A bad scenario for the UK

Last Thursday, the Bank of England produced much more dovish forecasts than in February. Inflation would fall to the central bank’s 2 per cent target by the start of 2027 rather than at the end of that year. And this benign prediction was based on plugging into the model interest rate assumptions that were around half a percentage point lower than in February.

In short, a big dovish change in forecast with lower interest rates and lower inflation, although the BoE’s communication sought to bury that fact.

Instead, the BoE unveiled its first set of economic scenarios, designed, it said, “to provide an illustrative quantification of how alternative economic mechanisms could result in plausible different paths for the UK economy over the forecast period”.

In the first scenario, households and companies were less certain and spent a little less and the BoE models were tweaked so that higher unemployment would bring down inflation faster. Let’s call it a disinflationary scenario.

The second is best described as an inflationary scenario, based around a supply shock. Administrative price rises would prove more persistent than the BoE usually expects. This assumption was coupled with weaker productivity growth.

As you might have already guessed, the inflationary scenario is a little inflationary and vice versa.

The BoE did not, initially, say how it thought policy would react with interest rates in either scenario, except that they would be higher in the first and lower in the second. Deputy governor Clare Lombardelli filled in this gap at the Bank of England watchers’ conference on Monday.

She said the inflationary scenario would require a 0.1 percentage point rise in interest rates over the coming year compared with the central forecast baseline, rising to 0.3 percentage points by 2028. The disinflationary scenario, being a demand shock, would require a slightly larger response of about a 0.45 percentage point cut relative to the baseline by 2027. These responses were smooth and indicated that the BoE also assumed it would be slow to learn the world had changed.

It is useful to chart the scenario outcomes compared with the inflation forecasting errors that the BoE has made in the past, as I have done below.

As is clear, the scenarios are well within the forecast band that the BoE only achieves 30 per cent of the time. Roughly speaking, Britain’s MPC can expect the real world to be more extreme than these scenarios about eight times in 10.

That is an overestimate of the real difference with the central forecast, because the BoE’s response with rates would reduce the ultimate forecast differences.

Governor Andrew Bailey and Lombardelli explained that there was no one on the MPC that actually believed these scenarios reflected reality. They were just a pick from a mix of possibilities.

I was left wondering what is the point of all this work?

Lombardelli said on Monday that the scenarios helped the committee to understand the world better.

I am far from convinced. In truth, even though she claimed otherwise, the scenario outcomes were little more than a sensitivity test of a little more and a little less inflationary model assumptions without any reaction function.

In England, schools used to be rated by government inspectors on a four-point scale from outstanding to inadequate. This innovation from the BoE would deserve a “requires improvement” grade, the third on the scale. Having spoken to many current and former MPC members at the conference, that was a common view. Lombardelli herself said the BoE scenarios “will not be the last word, far from it”. It seems she agrees.

There’s more . . . 

I am not going to change the grade I’ve given these scenarios, but you might think I’ve been a tad generous.

The BoE presented its scenarios as something new for the MPC. I found that surprising because detailed scenario analysis is normal in other central banks and I had not thought the BoE was so far behind the curve.

Philip Lane, European Central Bank chief economist, recently described the way it thinks about scenarios at a Peterson Institute event, making the point that scenarios should be about a “whole narrative” on a “particular way the world could turn” rather than “routine” sensitivity analysis looking at changing a few parameters in a model, as the BoE did.

The US Federal Reserve produces deep scenarios with illustrative monetary policy responses for every meeting of the Federal Open Market Committee. We know this because it publishes them in the Tealbook with a five-year lag. In January 2019, for example, there were seven scenarios, which at times produced interest rates 4.5 percentage points different from each other.

I do not think the BoE is as weak in its analytical prowess as Lombardelli suggests.

Why? Thanks to Fed heir Kevin Warsh, the BoE also publishes transcripts of its MPC meeting and the material provided to the MPC with an eight-year lag.

As long as you manage to ignore the painfully cringe metaphors, the analytical quality is high. In May 2016, for example, BoE staff provided the MPC with detailed scenarios of the likely economic outcomes of a leave vote in the following month’s EU referendum. They’re pretty good even with the benefit of hindsight. After the referendum, staff later followed this up with an early assessment of the consequences alongside modelled effects of different possible monetary policy responses.

Scenarios are therefore not remotely new for the BoE or the MPC and what staff have provided internally in the past is much better than what was published last week.

BoE still got the hump

In February, the BoE introduced a predicted “hump” in its short-term inflation forecast for 2025. The largest driver for this was higher energy bills resulting from wholesale gas prices rising from an expected £1.01 per therm in the November 2024 forecasts to £1.15 in February.

MPC member Megan Greene told the conference that the hump was “way smaller” in the latest forecasts than in February because energy prices had fallen back.

The latest forecast has energy prices this year below the November 2024 expectation at £0.94 a therm (current price £0.84). Greene is therefore correct about the conditioning assumption, but she was wrong about the hump being “way smaller”. It is barely flattened, as the chart below shows.

I thought that looked a bit weird and there was nothing in the Monetary Policy Report that explained it, so I got in touch with the helpful staff at the BoE to find out why its forecast still has the hump.

It turns out that this was caused by energy, but is now driven by a few new expected price rises in the short-term forecast and a bunch of unfortunate technical stuff that the BoE needs to do to smooth the connection between its short-term and macro inflation models, which has an unusually large effect.

You do have to wonder whether the BoE was slightly lowballing the short-term inflation forecast in February, not to scare the children, and highballing it now, so it can have good news to impart later in the year.

What I’ve been reading and watching

• A question lots of countries must answer is whether to stand up to Donald Trump’s bullying on tariffs or just suck them up, since retaliation hurts your people most of all. There was a great good-natured dispute about what to do between Larry Summers of Harvard and former IMF chief economist Olivier Blanchard at a recent Peterson Institute event. You won’t get an answer, but you will be better informed

• The ECB’s Isabel Schnabel warns again about cutting European rates too far, too fast

• My colleagues at the FT’s Monetary Policy Radar and I answered your questions on central banks in a live Q&A last week

• As it held rates again, the Fed last week warned about risks of higher inflation and lower activity, exactly the combination that makes things difficult

A chart that matters

While we are looking at the UK, its Office for National Statistics recently published an impact assessment of using retail scanner data to collect food and drink prices, comparing the results with the existing method of sending an army of people into supermarkets with clipboards.

In the post-pandemic period, scanner data would have lowered measured inflation significantly, as the chart shows. It is nothing like the error in the UK’s legacy retail prices index at the base aggregation level that can add almost a percentage point to inflation, but overstating inflation 0.1 to 0.2 percentage points is not great, as the chart below shows.

While the good news is that the ONS will start using scanner data from March next year, the bad news is that inflation has again been overestimated in the UK. For more on this read my colleague Louis Ashworth in FT Alphaville.