Surveys, forecasts and scenarios: setting UK monetary policy under uncertainty − speech by Dave Ramsden

Thank you for the invitation to speak at Stellenbosch University today. I’m visiting South Africa in my capacity as a Deputy Governor of the Bank of England, attending the bi-monthly meetings of the Bank for International Settlements, starting later today in Cape Town. This morning I’m speaking as one of nine members of the Bank’s Monetary Policy Committee (MPC), which has responsibility for setting monetary policy in the UK, with the primary objective of keeping UK inflation at 2% sustainably over the medium term.

In my speech today I want to set out how my views on monetary policy in the UK have evolved over recent months in response to my changing assessment of the outlook for the economy. That could sound like a relatively narrow focus but I hope my focus on the challenge of setting monetary policy against a back-drop of heightened uncertainties is of wider relevance.

Uncertainty is going to be a recurring theme of my speech. There are three dimensions that I’m going to bring out. The majority of my speech is going to be devoted to the prevailing uncertainty about the state of the UK economy; in particular the state of the UK labour market and the persistence of inflationary pressures. Most economies face some of the same uncertainties given the huge shocks that have hit the global economy but the UK is experiencing more than most.

The second aspect of uncertainty is about global developments, whether that be geopolitics or trade and financial fragmentation. The UK is a relatively small open economy so these matter and I will return to this aspect towards the end of my speech.

The third dimension is the impact domestic and global uncertainty has on the actions of businesses and consumers and what that means for the outlook for the economy.

The key conclusion I draw from the current elevated degree of uncertainty is that it increases the range of plausible states that the UK economy might end up in in the medium term. Hence the focus of my speech: on surveys, about the signals we can take from them about the current state of the economy; about forecasts of the most likely future state of the economy; and the importance of thinking through different possible scenarios and what they imply for the setting of monetary policy.

This approach, and particularly the role scenarios can have in informing monetary policy decisions, is entirely consistent with the recommendations of Dr Bernanke’s review^{footnote [1]} of the Bank’s approach to forecasting. I won’t say anything more today about the Bank’s wider Monetary Policy Transformation Programme, other than that I support strongly the Bank’s direction of travel, which will better equip us to set and to communicate monetary policy in an increasingly uncertain world.^{footnote [2]}

The current state of the UK economy…

Throughout 2024 the MPC’s projections consistently pointed to a gradual normalisation in the UK economy, with domestic inflationary pressures easing as restrictive monetary policy increasingly took effect to restrain demand.

Table 1: February 2025 Monetary Policy Report Forecast summary^(a)

As Table 1 shows, comparing the MPC’s forecasts from a year ago for 2025 Q1, the current quarter, with the forecasts published earlier this month, that is largely what has happened, at least when looked at in annual growth rates space. GDP growth was quite volatile around the start of 2024 but has been sluggish since, slowing a little more than expected to just over 0.1% a quarter on average through the second half of 2024. Unemployment has risen slightly as a small amount of spare capacity (or excess supply relative to demand) has opened up.

Inflation fell back to the 2% target in the middle of 2024 as past increases in energy prices partially unwound. But the MPC has consistently forecast that this wasn’t a sustainable return to the target, as energy prices would stop being a drag on inflation once this unwind dropped out of the annual comparison, and the persistence of domestic inflationary pressures would be more evident, leading to a pick-up in headline inflation. That has happened, with headline CPI inflation rising to 3.0% in January 2025 from a low of 1.7% in September.

So far so seemingly on track, which doesn’t seem consistent with the emphasis I’ve put on increased uncertainty. That’s mainly because the challenges with understanding what is going on in the UK economy relate above all to the structural performance of the economy and in particular the labour market. This is most evident from the penultimate row which shows private sector average weekly earnings growth, which is now expected to be 6¼% in 2025 Q1, similar to the level a year ago and fully 2 percentage points higher than was forecast at that time. How to interpret this development is one of the key questions I will return to in the next section.

…and the importance of surveys in informing the short-term outlook

The key issue for the MPC in formulating its forecasts has been to judge how much persistence to assume in domestic inflationary pressures. We’ve framed this judgement in terms of three cases to illustrate the range of possible outcomes on the degree of inflation persistence. In case 1, the unwinding of global shocks that drove up inflation in 2022, and the resulting fall in headline inflation, should feed through to weaker pay and price-setting dynamics meaning inflationary pressures dissipate more quickly. In case 2, on which the MPC’s forecasts have been based, a period of economic slack is required in order for pay and price-setting dynamics to normalise fully. In case 3, structural shifts in wage and price-setting behaviours, following the major supply shocks experienced over recent years, mean inflationary pressures are even more persistent.

For my part, as set out in a speech I gave last autumn^{footnote [3]}, the MPC’s forecasts throughout 2024 presented a plausible outlook, consistent with case 2. But I also believed it was at least as likely that domestic disinflationary pressures would be stronger than envisaged in the MPC’s published forecast, which would result in headline inflation staying closer to the 2% target throughout the first part of the forecast and falling below 2% more materially later on, consistent with case 1.

The evidence on the labour market has been key to the MPC’s and my assessment and in recent months there has been a significant amount of new information to absorb. This evidence takes two forms. First, some concerning developments in short term indicators, particularly on wages, as I have already highlighted. Second, how these fit into our understanding of the structure of the labour market and the economy, where staff analysis from their annual supply stock-take has fed into the latest forecasts.

This new evidence leaves me with four interlinked questions, which all featured in the MPC’s latest round of monetary policy deliberations.

• First, how resilient is employment and what is the trend in labour demand?
• Second, how has productivity growth evolved and what are the implications for the growth of potential supply or speed limit of the economy?
• Third, what is the starting level of spare capacity?
• Fourth, what are the implications for the short-term path for earnings?

The main analytical challenge in responding to the recent data is how to extract a signal from the volatility, or noise. And this is particularly the case in the UK at present where a lot of the official series are exhibiting unusual volatility. The shortcomings of the LFS data are well documented^{footnote [4]} but other official series such as the AWE and also GDP, as shown in Chart 1, appear to have become more volatile.

Chart 1: Headline GDP has been more volatile than underlying survey indicators

3-month on 3-month growth in GDP and quarterly GDP growth implied by business surveys

In approaching these questions I’ve therefore found it even more important than usual to look at a wider range of surveys and other data sources carried out in the UK by the Bank and other bodies, as well as the official data produced by the ONS. The first type asks the business or household what they have done over a period in the past. This is the basis for most official data. Some official data isn’t produced from surveys at all but is compiled directly from administrative sources, such as the RTI data compiled direct from HMRC tax returns. The second type of survey focuses more on future intentions, what households and businesses plan to do over some future period. Examples include the Bank Agents’ survey of future wage settlements. The third type focuses more on sentiment, how businesses and consumers are feeling, such as the GFK consumer confidence survey in the UK.

On the first key question, regarding the resilience of employment, the most recent data from the LFS shown in Chart 2 suggests employment grew by 0.3% in the three months to December 2024. But this series has been volatile and most alternative sources such as the HMRC RTI data suggest a weaker picture. Bank staff can pull together the signals from a range of sources into indicator-based models of the economy (using MIDAS techniques), as shown for GDP in Chart 1 and for employment in Chart 2. These models give a clearer sense of the underlying picture, as well as pointing to the immediate outlook a quarter or two ahead.

Chart 2: Resilience of employment?

Measures of employment growth

As shown in Chart 2, the Staff’s model suggests that underlying employment growth has softened in recent quarters to close to zero. It takes a signal from a range of indicators, including the composite PMI employment balance, which fell to 43.5 in February 2025, among the lowest levels since the series began in 1998. This and other more intentions-based surveys may also have been affected by the expected impact of, and reactions to, the October 2024 Budget measures, which will significantly increase the amount of NICs paid by employers from April 2025.

Turning to the second question on productivity and potential supply growth, the LFS data on the level of employment has recently been revised up significantly, to take account of the latest, higher estimates of UK population. Because there hasn’t been a corresponding revision in GDP, estimates of actual and potential productivity have been revised down significantly. Output per worker is now estimated to have been flat in 2024 and potential productivity growth is estimated by Bank staff to be 1¾% lower than estimated a year ago, and to have fallen by ½ per cent in 2024 as shown in Chart 3.

This shows a significantly different picture for the recent path of productivity, which can’t be explained by any of the usual drivers of productivity Bank Staff focus on. Stepping back, this adds to the existing puzzle of weak UK productivity growth since the global financial crisis, which has been the subject of a lot of analysis and commentary over the years^{footnote [5]}. The path of productivity growth is the key element in calculating potential supply growth, the rate at which the economy can grow without triggering inflation, effectively the economy’s speed limit.

Chart 3: Evolution of productivity?

Potential productivity per hour

This matters for addressing my third question, on the starting level of spare capacity in the economy. A key MPC judgement underpinning our February 2025 forecasts was that the majority of the recent weakness in GDP reflects the weakness in productivity and therefore in supply. That means the sluggish growth in GDP is assumed to have led to only a small margin of spare capacity opening up to date, as shown in Table 1.

Another useful way of thinking about spare capacity is through the lens of labour market tightness. A key metric I have relied on to judge this has been the vacancies to unemployment ratio. We have seen vacancies decline markedly from their peak in mid-2022 and the ratio of vacancies to unemployment shown in Chart 4 is now back to levels seen in 2018. This is consistent with other indicators on labour market tightness and has led the MPC to assess the labour market to now be in broad balance. But with the pace of decline in the V/U ratio and the weakness evident in surveys of employments intentions, labour demand could continue to ease much more materially in the near future.

Chart 4: Emerging spare capacity in the labour market?

Vacancies to unemployment ratio and estimated equilibrium values

What then are the implications of these developments for wages, my fourth question? Whilst earnings growth was still elevated in 2024, it had been steadily declining. However, in 2024 Q4 annual growth in private sector AWE rose to 6.2% from 4.9% in the three months to September. And as Table 1 showed it is forecast to stay at that level in 2025 Q1, a full two percentage points higher than the forecast a year ago.

This is a big forecast error in a key variable. And while Chart 5 shows the official data has again been volatile, the Bank Staff’s indicator model has also stopped falling. I am not discounting the signal from these indicators but they are a relatively backward looking gauge, relating to the earlier tightness of the labour market.

In assessing where things are heading in the future I place more weight on the information we have from the Bank of England’s Agents’ Pay Survey^{footnote [6]}. This survey has a good track record as a predictor of how wages are likely to develop over the year ahead, as shown in Chart 6.

Chart 6: The Agents’ pay survey has a good track record

Measures of annual pay settlements

The latest Agents’ Annual Pay Survey, carried out late last year, suggests average expected pay settlements in 2025 of 3.7%. That’s quite a bit higher than the bottom half of the 2%-4% range, where I judged expected settlements were likely to settle. But it still marks a notable decrease from the average actual settlements figure of 5.3% in 2024. The MPC has taken a strong signal from this survey, as well as from the Bank’s Decision Maker Panel survey, for the first year of the earnings forecast, which I will return to in the next section.

The MPC’s latest forecasts for the UK economy

Although there remains considerable uncertainty about the current state of the UK economy the emerging answers to my four key questions help establish a starting point for the MPC’s forecasts for the next three years.

The MPC’s latest forecasts, which I summarised in Table 1, set out the most likely path for the economy over the medium term and provide the framing for our policy decisions. They are conditioned on a set of assumptions, and are produced using a range of forecasting models and techniques which go much wider and deeper than the indicator based statistical models I presented in the last section.

The majority of our forecasting models are estimated on past relationships between economic variables, and so rely on the past being a good guide to the future. Even at times of certainty and stability these models will perform less well the further into the future one goes. But at times when the economy is experiencing or dealing with the aftermath of significant shocks, when previously established relationships change, forecasting performance will deteriorate significantly. This is why the judgements made by the MPC, and all other forecasters, have become increasingly important in recent years.

As I’ve stressed, the key judgement for the MPC has been to decide the degree of domestic inflation persistence to assume; since last spring the MPC has judged there will be additional persistence during the first two years of the forecast over and above what previous experience would suggest. A reasonable question to ask is whether over the last year the evolution of inflation, and in particular the disinflationary process has been in line with this judgement.

To assess this I find it useful to focus on the relative pace of the increase and subsequent fallback of the different components or waves of CPI inflation. What is apparent from Chart 7 is the symmetry in the rise and fall of each component, with the exception of inflation in private sector rents.

The major rise and fall of the waves for energy, food and core goods has ended, with the recent pickup in inflation in these components. But the easing in services inflation, the component which most relates to domestic inflationary pressures, continues. Services price inflation did rise to 5.0% in January 2025 up from 4.4% in December, though that largely reflected a rebound in volatile airfares prices.

Adding all this up and based on current short-term trends and known developments we expect headline inflation to rise further in the immediate future, and peak at around 3.7% later in the year. One key driver is the further increase in household energy bills from April 2025. Regulated prices are also increasing, and, alongside the impact of government policies announced in the Budget last Autumn, will also add to headline inflation by the middle of the year.

In determining its inflation forecasts beyond the short term, set out in Table 1, the MPC has judged that the latest pickup in headline inflation will not lead to additional second round effects on underlying domestic pressures. This is clearly an important new judgement, and it is borne out of the assessment of the evidence which I have set out, in particular the looser labour market compared with the time of the very large external costs shocks. The pickup in inflation is also expected to be much smaller than in 2021-22.

This suggests that some of the dynamics in the inflation generating process that were observed following the previous large shocks are unlikely to reoccur. Given this context, as already highlighted the MPC has taken a strong signal from the Agents’ Pay survey and the DMP survey, such that earnings growth is forecast to continue to slow to around 3¾% by the end of 2025, as the labour market continues to ease and medium-term inflation expectations stay anchored. The challenges of forecasting earnings and, as a consequence, the extent of the signal taken from the surveys is illustrated in Chart 8. This shows that the 2024 Q4 official data on earnings, shown by the yellow line, was more than 2 percentage points above what could be explained by the Bank’s three main forecasting models for earnings, shown by the swathe.

Reflecting these judgements earnings growth is forecast to fall back to closer to inflation target consistent rates by the middle of 2026. Services are very labour intensive, so wages are the key determinant of services inflation.^{footnote [7]} This easing in domestic inflationary pressures along with a diminishing contribution from energy prices is what drives the forecast for headline inflation to fall back to close to the 2 per cent target from the end of 2026 as shown in Chart 9.

In terms of the forecast for activity, shown in Table 1, the sluggish growth in GDP is judged to have led to only a small amount of spare capacity opening up as growth in productivity and the supply capacity of the economy is assessed to have weakened sharply over the past year. GDP growth is forecast to pick up steadily from the middle of 2025 with growth of 1.5% in the year to 2026Q1. Continued household real income growth, as real wages continue to increase in aggregate, plus a small reduction in household savings rates supports private consumption growth. Housing activity is picking up which should support housing investment over time and business investment growth remains positive. In addition, there is a contribution of about ½% of GDP to growth from the higher public spending included in the October 2024 Budget. As I’ve already stressed there are considerable uncertainties around the path for potential supply, but the MPC has assumed that the recent weakness in productivity is not sustained and potential supply growth picks up. As a result, excess supply capacity increases from ¼% in 2025Q1 to ¾% in 2027.

The forecast remains consistent with case 2 that the MPC has identified. The emerging margin of spare capacity in the economy acts against the continuing second round effects in domestic prices and wages to help bring inflation back to target in the medium term. Restrictive monetary policy is also continuing to do its job; it’s noteworthy that the latest forecast is conditioned on market expectations for UK monetary policy which have tightened materially over the last year, with the implied path for the level of Bank rate in 2025 Q1 0.7 percentage points higher than was the case a year ago. The implied level of Bank rate at the end of the forecast period in 2028 Q1 is 4%.

Stepping back, I think the MPC’s latest forecast represents a realistic assessment of the outlook for the UK economy. But given the evidence that has emerged in recent months and the increasing uncertainties about the state of the UK economy and global developments it is clearly not the only outcome. The MPC has been collectively thinking about various scenarios throughout the last few years, as well as the more specific three cases framework more recently.

As part of my thinking and approach to the MPC’s latest forecasts and the February policy decision I’ve found it helpful to sketch out two very different qualitative scenarios for the UK economy, from the huge number of possible futures, which build on the MPC’s existing cases framing for inflation persistence, and explore the interactions with judgements on the labour market and the demand and supply sides of the economy. In headline terms I think of them as a Weaker Demand and a Weaker Supply scenario.

Taking each in turn, weaker demand could emerge for example due to increased uncertainty about the outlook for the economy; the third dimension of uncertainty I highlighted at the start of my speech. Household saving might stay at the same high level for precautionary reasons and so not help support consumption growth. And business investment growth might not be sustained if businesses postpone investment in the face of growing uncertainty. If demand is weaker relative to supply then increased spare capacity would exert more downwards pressure on wages and prices. This process could either already be in train or it could quickly emerge if labour demand weakens more materially, more in line with the signal from some of the employment surveys. In addition, the impact in terms of reducing employment of the October 2024 Budget increase in NICs could be greater than assumed, relative to the other margins of adjustment assumed to be at work through wages, prices and profits.

I would be concerned under these circumstances that the interaction of weaker demand with a loosening labour market could be multiplicative, opening up an excessive degree of slack. This would result in wages slowing more sharply even than assumed under case 1, and inflation falling back to the 2% target more quickly and undershooting it in the second half of the forecast. Here the appropriate monetary policy response would be to reduce Bank rate more rapidly than otherwise. My concerns about the risks from this scenario playing out were what led me to vote in the minority for a 25bp cut in Bank rate at our December 2024 meeting.

Turning to the Weaker Supply scenario, it is entirely possible that current supply capacity is even weaker than we assessed it to be and may be more persistently weak through the forecast. This could result from the starting point for supply being lower, followed by a much weaker recovery in productivity than the MPC has assessed. A range of structural factors could explain this weakness. In those circumstances, the speed limit of the economy would be even lower, the labour market may remain tighter for longer, and wages could remain elevated leading to greater persistence in domestic inflationary pressures.

These are similar processes to those at work in the third case for the economy that the MPC has been considering. But if this was more clearly the context for the near-term increase in inflation the UK is experiencing then it would be more likely to lead to additional second round effects. In this scenario, policy would have to be more restrictive than otherwise to guard against the risk of medium-term inflation expectations becoming de-anchored, in response to a more prolonged period of above target inflation.

These scenarios stem from domestic uncertainties but increased uncertainties in the global economy also need to be factored in. Geopolitical uncertainties have been heightened for several years. In addition, there is significant uncertainty around the next steps the current US administration will take with regards to tariffs, and the potential actions of other countries in response. At present we do not have any specific enough information to build into our forecasting models. However, it is clear that uncertainty about global trade policy is already heightened. Chart 10 shows that, on one measure, trade policy uncertainty had already reached record highs in January 2025. And this may already be impacting on the global and UK economy via financial markets and via confidence channels.

Chart 10: Trade policy uncertainty has reached record highs

Share of articles in selected publications discussing trade policy uncertainty

At this stage, as detailed in the February 2025 MPR, we have focused on the main potential channels for the impact.^{footnote [8]} On combining these channels our clearest conclusion is that greater trade fragmentation will likely lower UK economic activity, as barriers to trade inherently weigh on global demand, weakening demand for UK exports, and disrupting supply chains. But the impact of US tariffs could be inflationary or disinflationary for the UK depending on other countries’ trade policies and the relative strength of different transmission channels. Depending on the specifics of the trade policy, and then how the UK and others respond to that policy, prices in the UK could go up or down.^{footnote [9]}

Conclusions and implications for monetary policy

Before concluding, I think it is important to be reminded of where we are starting from. UK headline CPI inflation was 3% in January, above the 2% level it averaged from April to September last year, but only slightly above the rates in the US and Euro area, which have also risen recently. UK inflation is expected to rise in the short term to around 3.7% and is then forecast to fall back to close to the 2% target in the second half of the forecast period.

To conclude, in assessing the state of the UK economy, compared with my position throughout last year I am now less certain than I was about the outlook for the UK labour market, and its implications for future inflation persistence and growth. Because of the evidence of recent months I no longer think that risks to hitting the 2% inflation target sustainably in the medium term are to the downside. Instead, I think they are two sided, reflecting the potential for more inflationary as well as disinflationary scenarios. I do, though, think the core disinflationary process remains intact.

Given the increased uncertainty and risks to inflation on both sides – from the near-term outlook to inflation, and from developments in the global economy – I am even more certain than I was that taking a gradual and careful approach to the withdrawal of monetary restraint is appropriate. This means judging the evidence afresh at each meeting to ensure the MPC sets monetary policy to meet the 2% inflation target sustainably in the medium term.

As a keen mountain climber, I was very pleased when the MPC’s policy of maintaining Bank rate at a level of 5.25 percent from August 2023 to August 2024 became associated with South Africa’s Table Mountain. The MPC is now on the descent path having cut Bank rate three times by a total of 0.75% since August 2024, most recently by 25bp at its February 2025 meeting. I know from my own experience that great care needs to be taken on the descent from a mountain; tiredness often sets in, concentration can lapse, obstacles have to be bypassed and at the end of a long day the weather can deteriorate. A gradual and careful approach is always needed on the way down a mountain to ensure a safe descent and a successful outcome. But that doesn’t always mean the descent has to be slow. There may be circumstances when a slower than expected descent is justified but there will also be times when conditions require that the pace has to quicken.

With thanks to Ed Kent for his assistance in preparing these remarks, and to my fellow MPC members and colleagues, including Hassana Babangida, Alan Castle, Harvey Daniell, Giulia Gardin, Josh Lillis, Dan Steel, Michal Stelmach, Lukas von dem Berge and Jamie Walkington for their helpful contributions.