Can increased public spending pay for itself?
Since the election, we’ve seen several proclamations of the ‘end of austerity’. As TUC post election poll (among others) has showed, there is now a clear majority in favour of ‘Maintaining decent public services even if that means my taxes go up’.
This is a far cry from some of the rhetoric during the election campaign, when talk of ‘magic money trees’ was used to dismiss anyone calling for higher spending.
But it’s important to recognise that there are two separate debates about how to fund public spending happening here. First, whether the public are willing to pay higher taxes in order to fund additional spending. And second, whether an expansion in public spending financed by borrowing, could, in fact, lead to higher economic growth, therefore ‘paying for itself.’
The second of these arguments found some surprising support during the election campaign from Julian Jessop, the chief economist at the right leaning Institute for Economic Affairs. His title ‘why there really aren’t any magic money trees’ detracts from a more subtle position – that there is a case for government consumption and investment expenditures to boost growth, but a rejection of doing so at present.
In this post I set out the reasons why the IEA were right to argue that spending can boost growth, but wrong to suggest that now isn’t the time. I approach the argument on the technical (but not difficult) territory of (1) the multiplier, (2) whether spending pays for itself, (3) issues of capacity and (4) ‘crowding out’.
1. The multiplier
The reality or otherwise of the ‘magic money tree’ boils down to whether (and by how much) government spending will strengthen the economy.
The IEA blog recognises “there may be times, for example, when the economy could benefit from a fiscal stimulus financed by a temporary increase in borrowing”.
The ‘multiplier’ determines the strength of the boost to the economy from each pound of additional spending. The IEA argue that:
A fiscal multiplier of 1x, for example, would mean that a £100bn increase in public spending financed by borrowing would be expected to increase GDP by £100bn. Many studies have found that fiscal multipliers are usually below 1x during expansions but between 1x and 2x during recessions.
The controversy over the size of the multiplier effect has been downplayed here. We know from recent experience that the extent of the boost we could expect to economic growth from any additional investment goes beyond ‘recession’ and ‘expansion’.
With the worst of the financial crisis over and the economic decline arrested, from 2010 policymakers (across the world) began to make cuts in government spending. But the detrimental impact of these cuts on national economies was greater than initially expected, and international organisations quite quickly recognised that multipliers – that is, the size of the impact of cuts on the economy – were underestimated.
The case for the multiplier was first set out by Keynes and his collaborator Richard Kahn. They argued that the multiplier could be calculated by estimating the marginal propensity to consume (mpc) – that is, the extent to which any additional pound earned will be spent in the economy. An adjustment is made for the marginal propensity to import (mpm), in order to look at the impact only on domestic goods and services. The equation looks like this:
Multiplier = 1 / (1 – mpc + mpm)
In 2009 analysis (here), I estimated that this this leads to a figure of around 1.5 for the UK, with the mpc around 2/3 and the mpm around 1/3.
Therefore the multiplier = 1 ( 1 – 2/3 + 1/3) = 1.5
This conforms reasonably well to an assessment of the damaging effect of spending cuts on economic growth in the UK since 2010 (more detail is set out here). Looking at the impact of spending cuts across all OECD countries suggests an average multiplier closer to three (here). But in the UK the OBR has resolutely stuck to estimating the size of the impact of fiscal policy on the economy using multipliers of one (on investment spending) and less than one (on everything else).
A distinction between investment (capital) and other (current) spending is often made. And it is sometimes claimed that only capital spending offers the possibility of increasing growth. But this is to misunderstand the nature of the multiplier process – at least as understood in traditional Keynesian economics. The multiplier process operates through increasing aggregate demand. The increased employment and/or income resulting from government spending puts extra money into people’s pockets which is spent, so boosting activity across the whole of the economy. It doesn’t matter how that money is put into worker’s pockets. So long as there is capacity (see section 3) increasing nurses’ salaries will be beneficial to the economy. Infrastructure spending may have additional supply-side effects, strengthening the economy going forwards. This is all well and good, but these additional supply side effects are not necessary in order to argue that increasing spending has the ability to boost growth
2. Spending paying for itself
The full story of the ‘magic money tree’ follows because if spending has a multiplied effect on the economy, it might pay for itself. Any new economic activity will mean higher government tax revenues and reduced expenditure on welfare that together offset at least some of the original outlay on expenditure.
Julian offers a ‘hat tip to Professor Keynes’ for seeing spending will strengthen the economy. But he goes on to contest Ann Pettifor’s claim that …
“an increase in spending on public sector wages could be entirely self-financing, because of the resulting boost to economic activity and tax revenues”.
… as follows:
For this to be true, the fiscal multiplier would have to be implausibly high (at least 3x, assuming a tax/GDP ratio of 35%). And if this were the case, the government would have nothing to lose from awarding 100% salary increases across the board.
Here the idea of a ‘fiscal multiplier’ conflates the multiplier effect on the economy and then the effect on the public finances. But the basic point is right: revenues will be changed by applying the present share of government revenues in the economy to the increased level of activity. Though the IEA omit the reduced spending on welfare payments (housing benefit, tax credits, unemployment benefit and the like) which are around 7% of GDP (excluding pensions). On the basis of the multiplier and these simple ratios, Julian’s example of a £100bn stimulus would increase GDP by £150bn, taxes would be raised by £52.5bn and welfare spending reduced by £12.5bn. So the shortfall between the spending and revenue changes would be only £37bn. And note that the break-even multiplier would be around 2.4 not 3. But even at face value this is a far cry from the OBR calculations which effectively have no net revenue change.
But there are other effects that are likely to compound the beneficial effect. First any increase in the deficit as a share of GDP would be moderated, because GDP would be higher thanks to the aggregate multiplier effect. Second, in an ideal world, any stimulus would be as part of a global stimulus and UK exports would be advantaged as other countries’ imported more. Third – and most important – the whole point of fiscal initiative is to revive confidence in the economy more generally, so that businesses are confident to expand investment (which is currently flat-lining).
Julian also neglects to point out that so far Pettifor has been proved right but in the reverse direction. (See the Economic Consequences of Mr Osborne, co-authored by Professor Victoria Chick from UCL and myself.) The damage caused by cuts has meant greatly reduced tax revenues and increased welfare expenditure than would otherwise have been the case, so that the public finances have not been repaired. Despite the eighth years of spending cuts, deficit reduction projections have been missed consistently, and the public sector debt ratio has continued to increase not fall (see here for more detail).
Public debt as % GDP
Source: ONS & OBR
A second argument made against increased fiscal expenditure to stimulate economic growth is that the economy is at capacity –that is, it cannot grow any faster because we have reached full employment. Under this view any efforts further to expand activity will lead only to inflation and damage to the economy. Julian is far from alone in claiming “the UK is now close to full employment” – i.e. is at capacity.
The obvious point is that much of the work created over the past five years is insecure work, and we are in the longest pay squeeze in living memory, as the IEA recognises:
At this point, somebody usually observes that the current high levels of employment have only been achieved at the cost of sluggish growth in real wages. This sluggishness is partly due to Conservative policies (including the ongoing public sector pay freeze), as well as the drop in the pound following the Brexit vote. But the fact remains that there are now many fewer people looking for work and little spare capacity in the economy as a whole.
But ‘somebody’ has a genuine point. The notion that there are fewer people looking for work does not necessarily mean that there is little spare capacity. The proposition that unemployment around 5% is equivalent to full employment may be widely shared but that doesn’t make it a fact. For around 25 years after the war unemployment did not rise above 3% and inflation was modest in most years. We are now in uncharted waters. The coincidence of feeble economic growth, high employment growth and low wages should mean that we need to review ideas of capacity, At the very least, most recognise that the role of insecure work means we do have to review our ideas of full employment. For example, Michael Saunders’ of the MPC’s January 2017 speech [link] on ‘the labour market’ made the case for going (a little) below 5 per cent.
[More technically] The argument that the economy is not yet at capacity is re-enforced by considering GDP growth in its own right. At the start of austerity, the OBR’s June 2010 forecast was for growth of around 2.5% a year. Instead growth has been around 2% a year. The way the OBR operates is effectively to write off this lost growth as never to be recovered. Because they do not accept that cutting government demand is damaging the economy (which follows from their low multipliers), they are then obliged to regard the weakening of the economy as structural, i.e. as the result of underlying (supply-side) flaws in the economy, rather than stemming from a lack of demand. (See here for fuller discussion and in particular the role of productivity on Royal Economics Society website).
If, as I argue, the multiplier is significantly higher than the OBR recognise, then there is idle economic capacity sitting there waiting to be used. There is therefore the prospect of good work that is presently being denied as a result of the estimate that policymakers have chosen.
4. Crowding out
A further argument often made against expanding public sector activity is that it will ‘crowd out’ private sector activity. In general terms this means drawing workers, money and investment into the public sector deprives and so damages the private sector. Clearly this is the antithesis of the idea of expanding public sector activity to support the private sector.
Commonly this is refined to the idea that public sector borrowing will ‘crowd out’ and therefore lead to higher interest rates on private sector borrowing. Outcomes since the start of the crisis categorically refute this notion. Even while the cuts failed to reduce borrowing to a sufficient extent to halt rises in the public debt, interest rates have fallen to historic lows.
Under present conditions central bank purchases of government debt under the quantitative easing programme, effectively mean that the Bank of England (and other central banks) are acting as a buyer of last resort for government debt. But the same mechanism could have protected (if necessary) against any rise in interest rates in the context of an increase in loan-financed expenditure. But overall under the spending approach the public debt would be far lower than it is now, and any (perceived) dangers of crowding out diminished. Plainly this argument is contingent on the higher multiplier, but inevitably all these arguments are inter-related.
(Whether quantitative easing is necessary to support loan-financed expenditure is a separate question, discussed here.)
5. In conclusion
To close, the IEA worry that the pendulum might swing too far the other way – i.e. we might enter a dystopia of reckless spending. This still seems unlikely, notwithstanding the shift in debate following the general election. And to be clear just ahead of the general election the OECD confirmed that the present policy still constitutes ‘consolidation’, i.e. cuts.
This issue really matters. For years public services have seen their budgets cut and public sector workers have had their pay reduced. But if the argument above, first put forward by Keynes et al, is right then this is not only needless but senseless. The correct way to improve the public finances is instead to increase rather than reduce government spending.
Two years ago Simon Wren-Lewis – who has consistently opposed austerity policies from academia – wrote:
My own best guess would be that the multiplier has been larger than one … but I have never suggested that I know with certainty what the size of the multiplier has actually been. However there has, to my knowledge, been no public debate on these terms.
Given the fundamental importance of the multiplier, this is a bizarre state of affairs and does not reflect well on public debate. Perhaps finally things are turning around. Ironically the IEA point the way to the right way of thinking about the impact of public spending on the economy and public finances. This approach would permit a more fruitful discussion of whether now is the time to invest in infrastructure and public services. Given the terrible wage performance, rise of insecure work and seemingly unceasing weak and unbalanced growth, this is a discussion that is long overdue.