consumers' wellbeing will by 10% multiplied by the percentage of their income they were spending on white bread. But some consumers will like brown bread nearly as much as white bread, and switch to this. CPI is a measure of inflation which is theoretically much closer to the utility loss associated with price changes because it takes this into account. I present both here because personally I am sceptical of the scope for substitution in the healthcare sector. Frankly, if the price of insulin goes up, a diabetic can't do much substitution.
Healthcare cost-adjusted spending per capita: this is as before, but takes the healthcare component of CPI only. In theory, this should account for the differential cost of providing NHS services over time. In normal times, we would expect this to be higher than CPI, although the public sector wage freeze may have had some impact in holding down healthcare inflation in recent times. This is theoretically a better measure, but I want to present both in order to allow for a comparison: we want to see how much the inflation which is specific to healthcare is eroding the effective spending.
Healthcare cost-adjusted spending per pensioner: here I use the number of pensioners as a crude measure of need in order to correct for the fact that population growth has not be agnostic with respect to the need it gives rise to. This relies on the assumption that the healthcare needs of the average pensioner have not changed over time, which doesn't strike me as a more heroic assumption than that the healthcare needs of the average person have not changed over time.
We have a range of measures here, ranging from the most generous to the argument that healthcare spending has increased to the least generous. These are plotted together below. I use the "Expenditure on Healthcare" series from the ONS to do this. Theoretically, "UK Health Accounts" is superior, but this series as published only goes from 2013, and there are technical reasons why it is not straightforward to convert one to the other. Primarily, the latter series also includes all social care spending, but conversion would require adjusting for the small amount of social care spending which is centrally funded. More on the implications of this later. I calculate each series as described above and index them to 1 in 2010 in order to facilitate comparison. In order to extend the series to 2019 I allow healthcare spending to increase by 3.4% per year as per the OBR's assumptions and subsequent revisions and inflation measures to increase up to the end of 2019 at the average rate for 2019 as a whole.
The general picture is more-or-less as we would expect. Adjusting spending by health-specific measures of inflation gives us a smaller increase in healthcare effort, and adjusting by the percentage of the population over 65 shows a marked decrease in healthcare effort since 2010 (nearly 10%). One surprise is that inflation as measured by the GDP deflator appears to be consistently lower than as measured by CPI. This implies that consumers' baskets in general have increased significantly more than for the economy as a whole over the period.
This gives us a provisional answer to the question, "why does the NHS feel as though it is underperforming, even with increased funding?". Based on the above, it is primarily to do with an ageing population. Were the percentage of the population over 65 not increasing over time, spending on healthcare in the UK would be broadly flat in relation to need, but increases have not compensated for the increased need from an ageing population. This gives some support to Boris Johnson's claim that the NHS is well-funded but demand is increasing, but not to his implication that all we need to do is cut immigration and everything will be fine. Immigration turns out to be roughly neutral here, as the demographic breakdown of immigration to the UK turns out to be broadly in line with the UK population as a whole (with about 20% over 65), and it is not the increase in population per se which is the problem.
Two issues remain. Firstly, as noted above, the series presented do not include social care, which is predominantly provided by local government. Cuts to local government will therefore have an uneven impact on social care budgets across the country, and these will have a disproportionate impact on the NHS as its resources are used inefficiently to cope with increased demand for which it is not well designed (A&E wards should not be dealing with elderly patients who fall because the resources were not available to make the requisite adjustments to their lifestyle to prevent excessive risk-taking). Partly, this is a limitation of the dataset. However, even were we to include social care spending in the data, this would not account for the decreased efficiency with which public money is used when it is reallocated from social care to emergency care.
Secondly, even on the most optimistic measures, it is undeniable that healthcare effort increases have more or less stalled since 2010. Some might argue this is a positive, given the difficulties governments appear to have across the world in restraining healthcare spending. In fact, the NHS is comparably excellent at constraining spending because it has a central budget which governments can squeeze. But there is another interpretation of the persistent increases in healthcare spending: as countries get richer, the returns to higher and higher spending decrease. It might be very nice having one holiday abroad per year, but having five isn't five times as good; even the most die-hard petrolhead won't get that much pleasure out of her fifth Ferrari. On the other hand, it is almost impossible to enjoy one's life unless you are alive and healthy. Given this, it might be rational to think that as societies get richer, they ought to devote a larger and larger share of national income to health and healthcare. Indeed Hall and Jones calibrate the extent of this and suggest that even the astronomical increases in healthcare spending in the US might be optimal from a standard economic perspective. Given this, the stalling of the increase in healthcare spending would be a significant welfare loss, as UK patients lose out on the efficient application of new medical technology.
So in conclusion, why does it feel like the NHS is underfunded despite claims are large spending increases? The answer is that higher spending increases would have been required to compensate for the UK's ageing and expanding population. This does not mean that the NHS is a "black hole". It means that a government that wants to keep its population healthy needs to increase spending by more than it has done. And, as discussed, it is probably optimal to have spending increases even greater than this in order to take advantage of the potential welfare gains of new medical technology and economic resources.