How frequently should the OBR produce forecasts?

There is much to criticise about the way in which UK fiscal policy is conducted. One particular issue is the amount of policy volatility and the fact that, twice a year, the Chancellor makes rushed decisions about tax and spending in response to relatively small movements in the Office for Budget Responsibility (OBR)’s central forecast. For example, despite her promise to hold only one fiscal event in the autumn of each year, Rachel Reeves announced a fiscal tightening in the 2025 Spring Statement because a modest deterioration in the OBR forecast meant that she would otherwise miss her ‘cast iron’ fiscal rules. Somewhat ridiculously, this tightening appears to have been precisely calibrated to return to the exact same amount of so-called ‘headroom’ against those rules.

The IMF recently suggested that one option for reducing policy tinkering in the spring would be ‘assessing rules only once per year at the time of the fiscal event’. One backbench Labour MP similarly recently argued that ‘we should move to one main OBR forecast per year, aligning it with the [Budget]’. The Institute for Government has gone further and suggested not only that the role of a second forecast be downgraded but that the OBR should publish only one economic and fiscal forecast per year.

We are in a bad equilibrium, and there is merit to arguments for downgrading the importance of fiscal forecasts that happen outside of the Budget. There is no unambiguously ‘right’ option available, and it is a delicate environment in which to make changes to the fiscal framework. Our view is that, on balance, there are strong arguments for retaining the longstanding practice of having two forecasts per year, and that alternative means of mitigating the (very real) policy volatility problem are preferable. Particularly in the current environment of high and volatile UK gilt yields, we would also caution that any reforms to the fiscal framework need to be handled very carefully – and especially so if they might be perceived (unfairly or otherwise) to be motivated by a desire to reduce fiscal transparency or to loosen the (already loose) fiscal targets.

The policy volatility problem

It is first worth reflecting on the nature of the problem that the ‘single forecast per year’ reform proposal is intended to solve. After all, UK governments have published two economic forecasts per year since the Industry Act 1975 was passed by Harold Wilson’s government.

The problem comes from the combination of having a hard, numerical, ‘bright line’ fiscal rule and then aiming to meet it almost exactly. Figure 1 shows that since 2022, successive governments have chosen to operate with far less ‘fiscal headroom’ than in the past. Macroeconomic and fiscal forecasts are inevitably volatile. The issue is that under the current set-up, forecast volatility feeds through almost one-for-one into policy volatility, because there is such a limited buffer against forecast fluctuations. This is exacerbated by the tendency of Chancellors to offset any improvements in the fiscal outlook with tax cuts or spending increases. So, twice a year, tax and spending policy is tweaked and fine-tuned in responses to movements in the OBR’s central forecast. Despite this government supposedly placing a great deal of weight on the importance of policy stability (see e.g. Rachel Reeves’s 2024 Mais Lecture), this way of operating all but guarantees the opposite. This is far from ideal and should stop.

Figure 1. ‘Headroom’ against fiscal targets in successive OBR forecasts

_{Source: Office for Budget Responsibility, Economic and Fiscal Outlook March 2025, chart 7.2.}

One reasonable conclusion might be that if the government is going to have hard, pass–fail numerical rules, it needs to operate with more ‘headroom’. There is certainly a case for that. The Chancellor could choose to build more ‘headroom’ at the Autumn Budget. But it is also worth considering other ways of addressing the issue through modest tweaks to the framework.

The tension between one fiscal event and two forecasts

There is a tension between having a single fiscal event per year (an autumn Budget) but having two economic and fiscal forecasts (one alongside the Budget and the other being a spring forecast) – mostly because there is a risk that in the non-fiscal-event forecast, the government will feel compelled to adjust policy if it would otherwise be on track to miss its fiscal rules. Moving to a single forecast per year would resolve this tension. It would also eliminate the temptation for Chancellors to announce shiny, new policies in the spring (which will exist regardless of how much ‘headroom’ they have). Having a single forecast should, therefore, reduce the amount of policy volatility – see here for a longer version of this argument.

This would come with downsides, however. Most obviously, it would come at the cost of reduced fiscal transparency. As noted by Richard Hughes, Chair of the OBR, in his recent evidence to the Treasury Select Committee, ‘doing two forecasts a year is international good practice … were we to reduce the number of forecasts to one, that would make us one of the least fiscally transparent countries in Europe and of any major advanced economy’.

There are other advantages to having a second forecast in the spring, even if there is only one fiscal event each year in the autumn. An important one is that there are circumstances in which the government would, for very good reasons, want to make policy changes between fiscal events (e.g. if there was a large shock to the financial system or the outbreak of a severe pandemic) and having a second forecast might make it easier to do so without the need to hold an ‘emergency’ statement. Another is that if the UK suffers adverse economic developments (e.g. a gradual deterioration in the global trading environment), having these flagged in a spring forecast update could – if the government could stick to having only one fiscal event a year – provide more time for the Treasury to prepare policy options, and more time for the Chancellor and Cabinet colleagues to decide how best to respond in the autumn, rather than have this decision-making process rushed and crammed into a short pre-Budget period. It also provides MPs – and those outside parliament – with information that enhances their ability to engage in policy debates. Having more information is valuable.

Given these arguments, it is worth considering whether the problem could be dealt with without scrapping the second OBR forecast.

Other options

One option, which is what we interpret the IMF as having advocated, is to change the current framework so that the OBR continues to produce a second forecast in the spring, but only assesses the fiscal rules formally each autumn. This would be nuanced. In the spring, the OBR forecast would effectively answer the question: ‘If a Budget were held today, would the government be meeting its rules?’. But because the rules were not being formally assessed, they could not, at that point, actually be breached. The government could then take stock, and make any necessary policy adjustments in the autumn, when the OBR would produce a ‘major’ or ‘main’ forecast. There is much to be said for this in theory. In practice, the question is how external commentators and – in particular – financial market analysts would react if the government was judged to be on course to miss its rules on the basis of the spring forecast, even if that did not constitute a formal breach. It would carry risks, and those risks are likely greater now than had this change been introduced from the start of the parliament.

A second option would be to bring forward a change that is already legislated to come into effect from Spring 2027. From that point, the target for current budget balance (the fiscal rule which currently binds) will be expressed as a range: balance will be defined as including deficits up to 0.5% of GDP. The framework explicitly stipulates that if the current budget is in deficit (of up to 0.5% of GDP) between fiscal events (i.e. in the spring forecast), it would need to be returned to surplus in the autumn – but no immediate policy correction would be required alongside the spring forecast, as the government would still be meeting its rule. The Chancellor could simply say that appropriate action would be taken in the autumn (at which point, if the current budget would otherwise be forecast to be in deficit, a fiscal tightening would need to be announced). This was explicitly justified on the grounds that it will ‘support the government’s commitment to a single fiscal event every year by avoiding the need for policy adjustment at forecasts outside of fiscal events’ – i.e. on the grounds that it will reduce policy volatility.

This is an elegant and sensible solution to the tension between having a single fiscal event but two forecasts each year. If at the second forecast the government is slightly on the wrong side of where it wants to be, no immediate action is required. If it is a long way on the wrong side, with a current budget deficit of more than 0.5% of GDP, then the Chancellor does need to respond. But the government and the world still get the benefit of a second transparent forecast from the OBR. This makes a lot of sense.

The proposed range for the current budget is currently intended to apply at all fiscal forecasts. In our view it should only apply between fiscal events. If the range applies at the Budget, there is a risk that the bottom of the range (i.e. a deficit of 0.5% of GDP) becomes the new line against which ‘headroom’ is measured and obsessed about. If the extent of ‘headroom’ was the same as at present (i.e. around £10 billion) but against a deficit of 0.5% of GDP rather than against budget balance, the government would be just as likely to need policy action in the spring, only from a starting position of looser fiscal policy. The range should be there to help manage the fact that forecasts are inherently volatile, not to allow for higher borrowing.

The range is not due to kick in until 2026–27, and so will not be of use until the spring forecast of 2027. There was an opportunity to start using the range, and to reform its structure, before the draft Charter for Budget Responsibility had been passed into legislation (as flagged here), but this was missed, and we are where we are. There is a good argument for bringing its introduction forward (and, ideally, tweaking its design so that it only applies between fiscal events), but to do so would require a change to secondary legislation. It could be viewed as an attempt to fiddle with the rules in an attempt to ease fiscal constraints. Such a change would need to be carefully communicated and framed purely as a means of reducing policy volatility, and not as a backtracking on the government’s commitment to fiscal discipline. If the range were introduced from 2026, it could lead to a situation where the Chancellor receives a worse forecast in Spring 2026 (than in Autumn 2025) but does not have to announce a fiscal tightening at that point. If the forecasts improve by the autumn, no tightening would be needed and unnecessary policy volatility will have been avoided. If not, it would only be pain delayed, not pain avoided – i.e. fiscal policy would not end up being looser overall.

More broadly, with any move towards a single fiscal event while retaining two forecasts, there is a risk that, if a spring forecast showed the government to be on course to miss its fiscal rules, it would trigger months of furious speculation about the possible response. But we do not escape any such speculation at present. And if the government moved to publishing a single forecast, there would still be external forecasts that could trigger speculation. To avoid such speculation, there is no substitute for the government operating with more ‘fiscal headroom’ and providing more clarity over how it would respond in the event that a fiscal tightening were needed.

Conclusions

We are starting from a bad equilibrium, one in which the UK fiscal debate is over-obsessed with the unhelpful concept of ‘fiscal headroom’, and where inevitable volatility in the economic and fiscal forecasts gets mainlined into policy volatility at twice-yearly fiscal events. It is against this backdrop that some have suggested that the UK would be better off with only a single official forecast each year. Our view is that, on balance, the government should not sacrifice the fiscal transparency that comes with a second forecast. There is a strong case to adjust and bring forward the introduction of the range for the current budget, so that it applies in-between annual fiscal events from Spring 2026. The change would need to be communicated extremely carefully in such a delicate, febrile fiscal environment. But, with successful communication, it would be a sensible means of reducing potential policy volatility from Spring 2026. That in turn could help cement the Chancellor’s intention to have only one fiscal event each year. It could do this without a weakening of the UK’s fiscal position.

Finally, we would note that the UK has produced two forecasts per year since 1975 without running into so much trouble. The current state of affairs results in large part from the Chancellor’s decision to operate with such a small amount of ‘headroom’ against her pass–fail fiscal rules. There is a case for also addressing that in the autumn – especially if using the range for the current budget is not brought forward to 2026.