14 July 2025

The OBR is always wrong

Making economic policy by forecast is driving Britain in circles.

In Britain, checking the weather forecast is second nature. We obsessively consult our favourite weather app, ready to adapt our day based on whether sunshine or showers are predicted. But we understand that weather forecasts become less reliable the further out they go. If someone claimed it would rain at exactly 10:13am two Tuesdays from now, would we believe them? We might pack a brolly. But we certainly wouldn’t rearrange our life around that precise time.

We understand that weather forecasts are informed guesses. They deal in probabilities, not guarantees – and culturally, we use them with that understanding. But when it comes to economic forecasts, we seem to forget this common sense entirely.

Take this year’s spring forecast from the Office for Budget Responsibility (OBR). Headlines focused on a specific figure: the forecasted shift from a £9.9bn surplus to a £4.1bn deficit by 2029-30. The ensuing political and media response centred entirely on this very specific number. But hang on – this is a prediction about the state of a very complex economy four years in the future. Common sense, and our experience of the weather, tells us that surely can’t be right.

Well, the OBR agrees. In its 2023 forecast performance report, it wrote: “Single point economic forecasts of an uncertain future are almost certain to be wrong – a single point forecast of an outcome (such as the inflation rate next year, level of GDP growth in three years’ time, or for the size of the fiscal deficit at the end of a parliament) has virtually no chance of being correct [my emphasis].”

This is entirely borne out by any look at the history of OBR (or indeed any economic) forecasts. A comparison of the forecast with what actually happened will show the folly of focusing on any specific number. So where have these numbers come from and how should we be using these forecasts? Consider the OBR’s own chart (below) estimating the 2029-30 budget surplus or deficit.

It shows, with 80 per cent probability, that the outcome could be anywhere between a 4 per cent surplus and a 4.6 per cent deficit of GDP – a swing of nearly £300bn. Yet all the attention focuses on the central estimate: a £4.1bn deficit, which statistically has almost no chance of being the final result.

This central estimate is essentially the midpoint between extremes. It helps indicate the centre of possible outcomes but treating it as a definitive prediction is misguided. It’s like hearing a forecast for rain at 10:13am in two weeks and basing all your plans around that exact timing. That surely can’t be right. It would be more reasonable to say: “There’s a risk that the fiscal rules may be breached – we will adapt fiscal policy accordingly if these trends firm up.”

Another issue is our apparent blindness to the inherent limitations of OBR forecasts. Again, from the OBR’s own report: “We are no better equipped to see into the future than other forecasters… We are also subject to some constraints that many other forecasters are not, like the legal requirement to condition our central forecasts on stated (rather than anticipated) Government policy.” This is key. The OBR doesn’t have special information. It relies on publicly available data and is restricted to modelling only government policies that are fully detailed and officially in place when the forecast is made. As a result, their forecasts often miss major factors that might change economic conditions significantly.

For example, the OBR’s spring forecast could not account for the government’s planned £113bn increase in infrastructure spending, the industrial strategy or the greatest investment in social housing since the 1970s. Nor does it account for events that followed, such as Donald Trump’s proposed tariffs or new trade deals. These omissions mean the forecasts are partial and outdated even as they’re published. Much like a weather forecast, these economic projections are just snapshots. They offer insights, but they are not – and should not be treated as – precise guarantees to base our long-term policy decisions on.

A final problem lies in the schedule itself. The requirement for two OBR forecasts per year has led the government to effectively base its fiscal calendar around them. This year, for example, what looked like a full fiscal event happened during the Spring Statement, even though the new Labour government had committed to only one major fiscal event per year. Why? Because the OBR had to produce a forecast, and the government felt it had to respond with fiscal announcements. Changes to welfare policy in response to the spring estimate are a stark case in point and policy is still arguably being driven by the belief that we must “score” a specific number.

This back-and-forth between forecasts and policy reactions creates volatility, rather than stability. Both the International Monetary Fund and the Institute for Government have recommended shifting to a single major forecast each year, aligned with the autumn Budget, to encourage policy stability. This is a sensible change.

So what should we do? First, we must stop treating single-point estimates as gospel. They can be useful, especially to test different policy options, but they should not be treated as firm predictions of future outcomes. Second, we should adjust the way we report and respond to forecasts – focusing on ranges and probabilities rather than repeating misleadingly precise numbers. We must also clearly communicate the inherent uncertainties and the limitations of the data used. Finally, we should move to one main OBR forecast per year, aligning it with the national budget, and resist the temptation to craft policy around forecasts that will likely be outdated by the time they are acted upon.

In short, let’s apply the same common sense to economic forecasting that we do to weather forecasting. We don’t cancel weddings months out because there’s a chance of drizzle, and we shouldn’t restructure fiscal policy purely on the basis of a single number we know to be, at very best, an informed guess. Instead, we should remain flexible. Economic forecasts are valuable tools – but only if we remember what they are: guides, not guarantees.