The last half mile − speech by Alan Taylor

Speech

Good evening. It’s a pleasure to be here at Leeds University Business School and I’d like to thank the university, the school, and the organizers for making this event possible.

This is my first speech as an MPC member, so it’s fitting that I’m giving it here in Leeds, back home in Yorkshire. I stand here only a few miles from where I was born in Wakefield and from where I grew up in a small mining village. While I’ve spent most of my adult life living and working in the United States, I’ve returned to the UK regularly and retained a keen interest in the UK economy. For me it is an honour and a privilege to be engaged in public service at the Bank of England and to contribute to UK monetary policymaking.

Today I want to set out how I see the UK economy currently, how we got here, and the outlook for inflation and monetary policy.

While it is informative to know where we have come from, naturally my focus is mainly on where we are going next, not least given the usual lags in the transmission of monetary policy. But the past is not silent; and in macroeconomics, it is the only laboratory we have. Evidence from the past, joined with theory, informs my view of the future and hence how to set policy. And today, I want to try to take some insights from economic history – from both the recent and more distant past.

To be up front, my view is that we’ve made it to the last half mile on inflation, but with the economy weakening, it’s time to get interest rates back toward normal to sustain a soft landing.

To make my case, I’ll want to consider the dynamics of inflation that we are living through and the evolving outlook for the normalisation of policy in the medium term.

As may be apparent from my vote in December, my view is that the risks around inflation have shifted in the last 12 months. Inflation moderated faster than expected in 2024, and aside from one-time tax shocks and various base effects which our forecasts show will push upwards on inflation in 2025, I would expect the underlying trend of inflation to remain on track towards the 2% target from now on. Thus, as my title suggests, though the road to normalisation will be bumpy, we have moved along much closer to the end point – the last half mile.

However, the most recent data and forward-looking activity indicators present an increasingly gloomy outlook for 2025. The labour market is near balance, but is still loosening at pace, GDP growth appears to have ground to a halt in the second half of 2024, and with confidence indicators and business expectations veering to the pessimistic, in my view the risks are now more skewed to the downside. Put differently, and I will return to this framework later: in terms of probabilities, the UK macroeconomy seems to me to be in a different scenario now than it was 12 months ago.

It is often said (not least by themselves) that central bankers are data-dependent. But data are mostly backward looking, and sometimes murky. And at possible turning points we also need to be outlook-dependent, alert to both hard and soft signals, and taking note of sentiment and feedback from surveys and other sources. The MPC can also rely on the qualitative information we receive from our unique, excellent, and extensive network of Agents, such as the team here at the Bank’s hub in Leeds.

My goal in this speech is to walk through how my thinking uses all that information and brings me to an assessment of the balance of risks relative to our current forecast. I’ll try to explain why both the recent UK data and the outlook leave me alert to the possible need for a more activist, more rapid withdrawal of monetary policy restrictiveness in the next couple of years.

But let’s still be clear on the progress we have made so far. Inflation has fallen materially over the past couple of years, and there has been little sign so far of macroeconomic costs: no marked increase in unemployment, and no major downturn in GDP. We might say, therefore, that the UK economy, like other major economies, is still on course for a soft landing – something that looked unlikely at points over the last few years. As we continue our approach along the last half mile, there will be turbulence and cross winds. But, given considerable progress already, we need to be mindful to try to keep the touchdown smooth.

Where have we come from?

In my academic work, I often draw on lessons from economic history, and I will probably talk about that more in future speeches. In that vein, let me describe the historical backdrop very quickly with a focus on institutional details.

At a time of inflationary challenges, the most relevant era to look back to is the 1970s (which some of us are old enough to remember): the greatest peacetime inflation in history, not just for the UK economy but also for most other advanced economies. The main thing in common with the 1970s is the origin of the shock – an energy price shock that is, to first approximation, completely external to the UK economy. However, the nature of the energy shock was different this time: in the 1970s, repeated and large oil-price shocks that largely persisted. In recent years, a large upward spike in oil and natural gas prices that have now fallen back towards more normal levels, in relative terms, although still not completely.^{footnote [1]}

There are also many differences with the 1970s, including the evolution of the global economy and the international integration of supply chains. But on today’s topic, for me a key difference with the 1970s is the macroeconomic policy regime, with changes to both the fiscal rules and the monetary regime, in the UK as in many other countries. Many of my colleagues on the MPC, past and present, have remarked on this, including former Deputy Governor Ben Broadbent in his final MPC speech. The current UK monetary regime, with the independence of the central bank and a formal inflation targeting remit established in 1997, stands us in great stead. And for the Bank, the remit set by the government for the MPC is clear on our duty: primarily a credible commitment and a policy objective to dampen inflation dynamics and keep inflation expectations anchored, while still being alert to potential trade-offs so as to avoid undue volatility in output.

This has been essential for the progress we have made so far towards a soft landing. The trade-off language in the remit is very important – as Governor Andrew Bailey explained in his speech at Jackson Hole last August – and it is most crucial when the economy is hit by large or persistent shocks or is in the gradual process of recovering from them, as recently. In a market economy, where prices and wages in any sector are set by buyers and sellers, and their reactions propagate shocks on to other prices and wage in other sectors, the Bank can’t instantly eliminate the inflationary consequences of shocks. But, with the tools we have, principally Bank Rate, our primary job is to react to and dampen those inflation shocks, whilst being aware of trade-offs, that is, the need to avert excessive swings in economic activity. As the Governor said:

The language of the Bank of England remit states that “the inflation target holds at all times”, which is essential. But it also states that “the actual inflation rate will on occasion depart from its target as a result of shocks and disturbances”.

It further adds that “attempts to keep inflation at the inflation target in these circumstances may cause undesirable volatility in output”. In 2013 the remit was amended by adding that where shocks are particularly large or the effects of shocks may persist over an extended period (or both) then “the Committee is likely to be faced with more significant trade-offs between the speed with which it aims to bring inflation back to target and the consideration that should be placed on the variability of output”.

That longer-run institutional context sets the stage for an analysis of the current inflation. The story of inflation over recent years is well rehearsed, so I will describe this briefly. Chart 1 shows annual inflation according to the CPI, alongside successive forecasts by the Bank of England in recent Monetary Policy Reports.

The initial increase in inflation in 2021 was a surprise, with prior forecasts (up to May 2021) not showing any meaningful increase in inflation. While accounts differ, I think it is clear that the initial burst of inflation was driven by a combination of factors: a rebound in global demand after the pandemic (especially in the US, with a large fiscal impetus), combined with still impaired supply (notably in China, but also more broadly), leading to a sharp increase in the price of certain types of goods, energy, and commodities. This alone would probably have been transitory. The Bank’s forecasts from August 2021 to February 2022 embody this period.

Chart 1: Annual CPI inflation and successive MPR forecast vintages
Forecast vintages from May 2021 to November 2024 Monetary Policy Reports ^(a)

The Russian invasion of Ukraine in February 2022 compounded the problem and led inflation to peak higher and last longer. The MPC’s May 2022 forecast is the first to incorporate this dramatic shift in circumstances, and it turns out to have quickly adjusted forecast paths to something that actually closely resembles the eventual shape of inflation that we’ve experienced. Given the importance of energy prices in driving inflation up and back down, differences in the conditioning path for energy prices had material bearing on the forecast errors around this time, a topic I will return to in a moment. As such, there is some natural variation in the forecasts from May 2022 onwards reflecting differences in the conditioning assumptions for energy. But from that point on^{footnote [2]}, forecast errors have been much smaller and more two-sided, in contrast to the larger and one-sided misses prior to that point.

More recently, in 2024, inflation has turned out somewhat below prior Bank forecasts, with inflation falling sharply to near (or even, briefly, below) the 2% target quite recently on account of a fall in energy prices and so-called base effects. The November 2024 MPR forecast sees a rebound in inflation over 2025 as the drag from falling energy prices fade, and as various one-time tax shocks hit, before inflation returns close to target in 2026-27. That rebound serves as a caution, and reminds us of how bumpy the last half mile of the road can be.

Chart 2 shows inflation rates for different parts of the CPI basket: energy, food and non-alcoholic drinks, other goods, and services. This illustrates clearly the different waves of inflation seen in different parts of the CPI basket over recent years – first energy and other goods, then food, then services. This was also noted by my MPC colleague Deputy Governor Dave Ramsden in his speech here in Leeds in November 2024.

Services inflation is still well-above pre-pandemic averages and target-consistent rates, but now appears to be returning slowly to those levels. Meanwhile inflation for energy, food, and other goods have all returned to roughly normal levels.

One might ask: why should an energy price shock, such as the one we saw in 2021 and 2022, lead to higher inflation of food, goods, and services? There are several possible mechanisms, which are set out as follows.

1. Global demand – Energy prices reflect the balance of demand and supply for energy, and the demand for energy in turn reflects broader demand conditions in the global economy. So higher energy prices might occur simultaneously with stronger global demand, leading the price of traded products to rise alongside the price of energy. This is in essence another external shock.
2. Energy as an input – Energy is an input into the production process of most firms and industries across the domestic economy, especially for food, goods, and transportation. When the price of energy rises, the costs of businesses rise, and they might pass those higher costs onto consumers in the form of higher prices of domestically produced goods and services, both traded and non-traded.
3. Inflation expectations and wages – Energy is consumed directly and indirectly by households, so an increase in energy prices makes households worse off. Workers might therefore bargain for higher wages to compensate, on account of either their lost purchasing power or the expectation of higher inflation in future. Energy prices are particularly salient in the formation of household inflation expectations. Higher wages are higher costs to businesses, who might in turn raise prices.
4. Policy responses – Governments might be inclined to intervene in the face of high energy prices to alleviate the effects on households and businesses. Providing subsidies, providing cash transfers, or implementing price caps on energy, all have the effect of boosting real incomes (directly or indirectly) relative to the situation without the intervention, and so increasing the level of aggregate demand which will tend to increase inflation.
5. Relative prices – When the price of energy or other downstream products increases, consumers will tend to buy less of them, but they might then buy more of something else in substitution. This shift in demand will spread the inflation out across a wider range of goods, at least until the relative prices have normalised.

Which of these mechanisms have been at play over the course of the past few years? In short, all of them, to differing degrees.

Clearly it is inevitable that an energy price shock will propagate through the economy via different channels and push up inflation more broadly. But several questions remain. Which non-energy products are most affected by an energy shock, and by how much? What is the timing of the propagation to the other products? And was the propagation of the latest energy shocks typical or unusual? I turn to these questions now.

Some empirical evidence on the impact of energy shocks

Let us start with a stylised example, which I’ll then test with some data. Chart 3 shows a series of arches, which illustrate the response of inflation of different sorts of products to an energy price shock. Some respond quickly, some respond more slowly, but each product responds in a sort of arch – up then down. The aggregate response of CPI inflation then might be traced out as the weighted average of each of the component arches, illustrated by the dashed line in Chart 3. Notice that, as drawn here, even if each individual arched response is symmetric, the aggregate response is not symmetric: it has a right tail that stretches out.

At the beginning, immediately after the shock, the aggregate response (dashed line) rises rapidly, since all the products (solid lines) are responding to some degree. Later, as some products finish their arch, the aggregate response gets smaller. But the aggregate response doesn’t get back to zero until the last product – the one with the slowest response – gets back to zero.

Chart 3: Stylised illustration of the responses of the inflation rates of different products to an energy price shock ^(a)

What the path of these arches are in practice is an empirical question, on which I hope to now offer some insights. Using UK CPI data since 1988, aggregated into different categories which we might intuitively think should respond at different speeds, we can estimate the response ‘arch’ of different parts of the CPI basket.

To do this, we estimate local projections (Jordà, 2005) of the response of different categories of CPI products to an energy price shock. Here, we use CPI energy inflation as the shock, instrumented by the annual changes in monthly spot crude oil prices (Brent crude). More details are in an Appendix, including some robustness checks. The results here are robust to starting the sample in 1992 (the start of the inflation targeting era) and a few other breakdowns of the CPI basket.

Chart 4 shows the results for different types of products and are intuitive. After an energy price shock at time zero, consumer energy prices (left panel) react sharply before turning negative after around two years. However, the shock continues to pass through to other components of the CPI. Food prices react strongly and fairly quickly (right panel), with a peak impact on annual inflation 16 months after the initial shock, while goods prices react less strongly but still relatively quickly, the impact peaking after 18 months. The impact of the energy shock on food and goods prices is short-lived, fully fading after around 30 months (2½ years). In contrast, services and wages exhibit a longer-lasting pass-through, peaking at 17 and 18 months, respectively, and not fully receding for over three years. Interestingly, the estimated responses of services and wages are robust to using a pre-pandemic sample (see Appendix) so in that sense the stickiest inflation components, which have come to be of greatest concern during this inflationary episode, do not seem to be behaving very differently in recent years.

A different cut of the CPI data by import-intensity, which could be interpreted as a measure of how domestic a product is – i.e. low import-intensity means higher domestic content – is shown in Chart 5. Products with the lowest import-intensity (i.e. those with higher domestic content), which partially overlap with certain services, respond much more slowly than products with higher import-intensity. This suggests that an external energy price shock will continue to have an impact on headline CPI inflation for at least three years. There are numerous other ways to cut the CPI data, some of which are shown in the Appendix.

To sum up, this analysis suggests to me that much of the increase in core and services inflation over the past couple of years was the predictable propagation of an energy price shock in line with dynamics seen over a longer sample period. This is consistent with a range of other recent research, some of which is summarised in the Appendix.

Should we have been surprised?

Given that understanding of the typical pattern describing the domestic transmission of UK energy price shocks, should we have been surprised about inflation over recent years?

Unfortunately, knowing the response of the economy to shocks is only one part of the equation. One also needs to know what shocks will hit the economy and, by definition, these are unpredictable. As a result, any forecast must either assume that future shocks will turn out to be at their mean, i.e. zero, or else make some conditioning assumptions about the expected path of various economic factors under alternative shocks.

The MPC’s forecast is always a conditional forecast, at least in respect to some shocks, which means it takes as given the future outturns of various “conditioning paths”. These include energy prices, Bank Rate (based on the yield curve), and various asset prices. Forecasts conditioned on a path for energy prices which later turns out to have been very different to the actual realized (true) outcome, will naturally have problems. The bigger that difference, the bigger the problem, as in 2021-22. But that does not necessarily mean the forecasting model was wrong; rather, unpredictable shocks materialized.

Recent work by Kanngiesser and Willems (2024) explored this for the MPC’s forecast. They find a large and important role for these unpredictable energy shocks. In other words, had the MPC in 2021 known the true future path of energy prices (i.e. had we foreseen the Russian invasion of Ukraine and its impact), we would have made much smaller forecast errors and thus been less surprised by the extent of the peak in inflation. That is, fed the right energy shocks, the MPC’s forecast would have got much closer to the right outcome, but shocks by their nature cannot be known in advance.

Indeed, no forecast model in any discipline can succeed by that metric, and we have seen the same inevitable issues arise for other central banks in this same timeframe. In a recent speech, ECB Chief Economist Philip Lane explored such a counterfactual in more detail to evaluate the performance of the ECB’s forecast models of Eurozone inflation. Much of the inflation forecast error was explained by energy prices turning out differently to expectations, via both direct and indirect effects. Another key contributor to the error was food price inflation, which was much higher than expected, and might in part also be attributed to the higher-than-expected energy prices. Put together, Lane suggests that the path of inflation over 2022-2023 could have been fairly well forecast from the standpoint of late-2021, if there was perfect knowledge of the shocks.

All of this suggests that the economy has behaved broadly in line with historically-estimated economic relationships, albeit in the face of unusually large shocks.

Where are we now?

The above discussion sets the stage for where we are now, with still slightly above-target inflation, yet at a level that is no great surprise ex post given the large shocks we have faced and their slow propagation which is still playing out.

After briefly falling below target to 1.7% in September 2024, annual CPI inflation was 2.6% in November – a little above the short-term forecast in the November 2024 MPR. That forecast then has inflation increasing to a peak of 2.8% in 2025 Q3, before falling back close to the 2% target in late-2026 and 2027 (see Chart 1).

The apparent rebound in inflation in late 2024 and throughout 2025 is largely because the drag from falling energy prices disappears, and so headline inflation reverts to its ‘underlying’ rate, plus additional contributions from one-off tax changes introduced in the Autumn Budget 2024 (see Chart B in Box B in the November 2024 MPR).

Is this enough to be confident that the future path of inflation is returning to target?

To gauge progress, what we ideally want is a measure of the true underlying inflation trend purged of all noise and one-off shocks. This is as much an art as a science, since that concept is unobservable. But plenty of proxies and estimates exist.

One approach would be to strip out the predictable pass-through of external price shocks (notably energy), to leave the remaining ‘domestic’ inflation. That is the spirit in which the analysis presented earlier is helpful – if elevated services inflation is largely due to the normal elongated passthrough of a (large) energy price shock, then it might not be indicative of high ‘underlying’ inflation as some have feared.

Statistical approaches to estimating underlying or ‘core’ inflation also exist, including exclusion-based measures (excluding external or volatile prices, like food and energy) or distribution-based measures (such as the median or trimmed mean across detailed CPI products). Most of these measures, at both annual frequency and higher-frequencies, are currently around 3% and I would expect them to continue to fall back towards 2%.

Looking to independent calculations is also useful as a check. Here, a good example is the new composite core inflation measure constructed by Chris Giles and Joel Suss for the Financial Times, which they call “FT core inflation”. This seeks to amalgamate various other ‘core’ indicators in a statistically optimal way. Chart 6 reproduces the FT core measure for the UK, measured on a 12-month (annual), six-month annualised, and three-month annualised basis. Here some progress on normalisation can be noted: while the 12-month and six-month inflation measures sit just outside the inter-quartile range, the latest three-month reading has now fallen within that putatively normal range, where that range is computed on the pre-pandemic sample that covers the inflation targeting era from 1992 to 2019. The trend towards 2% across these time horizons is also clear.

What about the outlook of other economic indicators?

The November 2024 MPR forecast for unemployment rises very modestly to 4.4% by 2027, which would be an increase of only around 0.2pp from current rates (see Chart 1.3 in the November 2024 MPR). The unemployment rate is currently hard to determine due to well-known issues with the Labour Force Survey. But looking through recent volatility and drawing on other data sources, it is clear that the unemployment rate remains historically low as of now. Nonetheless, recent disappointing data outturns, such as 0% GDP growth in Q3 2024, as well as a variety of survey data on sentiment of a forward-looking nature (hiring and investment intentions, etc.), point to rising concerns here.

Taken together then, the MPC’s November 2024 MPR is still a forecast for a soft landing, with inflation returning to target and no recession or sharp rise in unemployment. If that is achieved, history may look back on this period as a success.

Where next? The scenarios for policy in 2025 and beyond

I have summarised where I think we have come from and where I think we currently are in the process of inflation returning to target. I have also laid out my interpretation of current data and the near-term outlook. It is therefore time to conclude by laying out what I see as the implications for policy going forward.

Now the usual caveat applies: prediction is very difficult, especially about the future. As we have just seen, if we know what shocks the future would bring, the forecast models we use could do reasonably well. But we don’t.

Thus, to frame the risks around the central forecast I will use a form of scenario analysis. This follows approaches that the Bank has begun to develop of late, one of the many reforms to our policymaking toolkit that is under development following the wise and influential Bernanke Report published last year.

In recent MPC rounds we have considered three stylised “cases” as describing the possible dynamics of the UK economy and inflation persistence (see Box B of the November 2024 MPR for details). Now, different cases can mean different things to the nine members of the committee, so today all I can do is give you my own interpretations.

In case 1, the economy is viewed as having suffered a large external inflation shock; but although large, the shock has normal persistence, and the economy is adjusting back to a normal steady state at the usual speed, assisted by a normal restrictive monetary policy response in line with the standard reaction function, and in this case no unusual deviation of policy action is needed to get inflation to target in a timely manner. In the long run, there has been no structural change to the economy and the policy rate returns to its former neutral level.

In case 2, the economy is adjusting back to a normal steady state, but persistence is somewhat higher than normal, possibly due to temporary shifts in price- and wage-setting dynamics, so a more restrictive monetary policy response in warranted, which could entail slower growth and the opening up of some slack in the near term to get inflation back to target in a timely manner. In the long run, there is again no structural change to the economy and the policy rate returns to its former neutral level.

In case 3, both temporary shifts and long-run structural changes in the economy bear down. Inflation is highly persistent. More restrictive policy is warranted in the short run to bring inflation down, due to higher demand pressures, or higher price- and wage-setting pressures, or perhaps a higher natural rate of unemployment. In the longer run, a new higher neutral policy rate is warranted if the real natural rate (r*) has risen due to local or global shifts in the demand and supply of assets (I shall return to r* shortly).

Presented in these broad terms, and based on my own interpretations of each case, I would say that my probabilities would have been placed about equally across all three cases 6-12 months ago.

But right now is quite different, and with the recent flow of data and sentiment indicators, I see signs of weakening demand and labour market softness that make case 3 less likely and cases 1 and 2 more likely. I would perhaps assign a 40-40-20 probability split across the cases 1 to 3. Incidentally, that kind of probability distribution is not far off what the Bank hears from its own surveys of market participants (e.g. the December 2024 MaPS results) and what the Bank’s Agents hear from contacts when asked about the likelihood of each case. But I additionally would view the weight on case 1 rising over recent months.

We know the economy has been slowing progressively as 2024 unfolded, coming to halt by Q3. But there are important signs that this is to a large extent driven by faltering demand. The simplest intuitive sign here is that activity has softened at the same time as inflation, both headline and core, has moderated. A more technical analysis, using a model that identifies demand and supply shocks statistically based on output-price correlations, can translate this intuition into a decomposition of implied demand and supply shocks. In her recent speech, Deputy Governor Sarah Breeden highlighted the gradual shift in the estimated shocks from this analysis over the last 12 months, where the distribution of demand shocks has been gradually sliding into negative territory over the course of the past year. Chart 7 replicates that chart, but splits the shocks into UK domestic demand shocks (left panel) and global demand shocks hitting the UK (right panel). In the UK, a demand softening seems to have been under way; the key question then is how far it might go.

Now let me translate all of that into policy space and the prospective path of interest rates. Chart 8 shows some illustrative paths for Bank Rate under the three cases described above, as well as a downside scenario path which I will discuss shortly. I will stress that these paths are my own views and not those of the committee, and are indicative rather than commitments, intended to show broadly how different scenarios could play out.

Looking to the current calendar year, the middle case, case 2, has been identified (by many observers) with a “gradual” pace of interest rate cuts of one per quarter, or 100 basis points by December 2025. In that sense case 2 has also been viewed (by some) as the central case. As Chart 8 illustrates, I would judge around 100 basis points of cuts by the end of 2025 to be appropriate under case 2, the central case, and I expressed this view of “gradual” at the Treasury Select Committee hearing in November.

However, my own view, with the probability of case 1 rising, and the risk of demand stalling, is that we could then need a more accelerated pace of rate cuts, perhaps 125 or 150 basis points in the coming year. Under this view, the economy might face adverse demand pressures potentially on many fronts, while supply is less perturbed. To me this pessimistic scenario of case 1 has been becoming a more likely scenario given recent data.

The bulk of the range of City economists’ forecasts for Bank Rate at the end of 2025, compiled after the publication of the November 2024 MPR, fall in the range of 3.0% to 4.0%. The illustrative path for case 1 in Chart 8 aligns with the lower half of City forecasts of the rate path for 2025 (ending at 3.0% to 3.5%), while the top half is more aligned with the illustrative case 2 or ‘gradual’ path (ending at 3.5% to 4.0%). Overall, this set of City forecasts seem to bracket cases 1 and 2 as I have depicted them in Chart 8, again in line with the mass of probability falling mainly on these two cases.

But there is a clear risk here. If we were to misjudge, or be too late to recognise if this case 1 possibility were developing as the true scenario, then we would be using case 2 policies in a case 1 world – and we would run the risk of undershooting the inflation target while falling into an unwanted recession. And with that risk comes potentially high costs, including longer-run damage to the economy (see Jordà and others, 2024), and a new set of stabilisation challenges for monetary policy in the years to come. A hard landing it would be.

Of course, there are other risks. We could end up with case 1 policies in a case 2 world, for example. And the risk of case 3 still isn’t necessarily zero either, so the Bank needs to be extremely alert to all real-time and forward-looking signals to get a clear read. There will be bumps on the road in the final half mile, as there must be. Coming steeply down from high inflation, the direction of travel is firm, and upticks are rare; when you are at or near target, the trend should be flat, but the ups and downs of shocks will cause both upticks and downticks from time to time.

Many of these unwelcome bumps are familiar and foreshadowed – some arise from lumpy administered price changes (such as electricity and other utilities) or unexpected one-time tax measures that phase in (like, VAT and NICs changes) or don’t phase out as expected (like, fuel duty). Some bumps may be large in the context of a 2% inflation target, as the November MPR forecast indicated. Indeed, we could possibly end up in uncomfortable trade-off territory very soon, if the economy weakens while inflation bumps up.

Chart 9 (reproduced from the November 2024 MPR) shows the impact of the Autumn Budget 2024 on the Bank’s inflation forecast, via both taxes and spare capacity. A lot of noise in the future inflation path comes from these sources, but we should be careful not to confuse those disturbances with underlying inflation. And for that, as I have argued, the path remains slowly but steadily down towards target. It does so in line with historical norms, and is forecast to converge smoothly without going too far into an undershoot.

As to where policy rates finally end up – the neutral level – under each of the cases, time is short and that is a topic for another day. But I will make a few comments here, where I base my view on r*, defined as the long-run real natural rate that prevails when all disturbances, real and nominal, have subsided and the economy is back at its non-inflationary and equilibrium-employment steady state.

Under the assumption of no long-run structural changes in the economy – that is, for cases 1 and 2, at least as I have defined them – I take my reference off the shelf from recent private-sector central estimates of the UK real natural rate, which are in turn based on a range of models from academic research, including my own work (Davis and others, 2024). These central estimates levels of r* range between 0.5% (PIMCO) and 0.75%–1.0% (Goldman Sachs), just a touch higher than the r* estimates of those models a few years ago. Taking from among these 0.75% as a central estimate of r*, and adding the inflation target of 2%, I would see policy rates levelling off at a terminal rate of 2.75% in steady state under case 1 or case 2.

In my view, this level of a UK terminal rate makes sense as it sits somewhere between the Fed’s terminal rate (around 3%, as suggested by the Fed’s dot plot, or around 1% in real terms) and the ECB’s (around 2%, as suggested by Philip Lane in an interview for the Money & Macro podcast in July 2024, with reference to ECB, 2024 for a range of measures around 0% in real terms) which would be consistent with the recent historical pattern and with the UK having growth and demographic fundamentals somewhere in between the US and Eurozone.

What about case 3? There, it would then remain to be specified how and why structural changes have moved r* further up, and then the size of the shift would need to be estimated or calibrated. But in Chart 8, just for illustration, I sketch out a case 3 policy path where the terminal rate is shifted up a distinctly non-trivial additional 1 percentage point. But this remains an open question, and I am less convinced by case 3 at present.

Watching out for risks of a downside scenario

Finally, I would like to stress that one important aspect of scenario analysis is that we can develop and introduce new sets of scenarios as economic conditions evolve. The cases we look at today will likely not be the relevant cases in one, five or ten years’ time. The cases will change over time as the economy evolves.

Today, I am increasingly looking mainly at cases 1 and 2 as the most likely, and I just conditioned my paths and interpretations on no shocks, up or down, relative to a continued normalisation. But that may be wrong, and the downside skew I worry about may come into play. A new case may then materialise.

Should demand conditions continue to worsen, we could have a new case: one where output weakens further, material slack opens up, disinflation accelerates, and policy restrictiveness has to be replaced by further policy easing with rates quite likely below the neutral level for a time, as in past downturns. The dashed line labelled “downside scenario” in Chart 8 illustrates what this could look like, though naturally the precise shape would depend on the depth and length of the downturn.

The risk of this case may be small as of now, but the costs, both of undershooting the inflation target and of a recession, could be very high. As a final lesson from economic history, we know that growth is negatively skewed. Most expansions are slow and cumulative, a gradual climb up the stairs; but recessions can take hold quickly, sentiment can chill, and the descent is more like taking the elevator shaft. Unwelcome output volatility can emerge quickly and in turn a risk of inflation undershooting the target, as output dynamics feed into inflation dynamics.

Now, I will stress, by no means is this downside case currently my base case, nor maybe anyone else’s base case. But the risk of this eventuality has clearly been rising.

The MPC has noted the heightened uncertainty in the outlook, and I think we need to watch carefully for signs of ebbing confidence. One can easily see possible catalysts for this adverse scenario in 2025. Some are external, where geopolitical risk, including the possibility of various trade wars, could disrupt economic activity directly via new trade frictions, or even indirectly just through the uncertainty created. The global growth environment is already challenging, with slow growth in Europe and China.

But domestic shifts also need to be watched. For me, the biggest domestic concern is what we might term the cashflow squeeze. And, in my view, this squeeze is already being felt both by businesses and households on various fronts.

In line with the OBR’s analysis of the Autumn Budget, the changes to employer’s National Insurance contributions can be seen as being a tax on employment. This tax can be absorbed on a number of margins, including prices, profits, wages, productivity, and employment, and MPC members stressed these possible impacts at our last meeting with the Treasury Select Committee in November.

I would emphasise a few important points here. First, the exact composition of these responses is very hard to predict, and we are carefully watching how the business sector reacts. Second, the full economic response isn’t just about business: even as these margins squeeze business cashflow in real terms, they also spill over into households’ real incomes. Thus, not just end demand for investment by firms but also private consumption may decline. Third, on the broader fiscal outlook, as the OBR has noted, while some spending in the Budget is rising to support demand in the short run, not far in the future the stimulus effect fades out.

But other forces are weighing on demand too, and I need to emphasise that, with its long lags, the restrictive stance of monetary policy also has weighed on UK businesses and households in 2024 and it will still add further pressure in 2025 to the cashflow squeeze in the near term. Any business trying to finance new investment after a pause during the pandemic years now faces a tougher interest-rate environment. And among households, 50% of mortgage accounts are expected to see an increase in payments by the end of 2027 (Financial Stability Report November 2024) which is expected to keep the overall household debt-servicing ratio rising for some time to come. It is therefore not the case that the effects of restrictive policy have largely played out at this point. I believe there is quite a bit more drag yet to be felt.

This is why, on multiple fronts, UK businesses and households could face a near-term cashflow squeeze, and we need to keep a careful eye on this important potential downside trigger.

If I may, on this specific point, I will end with an old Yorkshire saying, passed down in many a mining family like mine: “it all comes off t’ pick point.”

What’s that? This expression embodies a basic, common-sense economic concept: the resource constraint. The hard-won wages of piece work basically have to line up with household expenses, sooner or later. If some sudden essential costs rise, like taxes or debt service, then something else has to give.

In hard times – times of uncertainty or pessimism, when animal spirits are muted, when taking on debt or tapping into a buffer of savings seems just too risky, or infeasible – then incomings will have to match outgoings pretty tightly. Cautious households and businesses will adjust their budgeting. The resource constraint feels tighter.

That is the nub of a cashflow squeeze. Whether you were hewing away at a coal seam in the 1930s, or you are now drawing a wage packet or pension, or running a large factory or a small business – it’s still the same math, a grinding hand-to-mouth logic. It means having to manage carefully your budgets wherever you are on today’s economic coalface.

I fully appreciate these challenges for businesses and households and the headwinds they pose for the UK economic outlook, together with all the other emerging downside economic risks in the UK and around the world. Right now, I think it makes sense to cut rates pre-emptively to take out a little insurance against this change in the balance of risks, given that our policy rate is still far above neutral and would still remain very restrictive.

To reiterate: we are in the last half mile on inflation, but with the economy weakening, it’s time to get interest rates back toward normal to sustain a soft landing. It is this logic that convinced me to vote for an interest rate cut in December.

Going forward, I can also say that the MPC as a whole understands all of these risks, and collectively we will be weighing them and watching carefully, reading the data and listening as we go around the country.

And, in line with our remit, we will use our best judgement to adjust policy as necessary to fulfil our mandate. Our primary task is to return inflation to target in a timely manner and keep inflation expectations anchored. But we also need stay alert to possible trade-offs and the need to avoid undue volatility in output.

Thank you.

Thanks to Josh Martin and Vitor Dotta for help preparing this speech, to Derrick Kanngiesser, Davide Brignone and Michele Piffer for providing model forecast simulations, to Chris Giles and Joel Suss at the Financial Times for sharing their work, and to Andrew Bailey, Natalie Burr, Alan Castle, Swati Dhingra, Brian Gorst, Neha Jain, Neil Kisserli, David Latto, Catherine Mann, Huw Pill, Martin Seneca, Fergal Shortall, Andrea Sisko, and Tim Willems for helpful comments.