Quantifying The Continuing UK Public Spending Squeeze

By Simon Wren-Lewis, February 11, 2024

Simon Wren-Lewis is Emeritus Professor of Economics and Fellow of Merton College, University of Oxford. He Was One of the Seven Members of the UK Labour Party’s 2015-2016 Economic Advisory Committee.

In her October budget the Chancellor raised total government spending by 1.8% of GDP compared to the plans of her predecessors, which means that total public spending as a ratio to GDP stays pretty flat over the 5 year forecast period. (Anyone who calls this the share of government spending in GDP is either being sloppy or deliberately misleading. [1]) As a result, that ratio is planned in 2029/30 to be roughly where it was in 2022/3, the penultimate financial year of the Conservative government.

Just before the budget, I calculated that spending to GDP needed to be about 3% higher than this to ‘end austerity’, by which I meant return public service provision to a level similar to the final years of the last Labour government. However the method I used to get to that figure was based on some unfamiliar data sources, so I thought it might be useful to redo that calculation using more familiar definitions. This post does that.

What you cannot do, but which so many do, is look at the total ratio of government spending to GDP and draw conclusions from that alone. Invariably such exercises equate a constant level of this ratio over time with a constant provision of public services, but this is completely wrong. For example, total government spending includes the interest the government pays on its debt, but it would be silly to suggest that the public services have improved since the pandemic because interest rates on that debt have significantly increased. The ratio of total public spending to GDP was unchanged between 2007/8 and 2016/17, and as a result I still get some people telling me that therefore austerity over this period was a myth. Clearly we need to look at some level of disaggregation.

Before I do so, it is worth noting that it isn’t easy to get hold of historical time series that break down the public spending total into its components, including how much each department has received over time. The ONS do publish a quite detailed breakdown by spending function, but for individual years rather than as a time series. Thankfully the IFS has done much of the legwork here, and I’ll draw on this source extensively. Those who want to avoid my analysis and jump to my conclusions can skip to the text after the final table.

Even if we disaggregate total spending into its major components, it is still true that keeping some part of public spending as a proportion of GDP constant over time is not necessarily keeping the level of public provision constant. This point can perhaps be made most clearly by noting that as GDP grows the share of food purchases in total GDP tends to fall, but it would be daft to report this as implying that as economies grow people are getting hungrier. The two items of UK public spending that have shown the clearest trends in the past are health and defence. Here, from the IFS document referenced above, are those shares over time.

Defence spending is now almost a quarter of what it was in 1955, reflecting both the abandonment of the UK’s imperial role and the end of the cold war. Obviously this does not mean people feel less secure than they used to. Health spending as a share of GDP has been steadily going up over time for a number of reasons. To the extent that this is because the population is getting older and therefore needs more care, it does not imply that we are on average healthier but just that life expectancy has increased. A clear example of this point is that during the decade after 2010 the share of health spending in GDP only fell slightly, but waiting times for treatment increased much more dramatically, correctly suggesting the level of public provision of health services had deteriorated substantially.

The chart below compares the public spending ratio with and without health and defence spending, and they look quite similar. However this is largely a coincidence, and critically the future looks very different, because probably both ratios will need to rise in the future. As a result, since the start of the century when defence spending stopped falling, and for the foreseeable future, we should expect to see the total public spending ratio (and therefore for the share of total taxes in GDP) to rise over time. I apologise for going on about this, but this basic point is hardly ever made in mediamacro commentary.

Of course health and defence are not the only areas of public spending where spending as a share of GDP may rise or fall over time even though public provision remains the same. Another example would be the implications of baby booms for education spending. However the upward trend in health spending is so substantial that some analysis is required. We cannot just fit a trend line through the data, because for much of this period health provision has been too low. Health spending is unlikely to be too high very often because raising taxes to finance it is painful, but Conservative governments since 1979 have squeezed spending. For this blog post I’ve tried to get a handle on this by looking at total (public and private) health spending in other countries using OECD data.

In almost every OECD country where data is available from the early 1970s there is a pronounced upward trend in the GDP share. The trend varies across countries as we would expect, but an average across 19 OECD countries implies that health spending to GDP rose by 0.11% each year, with roughly an equal number of countries above and below this figure. (The UK number is 0.145%.) As the pressures leading to this ratio rising are common across most countries, it seems reasonable to use this 0.11% figure as a measure of the underlying trend in the health to GDP share.

All that remains to decide is when UK health spending was sufficient. I have chosen 2010, because waiting times for treatment at this point were tolerably low. This allows us to measure the spending gap in other years, and also to project this trend forward in assessing the adequacy of future plans. Of course this approach is very rough and ready. It may also be possible to reduce this trend in the future by increasing the amount of preventative care (see the discussion here), but that is not going to happen when health is underfunded because treatment will always get prioritised.

Finally I have used data from the IFS to compare GDP shares and ratios across categories of spending between 2007/8 and 2022/3. [2] The earlier year is a useful reference point because it is before the recession that followed the financial crisis (recessions tend to raise spending in some areas) and is not affected by subsequent austerity. As I have already noted, the current planned total spending ratio for 2029/30 is similar to 2022/3.

The chart above suggests health spending in 2022/3 was around 0.5% of GDP below what it needs to be to replicate 2007/8 levels of provision. Actual provision was undoubtedly worse in 2022/3 because of the hangover from the pandemic. Hopefully that will have dissipated by 2029/30, but equally the underlying upward trend in health spending will mean that spending will need to rise by just over 0.5% between 2022/3 and 2029/30. On this account the total public spending total needs to be 1% higher in 2029/30 to replicate 2007/8 levels of spending on health grounds alone.

Turning to education, allowing for changing student numbers still suggests that public provision in 2022/3 was significantly below 2007/8 levels. This assessment seems to roughly correspond with media reports of inadequate provision (see here for example). Turning to other social security, levels of deep poverty in 2022/3 were similar to 2007/8, but spending has fallen, suggesting another significant fall in public provision here. (The two child limit needs to end, for example.) Current problems in the provision of public order and long term care have been well documented. (Here is a very recent IFS report on justice.) If we allow for a public spending gap worth around 0.5% of GDP in each of these four sectors, then adding in health implies we need the total public spending ratio to be 3% higher in 2029/30 than current plans to replicate the level of public spending provision seen in the final years of the last Labour government. [4]

A 3% gap just happens to be the same number I calculated before the Budget using a rather different approach. Any increase in the defence spending share needs to be added to that. This analysis confirms that Labour’s spending plans remain substantially inadequate if the aim is to return public provision to levels seen in the final years of the last Labour government.

Of course current plans only go five years ahead, and Labour may be in power for longer than that. It was only in the second term of the previous Labour government that public spending increased substantially. There are various reasons why the political situation is similar and different to then. My more substantive concern is that the Chancellor shows no sign of having done the kind of analysis shown above, but I am happy to be corrected (in confidence) if I am wrong about this. [5]

My analysis suggests that repeating Labour’s commitments on the main areas of tax going into the next election would be very foolish, but I fear the political pressures to do so will be great, and that this government will succumb to those pressures. There is a nasty feedback loop here. The more public service provision fails to noticeably improve over the next few years the more Labour will feel it needs to make unrealistic pledges on tax to win the next election. One way of avoiding that was to have increased taxes by more in the last Budget.

Footnotes

[1] Calling this ratio a share implies that the rest of GDP is the non-government share, but it’s not because total government spending includes transfer payments.

[2] Data for overseas aid is for 2021/2 rather than 2022/3.

[3] There are many areas of public spending not included in this list. For example summing each column gives 32.4% for 2007/8 compared to 40.3% for total public spending. The total public spending ratio in 2022/3 was 44.8%. Compared to 2007/8 debt interest explains about half this increase, but I would really like to know what explains the rest.

[4] The level of debt interest in 2029/30 is expected to be only a little below 2022/3 levels, reflecting an assumption that interest rates will not fall back to levels seen in the 2010s. If this does turn out to be the case, higher debt interest should imply higher taxation rather than lower public spending. There is no reason why higher interest rates should shift voters preferences between publicly and privately provided goods, so saying that other public spending should fall to match higher debt interest is as absurd as saying spending on food should fall by this amount. Higher taxes allows the cost of additional debt interest to be spread across all areas of consumer spending.

[5] Attitudes within the Treasury, and in particular a belief that it is up to departments to make the case for higher spending, do not help here. Such an approach often precludes a common overall framework and leads to outcomes that can have more to do with the character of individual politicians than social needs and preferences.

——————