Winner of the New Statesman SPERI Prize in Political Economy 2016

Tuesday, 15 November 2022

Missing a fiscal rule does not make a black hole, and a financial crisis after tax cuts does not mean markets want spending cuts


Although the Autumn Statement has been postponed until this week, talk about black holes in the public finances continues unabated. I have always hated this term, which now seems almost de rigueur among economic and political journalists. A black hole in the public finances is just the amount by which the government is forecast to miss some fiscal rule. Connections between a real black hole and a black hole in the public finances is an empty set.

Why do journalists talk about mythical ‘black holes in the public finances’, rather than ‘forecasts of how much fiscal rules will be missed’? I suspect a headline involving the term ‘fiscal rule’ will turn many readers off, whereas everyone knows what a hole in a budget means, and calling it a black hole makes it sound serious enough to demand the reader’s attention. But this attempt to make the public finances sound like household finances is the essence of what I call mediamacro. [1] Unlike real budget holes, public sector black holes are only as good as the fiscal rules that give rise to them. [2] In the last week, thanks to groups like PEF and some good journalism, this fact has at last been given some media coverage.

While fiscal rules have advanced beyond the crude and harmful balanced budget at all times idea, there has been for at least a decade a serious disconnect between what politicians do and what sensible macroeconomics suggests. I wrote a short post about all this two and a half years ago, and will not repeat that here, but instead focus on the problems involved with two fiscal rules that may be the origin of black hole headline numbers: targets for the current deficit, and for a falling debt to GDP ratio. But before doing either, I need to discuss the overall macroeconomic situation. One of the reasons black hole language is an example of mediamacro is that it makes fiscal policy a matter of accounting, when it should always be about the health of the macroeconomy.

The macroeconomic situation

The big mistake Osborne made in 2010 was to try and achieve a deficit target while the economy was in a recession and interest rates were at their lower bound. In that situation deficit or debt targets should be thrown out of the window, and fiscal policy needs to take the place of interest rates in helping to achieve a strong recovery. Labour’s idea at that time, of trying to go less ‘far or fast’ with fiscal consolidation, would have been better than Osborne’s policy but still missed the key point. In a recession you need fiscal stimulus, and only when the recovery is complete do you start worrying about deficits. As generally in life, you do the important thing first, and the less important thing later.

The situation today is more complex than it was in 2010. We seem to be heading for a serious recession, but interest rates are well above their lower bound because the Bank of England thinks that is necessary to make the labour market less tight. That means that any immediate fiscal stimulus (that will impact demand next year, say) would be in danger of being counteracted by higher interest rates. Does that mean that some immediate fiscal consolidation, in the form of tax increases rather than spending cuts for reasons given here, to make a start on achieving a medium term target for the current deficit would be sensible as long as the Bank responds, as it should, by having lower interest rates?

This is where the complexity comes in. The Bank’s own forecasts suggest inflation will be going below target in 2025 or earlier, implying the Bank will be cutting rates in 2024 or earlier. Of course the OBR may come to a different view, but I doubt this as the reasons for a recession (higher interest rates today combined with a large income cut caused by higher energy and food prices) are pretty compelling. There has to be a clear danger that we may in a year or so’s time find interest rates heading towards their lower bound once again, and the economy will need fiscal stimulus rather than fiscal consolidation.

The Chancellor could do fiscal consolidation now, but then reverse this with fiscal stimulus when rates are at or near their lower bound. However such fiscal fine-tuning rarely works as intended. For example, we may experience a much deeper recession next year than the Bank is predicting, and that will only be intensified by any fiscal consolidation now. From a macroeconomic point of view the safer rule is not to tighten in a recession and to only raise taxes when the economy is healthy again. [Postscript A]

An impending recession also creates difficulties for the OBR and any fiscal rules. A good fiscal rule aims to hit some target for the current deficit (see below) when the economy is at trend. Normally forecasts have the economy going to trend after five years, which is why policy should aim to achieve a target in five years time. Yet if the economy is still at the bottom of a recession in 2025, will the recovery really be complete by 2027? If the forecast suggests not, then trying to hit a 5 year target makes no sense. [3] It would be far better to extend the forecast by enough to ensure the economy is expected to be running at trend for a few years at the end of that forecast, and delay fiscal consolidation until those years.

The government will almost certainly ignore these macroeconomic fundamentals, as they have mostly done since 2010. Even then, it could square immediate tax increases with good macroeconomic policy in two ways. The first is to raise taxes where the immediate demand impact is negligible, such as taxes on companies (like those producing energy, obviously), on the rich or on wealth. The second is to combine tax increases with larger increases in public investment. If fiscal rules constrain worthwhile public investment then they are poor fiscal rules. [4] Higher public investment is an excellent way of stimulating the economy, and is also necessary to rapidly green the economy.

Current deficit targets

One of the things George Osborne (or more likely Rupert Harrison and the IFS) got right in 2010 was to have a fiscal rule that focused on the current deficit rather than the total deficit. The difference between the two is net public investment (the current deficit + net investment = the total deficit). The big danger with targets for the total deficit is that the easiest way to achieve them in political terms is to cut public investment, but that is an anti-growth measure and a classic case of increasing the burden on future generations.

Targets for the current deficit just one or two years out make no sense, and my own view is that the best target is to set policy so the forecast meets the target five years into the future, as long as there is little chance of a recession where interest rates hit their lower bound in the meantime (see above). But what should that target for the current deficit in five years time be?

In the past governments have suggested a target of zero for the current deficit. This has the superficial attraction that the government is only borrowing to invest. However economists talk about the need for the public finances to be sustainable in the medium to long run, by which they mean that the ratio of some stock like public debt (or public sector net worth - see below) to GDP should be constant. If the government only borrows to invest (a current deficit target of zero), then this ratio will tend to fall over time because of nominal GDP growth. The higher the value of public debt, the more rapidly the ratio will fall as nominal GDP grows.

This means that having a target for a zero current deficit is arbitrary. What the target should be is a matter of economic judgement. Critical to that judgement is the observation that negative shocks to the public finances, like the financial crisis and the pandemic, tend to be larger than positive shocks. That means it is not very prudent to aim for, say, a constant ratio of some stock to GDP ratio, because in practice that will mean the ratio gradually rises following unexpected shocks. On the other hand many have argued that the climate emergency is exactly the time when it makes sense to let government debt to GDP rise.

Whatever your view, there are sound theoretical reasons for not planning on changing the debt/GDP ratio (or similar) too rapidly. In the current context I think a current deficit target of 1-2% of GDP would still be fiscally prudent. Any ‘black hole’ calculation based on getting the current deficit to zero in five years time is therefore questionable, even if we ignore the macroeconomic outlook. If hitting that target becomes the basis for not increasing spending on public health, it means the Chancellor is prioritising fiscal prudence over saving lives and expanding the economy. Fiscal prudence is just much less important than either saving lives or expanding the economy.

Falling debt to GDP?

Of the fiscal rules we have seen over the last decade or so, targeting a falling level of debt to GDP is one of the worst. The first reason is that It refers to government liabilities but not its assets. If households had a fiscal rule that their debt should always be falling they would never take out a mortgage to buy a house. Equally it is absurd to stop a government investing in worthwhile public projects just because this pushes up its debt, and the lack of public investment in what was until recently an era of very low real interest rates is one more failure to lay at the door of this government.

As I have already noted above, this point is recognised when we talk about flows, which is why we have targets for the current deficit rather than the total deficit. But that logic should be applied to stocks as well as flows. Targeting government debt relative to GDP, or its change, is inconsistent with targeting the current deficit.

The stock measure that is essentially the counterpart to the government’s current deficit is not government debt but the public sector’s net worth, which is the difference between all government assets and liabilities. This is because net investment funded by borrowing adds equally to the government’s assets and liabilities. The OBR has a nice working paper about public sector net worth, and its figure 2.3 makes this correspondence clear.

In 2020 general government net worth was large and negative according to the ONS, because debt greatly exceeded assets. The OBR can forecast the future level of public sector net worth just as much as it can forecast future levels of public debt. So why is all the talk about government debt rather than government net worth? Not only does this make no economic sense [5], having targets for debt rather than net worth can constrain otherwise socially beneficial public investment.

Should we therefore have targets for both the government’s current balance as a share of GDP and a target for public net worth to GDP to be falling at some date? The answer is no, because the change in government net worth is the current balance! This is a second reason why having a target for falling net worth/GDP or debt/GDP makes no sense. At best we have two targets for the same thing, and at worst we have two contradictory targets. Why not just have a target for the current deficit, which can be informed by where we want public sector net worth to move to over the long run? In my own experience, governments that have multiple fiscal targets are governments that are trying to please lots of different people rather than thinking things through.

Any ‘black hole’ based on the need to have debt to GDP falling at some date is therefore not a good basis for government decisions on spending and taxes. [6]

What about the markets

One of the many unfortunate consequences of the Kwarteng ‘fiscal event’ and the market reaction to it is that journalists, commentators, and some politicians will once again suggest that we need to make bad fiscal policy because otherwise the markets will punish us. Comments of this kind fail to understand why the market reacted the way it did recently. Worse still, the Prime Minister and Chancellor are pretending that the markets will only stay calm if there are large public spending cuts. This repeat of one of the lies of 2010 onwards is the opposite of reality: the market reaction to Kwarteng's tax cuts was in part because cuts to public spending were not credible.

I set out in detail the reasons for the market reaction to unfunded tax cuts here. In short, announcing tax cuts without saying if and when they would be supported by spending cuts, together with the likelihood that announced spending cuts could not be delivered in practice, greatly increased the uncertainty surrounding the future path of UK interest rates, which made sterling assets less attractive to hold.

The lesson from this was not ‘don’t run deficits’! The UK government in 2009, while the economy was collapsing and the deficit was rising rapidly, engaged in a fiscal stimulus and we didn’t get the market reaction we saw after the Kwarteng tax cuts. As Ben Chu has pointed out, other major countries have larger deficits and debts than the UK. Indeed the Eurozone has suspended their fiscal rules, so why cannot the UK do the same? The lesson from Truss/Kwarteng is to have clear and credible plans for both taxes and spending, and announce both at the same time. If the markets know what all parts of fiscal policy are doing and those plans are credible, this reduces uncertainty about future UK interest rates and keeps UK assets attractive.


Headlines involving black holes in the public finances are misleading at best. The numbers involved are only as good as the fiscal rules used to generate them. A fiscal rule for the current deficit to be zero may be too restrictive, and a falling debt to GDP target has many serious problems and should be abolished.

However, fiscal rules of this kind should not be chased just before the economy is likely to experience a recession which could be severe [7]. Although interest rates are not at their lower bound right now, as we have seen in the past rates can fall rapidly. Attempts at fiscal fine tuning (fiscal consolidation now and stimulus when/if rates fall to zero) are unwise. Those that appeal to recent market turbulence to say we have to have fiscal consolidation now and spending cuts are at best misreading what caused that turbulence, or at worst making up stories to achieve political ends.

Postscript A

After the global financial crisis, current data suggests UK GDP started falling in 2008Q2. Interest rates at the time were at 5%. Was that a good time for fiscal consolidation? How about 2008Q4, when rates averaged about the same as they are now? The Labour government wisely announced various forms of fiscal stimulus. Supposing instead they had raised taxes and cut spending. Rather than recovery through 2009, we would have had an even deeper recession.

While we hope that the coming recession will not be as bad as 2008/9, we cannot know that it will not be. Delaying fiscal consolidation costs very little, but fiscal contraction just before a recession can cost a great deal.

[1] While knowledge about fiscal rules is scarce, everyone knows that a black hole has to be filled with either spending cuts or tax increases. But those spending cuts or tax rises are only as sensible as the fiscal rules that motivate them, so if you want to say something intelligent about the consequences of black holes you should really have some view on how good the government’s fiscal rules are. In contrast the black hole terminology in effect tries to translate household accounting concepts to the public finances, plus some allusions to frightening solidity or mass. The black hole language is not a useful shorthand, but is instead misleading journalism.

[2] While some economists have made this point, they often add ‘and only as good as the forecasts used to produce them’. While technically true, I think this observation has less practical significance. The whole point of having an independent body to produce forecasts is that the government should accept them, unless they have a compelling forecasting reason not to. Ignoring the forecasts because they are very imprecise is just a recipe for deficit bias.

[3] One way to avoid that issue would be to target ‘cyclically adjusted’ deficits five years out, but cyclical adjustment has its own problems.

[4] Having an arbitrary target for public investment as a share of GDP, as this government has done, is particularly daft.

[5] One argument for focusing on government debt rather than net worth is that a government’s assets often do not produce a financial return to the government, while public debt has to be financed. While a government that borrows in its own currency cannot be forced to default, there might be some level of taxes where it chooses to do so, as the higher is debt interest the higher taxes have to be. However as UK levels of taxation are well below those of many other European countries, this argument hardly applies to the UK in the foreseeable future.

[6] These are not the only problems with falling debt to GDP targets. In the past such targets have failed because they have applied to a specific date, and are therefore very vulnerable to shocks hitting the economy on or just before that date. In addition, as Rob Calvert Jump and Jo Michell have shown, the target is very sensitive to whether debt held by the Bank of England is included.

[7] It is generally recognised that macroeconomic forecasts tend to miss recessions. The fact that one is being forecast now may well mean its severity is being underestimated. The combination of a large cut in personal incomes, higher interest rates, fiscal tightening and a possible recession in major Euro area countries is a powerful combination of deflationary forces.

No comments:

Post a Comment

Unfortunately because of spam with embedded links (which then flag up warnings about the whole site on some browsers), I have to personally moderate all comments. As a result, your comment may not appear for some time. In addition, I cannot publish comments with links to websites because it takes too much time to check whether these sites are legitimate.