Forecast Evaluation Report – January 2026

Foreword

On the basis of his thorough review of the Bank of England’s (Bank’s) macroeconomic forecasts and their role in policy preparation and communication (Bernanke (2024)), former Federal Reserve Chair Ben Bernanke offered several recommendations to improve the forecasting process underlying the decisions of the Monetary Policy Committee (MPC).

One such recommendation included the suggestion that ‘the staff should be charged with highlighting significant forecast errors and their sources, particularly errors that are not due to unanticipated shocks to the standard conditioning variables. Models and model components that may have contributed to forecast misses should be regularly evaluated and discussed’.^{footnote [1]}

This Forecast Evaluation Report is one part of the Bank of England’s response to Bernanke’s recommendation. Building on previous work by Bank staff,^{footnote [2]} the Report undertakes a thorough statistical evaluation of the accuracy, unbiasedness and efficiency of the forecasts published in the MPC’s regular Monetary Policy Reports. It also presents a more detailed narrative of forecasting experience and performance over recent years.

Bank staff plan to develop and extend these analyses in future years, with subsequent published reports to follow. To support that evolution and provide an opportunity for external experts to offer their insights on Bank staff’s underlying technical analysis, the Bank is also publishing a Macro Technical Paper describing the tools that have been used to undertake the statistical analysis (Abiry et al (2026)) and the code used in the implementation of the approach^{footnote [3]} in parallel with the release of this Report. The Bank welcomes feedback on both the Forecast Evaluation Report itself and the underlying technical material and code.

To ensure this evaluation matches best practice, Bank staff have embodied three key elements into their analysis. First, Bank staff have recognised the importance of adopting a ‘real-time’ approach, employing only the data that were available at the time the forecast was made when benchmarking against alternatives (and highlighting the role of data revisions). Second, Bank staff have recognised the role played by the MPC’s conditioning assumptions in driving forecast profiles and thus influencing forecast errors. (In the forecasts published in the MPC’s Monetary Policy Report, one such conditioning assumption is the future path of Bank Rate, which is derived from market-implied expectations.) And third, Bank staff have recognised that new shocks will inevitably affect the economy in the period between making any forecast and the outturn against which it is judged.

By recognising these three potential sources of forecast errors, Bank staff’s evaluation allows more focus to be placed on those errors that arise from potential shortcomings or gaps in the Bank’s models and analysis (rather than errors owing to unanticipated shocks, which cannot be entirely avoided). In real time, such model or analytical shortcomings are impossible to remedy entirely, but there is much to learn from evaluation of those errors with the benefit of hindsight. As is apparent from the quotation from his review above, this approach is very much in the spirit of Bernanke’s recommendation.

Even as the role of the Bank’s forecasts in preparing monetary policy decisions evolves (and the forecasts are supplemented by other complementary analyses) in line with Bernanke’s other recommendations (Bailey (2025)), an evaluation of how the forecasts published in the MPC’s Monetary Policy Reports have performed as predictors of economic outcomes can help identify strengths and weaknesses in the Bank’s understanding of the UK economy (and allow Bank staff to refine their approach to forecasting in the future). Embracing the feedback implicit in forecast errors lies at the heart of the Bank’s efforts to ensure monetary policy making in the UK is conducted within ‘a culture of continuous learning’ that leads over time to improvements in the MPC’s and Bank staff’s understanding of the UK economy and the transmission of monetary policy within it (Lombardelli (2024)).

From a policymaking perspective, the over-riding question is not solely, or even mainly, whether a forecast performs well on the statistical criteria of accuracy, unbiasedness or efficiency. Rather, a forecast that supports the monetary policy process needs to help the MPC form a view of the economic outlook, to prompt a rich discussion among MPC members that leads to appropriate monetary policy decisions, and to help communicate its decisions consistently with those views both internally within the Bank and to external stakeholders. The MPC’s job is to set good policy to meet the inflation target, and that is how it should be judged. Forecasts – along with other analytical inputs – are means to that end.

Learning from forecast errors not only helps us make better forecasts, it also helps us to develop a better understanding of the economy. It is that better understanding that is crucial to enhancing the analytical inputs that underpin the MPC’s monetary policy decisions. Publication of this Report represents one of many important steps in that learning process.

Huw Pill

Chief Economist and Executive Director for Monetary Analysis and Research

1: Executive summary

The Bank of England’s Monetary Policy Committee (MPC) sets monetary policy to achieve the inflation target of 2%.

The MPC’s policy decisions are informed by analysis prepared by Bank of England (Bank) staff. An important component of that analysis is forecasts of the economy. Forecasts will remain an important element of this input to the MPC’s policy decisions, even as their role evolves in the light of the Bernanke Review of forecasting for monetary policy making and communication (Bernanke (2024) and Dhami et al (2025)). Following that review, Bank staff are widening the range of inputs presented to the MPC, in part to include more regular assessments of risks, uncertainty, and alternative policy paths.

The nature of forecasts presented in the Monetary Policy Report has also changed. These forecasts are projections, based on a staff proposal, that a majority of the MPC agrees are reasonable baselines, rather than representations of the MPC’s ‘best collective judgement’ as was the case previously (Bailey (2025)).

Even as the role of the Bank’s forecasts evolves and they are supplemented by other complementary analyses, an evaluation of how these forecasts have performed as predictors of economic outcomes can help identify strengths and weaknesses in the Bank’s understanding of the UK economy. This Report finds that the Bank’s forecasts have been at least as accurate as the average of external forecasters or a range of alternative model-based approaches over the past decade.

Where forecasts have underperformed, analytical investment can improve the Bank’s understanding of specific sectors or macroeconomic relationships. Evaluating forecast performance is therefore one way to prompt the ‘continuous learning’ about the structure of the economy that the Bernanke Review recommends (Lombardelli (2024)).

This Forecast Evaluation Report draws lessons from both analysis of historical and recent performance.

Among several recommendations to improve the Bank’s forecasting and policymaking processes, Bernanke suggested that Bank staff should identify and explain forecast errors.^{footnote [4]} This Report addresses that suggestion, drawing lessons from both statistical analysis of longer-term forecast performance and a more detailed analysis of recent experience.^{footnote [5]} It evaluates the Bank’s forecasts of four key policy-relevant variables: CPI inflation, GDP growth, wage growth and the unemployment rate. CPI inflation is the MPC’s target variable. GDP, unemployment and wages relate to key channels in the domestic inflation generation process according to economic theory. The interactions between these variables can also shed light on structural features of the UK economy. While these four indicators are important, they do not represent an exhaustive list of the variables analysed when formulating monetary policy. Future Forecast Evaluation Reports may explore different features of the economy and include coverage of different variables as a result.

This Report assesses the historical performance of the Bank’s forecasts using a range of statistical methods.

This Report evaluates the statistical properties of the Bank’s forecasts along three standard dimensions: accuracy, unbiasedness and efficiency. Accuracy metrics consider the magnitude of forecast errors, ie whether the forecast has typically been close to the realised outcome. Unbiasedness examines whether errors have been evenly distributed in both directions, or instead have tended to lean in a particular direction, ie whether forecasts have tended to be systematically too high or too low. Efficiency investigates whether forecasts have made effective use of information, ie whether the forecast could have been improved upon by better exploiting information available at the time it was made.

While a good forecast in statistical terms is one that performs well on these three dimensions, from a policymaking perspective the crucial issues are whether the forecast helps the MPC form a view of the economic outlook, prompts a rich discussion among the MPC that supports appropriate monetary policy decisions, and helps communicate its decisions consistently with those views.

Forecast accuracy is therefore not the only or most relevant metric for assessing the value of forecasts in the policymaking process. This is particularly the case for the forecasts presented in the Monetary Policy Report, which are conditioned upon certain assumptions for key variables (including the future path of Bank Rate) that may not represent the MPC’s own views of how those variables will evolve. Nevertheless, valuable lessons can be drawn from this statistical analysis.

The Bank’s forecasts have typically performed at least as well as a range of alternative forecasts over time, and across the selection of variables reviewed here.

Bank forecasts are compared with a range of alternatives to permit an assessment of their relative performance. Among the alternatives considered we include forecasts constructed using an autoregressive (AR) model, a Bayesian vector autoregression model (BVAR), a mechanical application of the Bank’s dynamic stochastic general equilibrium (DSGE) model known as COMPASS (Albuquerque et al (2025)), and a survey of external forecasters (SEF) (Box A). By making a relative comparison, this exercise controls for the impact of changes in uncertainty or volatility on forecast errors, since all approaches are subject to these in the same way.

This comparison shows that none of the alternative forecasts considered would have outperformed the Bank’s forecasts over the past decade. While the accuracy of the Bank’s forecasts has reduced since 2020, that is also true of the range of alternative benchmarks. This result suggests that much of the deterioration in forecast performance since Covid can be attributed to heightened general economic volatility rather than specific deficiencies in the Bank’s approach or framework. This result is also consistent with the Bernanke Review, which found the deterioration in forecast accuracy since the pandemic had been similar for the Bank as for other major central banks (Bernanke (2024)).

The statistical analysis presented here identifies areas for potential improvement, particularly for labour market variables.

The relatively good performance of the Bank’s forecasts on average over a long historical window does not mean that there is no room for improvement or nothing to learn. Looking at forecast performance on a variable-by-variable basis, the Bank’s forecasting of labour market variables – specifically wage growth and unemployment – is an area that would benefit from further work. Forecasts for wage growth have exhibited signs of bias either side of the Covid pandemic, while the unemployment rate has been overpredicted since 2015. Statistical tests also point to inefficiencies in the Bank’s GDP, wage and unemployment forecasts. Related Bank research has also flagged potential for improvement in the economic relationships embedded in the Bank’s forecast machinery that may help account for this forecast inefficiency (Kanngiesser and Willems (2024)).

Unforeseeable events explain some of the largest forecast errors since Covid, but the persistence of inflation is harder to reconcile with standard forecast treatments.

This Report also explores some of the drivers of the Bank’s forecast errors over the past five years in greater detail, setting out seven key findings. As noted by Bernanke (2024), a series of large surprises – or ‘shocks’ – made forecasting more challenging than usual during that period, leading to larger forecast errors for central banks around the world. Model-based simulations – or ‘counterfactuals’ – of what the Bank’s forecasts would have been had it known about some of those shocks in advance can be used to estimate their contribution to forecast errors.

The set of shocks that hit during this period can be characterised with a model-based decomposition of the economic drivers that account for the deviations of inflation from the 2% target (Section 5). Adapting a model developed by Bernanke and Blanchard (2025) to the UK (Haskel, Martin and Brandt (2025)) reveals a prominent role for those shocks. These included the aftermath of the Covid pandemic, as bottlenecks in supply chains contributed to elevated global export prices; Russia’s invasion of Ukraine which drove up energy and food prices globally; tightness in the UK labour market which has driven up wages and inflation; and weakness in productivity which has pushed up on firms’ costs.

Surprises in energy prices and other global factors can broadly account for the peak inflation error. In late-2022, CPI inflation reached a peak of 11.1 per cent, more than 8 percentage points above forecasts from mid-2021, before Russia invaded Ukraine. Counterfactual analysis suggests that subsequent energy price rises can account for around half of this error (key finding 1). Other global factors explain most of the remaining error (key finding 2). Partly offsetting this, and in response to higher inflation, Bank Rate rose further than assumed in the Bank’s forecasts (which use financial market expectations for the future path of Bank Rate). And there is some evidence to suggest that tighter monetary policy affected the economy more quickly than in the past (key finding 3).

While the impacts of these external shocks can broadly explain the peak errors, they cannot fully account for the subsequent persistence of inflation. After 2022, the Bank’s medium-term inflation and wage growth forecasts proved repeatedly too low. Some of this strength in inflation likely reflected ‘second-round effects’ from higher inflation feeding back to stronger wage growth. It is also possible that there have been more long lasting, or ‘structural’ changes in price and wage setting in recent years (key finding 4).

GDP growth has been stronger than forecasts expected on average since 2021, which (along with slightly lower estimates of potential supply that we do not evaluate directly in this Report) has contributed to there being less spare capacity in the latest estimates than initially projected (key findings 5 and 6). Forecast performance for GDP growth – and for labour market variables – appears more favourable when assessed using initial data releases than when using later data vintages, but the latter may be a better reflection of the ‘truth’ as more information is incorporated into official statistics over time (key finding 7).

These findings will continue to inform improvements to the Bank’s forecast models and processes, as well as other inputs to policymaking.

A number of changes have already been made that address the results presented in this Report. For example, the treatment of energy within the Bank’s forecast machinery has been improved based on the experience of recent years (Albuquerque et al (2025)). The speed of monetary transmission within the Bank’s policy analysis toolkit has been adjusted to match the latest Bank staff analysis (Alati et al (2025)). And internal processes for monitoring supply developments have become more frequent, acknowledging the increasingly supply-driven character of economic fluctuations.

Future model development will seek to improve the Bank’s modelling and understanding of key economic mechanisms, including the labour market, wage-price interactions and inflation expectations, which should help to explain recent inflation persistence.

The remainder of this Report starts by giving an outline of how the Bank’s forecasts are produced and the role they play at the Bank (Section 2). It then gives an overview of the forecast evaluation approach undertaken (Section 3). It follows with a summary of some of the longer-term statistical properties of the Bank’s forecast (Section 4) before focusing in on some of the key findings from more recent forecast errors (Section 5). It finishes with some forward-looking takeaways (Section 6).

2: Forecasting at the Bank of England

2.1: How the Bank’s forecasts are produced

The Bank’s forecasts are constructed using a broad set of models, information and judgement.

The Bank produces forecasts for a range of macroeconomic variables that are relevant for monetary policy makers. These forecasts are produced four times a year and published alongside the quarterly Monetary Policy Report.

The Bank’s forecasts draw on steers from a range of models, data, surveys, analysis and intelligence from the Bank’s Agency network. These sources are also often supplemented by Bank staff and MPC judgements to adjust for model limitations or information that models may not account for directly. The precise approach and models underpinning the production of the Bank’s forecasts are often updated and have changed over time. Table 2.A gives an overview of how these forecasts are constructed currently.

The forecasts are conditional projections.

The Bank’s forecasts are conditional in nature, as they take several input paths as given. Conditioning paths for Bank Rate, world interest rates, global commodity prices, the exchange rate,^{footnote [6]} and wholesale energy prices are based on market expectations. Non-wholesale energy costs are assumed to evolve in line with Office of Gas and Electricity Markets (Ofgem) projections. Fiscal policy is assumed to follow announced government policy, as reflected in the Office for Budget Responsibility’s (OBR’s) projections.

An implication of this approach is that the resulting projections may not always correspond to the MPC’s single most likely expectation of the outcome. For example, if the conditioning paths for one or more of these variables did not represent the MPC’s own views of how they will evolve. Temporary inconsistencies between conditioning paths may also arise which can lead to forecast errors. For instance, if fiscal policy was widely anticipated to loosen relative to previous policy announcements reflected in the OBR’s projections, financial market participants might expect a higher Bank Rate path. As the Bank’s forecasts condition on these financial market expectations, that might show up as a drag on the forecast for GDP growth.

The near-term outlook is informed by the latest data and statistical models.

The first two quarters^{footnote [7]} of the forecast are informed by statistical models rather than the Bank’s structural forecasting framework. These models are purely data driven, extrapolating recent time-series dynamics and incorporating high-frequency information, and tend to deliver greater accuracy than structural models for short-term forecasting (Giannone et al (2016)). The first two quarters are treated as ‘constrained’ inputs to the medium-term forecast. For inflation, Bank staff’s preferred short-term forecasting approach relies on a collection of unobserved component regressions, modelling the time series properties of detailed CPI components (Esady and Mate (forthcoming)). For GDP and labour market variables, a staggered-combination mixed-frequency (SC-MIDAS) model combining information from survey and official data is used (Moreira (2025)). A range of other statistical models are also used regularly as complements or cross-checks to these.

The Bank’s medium-term forecasts are constructed using a structural economic model as the central organising framework…

The remainder of the Bank’s three-year forecasts, also referred to as the medium-term forecasts, are anchored in a structural model. Unlike the data-driven statistical models used to pin down the near-term outlook, these models try to capture the behaviour and interactions among households, firms and other economic actors.

Standard economic theory posits that there are some structural features of the economy that help determine the broad interactions between the variables evaluated in this Report.

• Fluctuations in GDP growth can reflect supply- and/or demand-type ‘shocks’ that cause the level of activity to deviate from its underlying ‘potential’.
• Unemployment responds to cyclical changes in GDP via an ‘Okun’ relationship, as well as structural changes to the ‘natural’ (or non-inflationary) rate of unemployment.
• Wages respond to demand conditions via a wage ‘Phillips curve’, as well as a role for behavioural relationships between workers and firms, and structural factors.
• Inflation is affected by demand conditions via a price ‘Phillips curve’, as well as cost-push supply shocks and behavioural dynamics, with a key role for the labour market.

Monetary policy makers need to take account of these relationships to achieve the inflation target, although economic models differ in their treatment of these relationships and all are a simplification of reality. Regular evaluation of forecasts for these variables can help shed light on these relationships and their changing nature.

While the approach to modelling these interactions has evolved over time, the current central organising framework for the Bank’s forecasts is COMPASS, a medium-sized, open economy New Keynesian DSGE model providing a coherent framework for understanding macroeconomic relationships and policy effects (Burgess et al (2013)). COMPASS has been updated over time with the latest iteration including an improved modelling of the role of energy (Albuquerque et al (2025)).^{footnote [8]}

…and incorporate forecasts from a range of other models that better capture certain features of the economy, alongside Bank staff and MPC judgement.

As no model can perfectly capture all relevant features of the economy, COMPASS is also augmented by a range of more specialised models. For example, COMPASS is supplemented by the ‘Post Transformation’ model. This takes outputs from COMPASS, along with other data inputs, and uses simple accounting and statistical relationships to produce forecasts for a wider set of variables, such as the unemployment rate.

A semi-structural model, known as the ‘sectoral model’, is used to capture the transmission of monetary and financial conditions to different sectors of the economy (Cloyne et al (2015)). Wage growth forecasts are informed by a suite of models linking them to their key economic drivers (Bank of England (2024)).

Forecasts for key supply-side variables are also produced using a combination of COMPASS and a suite of more targeted models. Given the UK is an open economy, Bank staff forecasts of international developments are another important input into the Bank’s UK forecast.

Finally, a range of cross-check models are also used regularly to help inform judgements and in some cases also interpret changes in the Bank’s forecasts (for example, Brignone and Piffer (2025)).

2.2: The role of forecasts in the MPC’s policy processes

The forecasts are projections that a majority of the MPC agrees are reasonable baselines.

The Monetary Policy Report forecasts are a central projection, based on a staff proposal, that a majority of the MPC agrees are ‘reasonable baselines’. Previously, these forecasts represented the MPC’s ‘best collective judgement’ (Bailey (2025)). This status reflects recent changes to the MPC’s approach to monetary policy making under uncertainty (Dhami et al (2025)).

The central projections in the MPC’s reasonable baseline include profiles for CPI inflation, GDP growth, the unemployment rate, and the output gap (a measure of the level of actual output relative to potential). Indicative projections for a range of other variables (including wage growth) are produced by Bank staff to be broadly consistent with these core central projections.^{footnote [9]}

As part of the Bank’s response to the Bernanke Review, forecasts are being increasingly supplemented by other inputs to policymaking.

The role of the forecasts in the MPC’s policymaking process has changed over time, most recently in response to the recommendations in the Bernanke Review. In the past, the forecast had been more central to the policymaking process, serving as the primary vehicle to help the MPC organise and consolidate its understanding of the economic outlook, as well as playing a central role in policy communications. Reflecting the need for greater flexibility in an uncertain environment, the MPC now emphasises that the forecast is one of several regular inputs to policymaking and communications, alongside a richer assessment of risks, uncertainty and alternative policy paths (Figure 1 in Dhami et al (2025)).

2.3: The importance of forecast evaluation

Forecast evaluation can lead to better-informed monetary policy decisions.

Given the relevant role the forecast still plays for monetary policy makers, regularly reviewing the differences between realised and projected outcomes – or forecast errors – is valuable (Lombardelli (2024)). By helping to identify the need for improvements to the Bank’s models, processes and judgements, as well as its understanding of the economy, regular forecast evaluation can lead to better informed monetary policy decisions. This is also important to support transparency around the policymaking process.

Forecast errors can arise from a range of sources and some are unavoidable, reflecting unforeseeable events.

Forecast evaluation can help to identify whether forecast errors have originated from unforeseeable circumstances, model shortcomings, or errors of judgement. This is important because not all errors imply problems with the forecast process. A significant proportion of errors can instead reflect unforeseeable events that hit after forecasts were made. An example from 2022 is when Russia’s invasion of Ukraine drove a significant rise in wholesale gas prices, leading to significantly higher inflation than forecast previously.

Systematic evaluation may prompt improvements in modelling and judgements…

Other forecast errors can stem from model mis-specification, for example using incorrect assumptions for a key parameter or economic mechanism. By flagging systematic errors in certain variables or relationships, forecast evaluation can help target upgrades to the Bank’s forecasting infrastructure. This may include investing in the main models feeding into the Bank’s forecasts or developing alternative models. Where model-based improvements are not immediately available, forecast evaluation can inform judgemental adjustments instead.

…or identify areas where the Bank’s understanding could be improved through the wider range of inputs to policy, including scenarios and assessment of risks.

Evaluation of the Bank’s forecasts can also improve understanding of how past economic shocks have propagated through the economy and to what extent they might continue to affect the economic outlook. Moreover, it can help to identify new trends in the economy such as changes in households’ or businesses’ behaviour. It can also help to flag emerging areas of uncertainty, which can in turn inform where scenarios or further exploration of risks could be particularly useful to support monetary policy decisions.

Various changes across time provide important context for the evaluation of the Bank’s forecasts.

A range of changes to the forecast process described throughout Section 2 mean that the forecasts evaluated in this Report are from a period when the Bank’s forecasts were both created and used differently. Some of the lessons from past forecast errors have also been incorporated in the Bank’s forecasting approach, as discussed in Sections 5 and 6 of this Report. While this is unlikely to impact the Report’s broad findings, it may mean that some of its results are no longer fully representative of the Bank’s forecasting capability today.

3: Overview of forecast evaluation approach

This Report evaluates the Bank’s forecasts of four key macroeconomic variables.

The Report evaluates the Bank’s ‘modal’^{footnote [10]} point forecasts for year-on-year (or four-quarter) CPI inflation, GDP growth and wage growth,^{footnote [11]} as well as the level of the unemployment rate.

These variables were chosen as they are of particular importance for monetary policy. The MPC’s target is for CPI inflation of 2%. GDP growth and the unemployment rate can help inform the assessment of whether spare capacity and associated inflationary pressures are building or fading. Wage growth has been monitored particularly closely in recent years, given the role of wages as a key driver of domestic inflationary pressures.

Bank staff have developed a new toolkit to support regular forecast evaluation.

To help embed systematic forecast evaluation, as well as support the production of this and future Forecast Evaluation Reports, Bank staff have developed a new forecast evaluation toolkit, detailed in an accompanying Macro Technical Paper (Abiry et al (2026)). The codebase and the underlying data are also being made available on the Bank of England’s public GitHub repository as an accessible package in the Python programming language.

The toolkit draws on best practice in data science to handle a large volume of data and forecast vintages, enabling evaluation of the Bank’s forecasts based on real-time information. This includes the ability to benchmark the Bank’s forecasts against a range of model-based forecasts produced on the basis of that same real-time information.^{footnote [12]} The toolkit implements the full range of statistical evaluation techniques used throughout Section 4 and can be extended with other evaluation approaches, variables or benchmark model comparisons.

This Report seeks to draw lessons from analysis of both historical and recent errors.

Previous forecast evaluations have tended to either focus on long evaluation windows (Independent Evaluation Office (IEO) (2015)) or zoom in more narrowly on recent errors (such as those included in the Monetary Policy Report),^{footnote [13]} but not both. This Report brings together insights from those two perspectives to draw appropriate lessons for the Bank’s forecast processes and modelling.

Section 4 establishes some longer-term statistical properties of the Bank’s forecasts over the past decade (2015–25). Section 5 deep-dives into more recent errors and their drivers, drawing on a range of more targeted approaches to highlight seven key findings.

The Report assesses historical forecast performance against traditional statistical evaluation metrics and alternative forecasts…

Longer-term forecast evaluation metrics can help to identify persistent issues with the forecast and place more recent errors in context. Section 4 draws on a range of standard statistical concepts of forecast accuracy, unbiasedness and efficiency to evaluate forecasts over 2015–25, at four horizons: the current quarter in which the Monetary Policy Report is published (‘Year 0’), and one, two and three years ahead (‘Year 1’, ‘Year 2’ and ‘Year 3’ forecasts). The latest forecast in scope for this evaluation is from August 2025. Technical details of these statistical approaches are laid out in Abiry et al (2026).

This Report also benchmarks the performance of the Bank’s forecast against a range of alternative model-based forecasts and projections from external forecasters (Box A). Comparison to these benchmarks helps to assess the performance of the Bank’s forecasts and identify whether they add value beyond alternative approaches. By analysing the relative performance of the Bank’s forecasts, this exercise also controls for the impact of any general increase in economic uncertainty on forecast errors.

Wherever possible, forecasts are compared with more ‘mature’ estimates of data, taken three years after the period in question. Where mature estimates are not yet available, the latest vintage of data is used. Official CPI inflation data do not change from first publication, but other variables can get revised as the Office for National Statistics (ONS) incorporates lagged source information and methodological improvements.^{footnote [14]} Revisions are typically hard to predict (Robinson (2016)) and policy needs to be set on the basis of real-time information. But three-year-old estimates are preferred for this evaluation because they provide a more definitive picture of the economy, and so a better basis for retrospectively evaluating the sources of inflation forecast misses. The impact of data revisions on forecast performance is explored further in Section 5.

The Covid-19 period is included in results unless otherwise stated. As GDP growth was exceptionally volatile during the Covid pandemic, the period from 2020 Q1 to 2022 Q2 is generally excluded from GDP evaluation metrics.

…and draws flexibly on other techniques, like model-based counterfactuals, to shed light on the drivers of recent errors.

An assessment of more recent errors cannot rely solely on statistical tests which require longer samples to produce meaningful results. Some conclusions will also be tentative, since more recent data outturns are more prone to revision. There is also unlikely to be a one-size-fits-all approach to investigating recent errors, with the most appropriate methods depending on the variable and horizon of interest, or the shocks at play.

Section 5 therefore considers a range of complementary approaches to analyse more recent errors. The section focuses on the period since mid-2021, which saw sharp rises in energy prices and increases in Bank Rate shortly after. The analysis includes model-based ‘counterfactuals’ to estimate the role of surprises to, or changes to the treatment of, conditioning paths such as energy prices. The section also draws on alternative models and staff analysis to discuss possible lessons from forecast errors, where these errors may have resulted from mechanisms underpinning the forecast or from mechanisms that may not be fully captured by standard forecast models. The role of other factors such as data revisions is also considered briefly.

4: Longer-term statistical properties of the Bank’s forecasts

This section analyses the statistical properties of the Bank’s forecasts over the past decade. It analyses forecast performance along three common dimensions considered in the forecast evaluation literature: accuracy, unbiasedness, and efficiency. It also benchmarks the Bank’s forecasts against a range of alternative model-based and external forecasts.

4.1: Forecast accuracy

Accuracy relates to how close forecasts have been to outturns.

A standard accuracy metric is the ‘root mean squared error’ (RMSE), which captures the size of a typical forecast error.^{footnote [15]} Table 4.A shows RMSEs for CPI inflation, GDP growth, wage growth and the unemployment rate over 2015–25 (top panel) and split into pre- and post-Covid samples (bottom panel).

Unsurprisingly, RMSEs generally rise at longer forecast horizons. This is expected as near-term forecasts incorporate available high-frequency information via statistical models and judgement, while medium-term forecasting is made harder by unanticipated shocks and wider uncertainties about key economic relationships and dynamics.

The Bank’s forecasts have become notably less accurate post-Covid, in part as economic volatility has also increased.

Comparing the pre- and post-Covid periods, the scale of errors has also increased across the board. The RMSE of one-year ahead inflation forecasts, for example, was 0.6 percentage points pre-Covid, compared to 3.7 percentage points post-Covid. Forecasts for GDP growth and wage growth have also become less accurate on average.

Larger forecast errors in the post-Covid period partly reflect heightened economic volatility, which has made forecasting more challenging. A further challenge over recent years relates to economic measurement. Smaller samples in the ONS’s Labour Force Survey have led to more volatile and uncertain labour market statistics. Meanwhile, some unusually large revisions to GDP have also contributed to larger post-Covid forecast errors, when measured against mature (three-years-old) data. This is discussed further in Section 5.

Alternative forecasts provide a useful relative performance benchmark.

To test whether the accuracy of the Bank’s forecasts could have been improved easily, it is helpful to compare them to alternative forecasts. Box A briefly describes the three considered AR, BVAR and COMPASS model benchmarks, alongside the SEF. Table 4.B presents their respective ‘relative RMSEs’, a measure of relative accuracy^{footnote [16]} where a number larger than 1 implies the Bank’s forecasts have been more accurate than the alternative. Relative forecast measures help to control for the impact of heightened economic uncertainty, which affects all forecasts produced on the basis of real-time information equally. Consistent with that, for this relative accuracy exercise the Covid period is included in the results for GDP growth.

The Bank has outperformed model-based forecasts for inflation and GDP but results for labour market variables are more mixed.

The Bank’s forecasts have generally been at least as accurate as model-based benchmarks. Relative performance is strongest for inflation and GDP at shorter horizons. For example, for Year 0 GDP growth, the three models would have had RMSEs around three times larger than the Bank’s, as Bank staff’s judgement helped to make short-term forecasts of GDP growth materially more accurate during the Covid pandemic. For CPI inflation at the Year 1 horizon, the AR and BVAR models would still have had 10%–20% higher RMSEs, but the improved COMPASS model would have been about as accurate as the Bank’s forecasts. For two- and three-years ahead forecasts, the Bank’s forecast performance is broadly similar to all the models.

The Bank’s wage forecasts have also improved on model benchmarks at some horizons, although less markedly over the near-term. Unemployment forecasts have underperformed relative to the AR model benchmark at horizons up to Year 1.^{footnote [17]} This partly reflects successive changes to the end date of UK’s Coronavirus Job Retention Scheme on which post-Covid forecasts were conditioned, but errors have also been relatively large in other periods.

The Bank has also been at least as accurate as external forecasters.

Like the Bank, external forecasters incorporate available high-frequency information into their forecasts, providing a like-for-like benchmark. The Bank’s forecast accuracy broadly matches the SEF, with relative RMSEs close to 1.^{footnote [18]} That means the Bank’s forecasts have performed just as well as an average across a much larger set of independent forecasters (which in theory should benefit from the ‘wisdom of the crowds’). As noted in the Bernanke Review, ‘historically, individual forecasters have had great difficulty systematically beating the consensus’ (Bernanke (2024)).

These results suggest there is no one readily available alternative forecast that systematically outperforms the Bank’s forecasts, but this does not rule out potential room for improvement.

These results for the past decade highlight the benefit of incorporating insights from multiple models, data sources and expert judgement in the Bank’s forecast process. But they do not mean that the Bank’s forecasting cannot be improved, or that accuracy could not have been higher during certain periods, for example if more specialised models were available to draw on or by applying different judgements.

4.2: Unbiasedness

A forecast is said to be unbiased if it is as likely to under- or over-predict subsequent data outturns.

Forecast unbiasedness can be tested by regressing forecast errors on a single constant term (Mincer and Zarnowitz (1969)), which estimates the average (non-absolute) error. A positive coefficient in this regression means that outturns have been higher than forecasts on average, and vice versa. Table 4.C presents unbiasedness results estimated in this way for each variable over 2015–25 and split into pre- and post-Covid samples.

Inflation and wage growth have been underpredicted on average in the past decade…

The Bank’s forecasts of inflation and wage growth display some signs of bias. For example, two-years ahead inflation forecasts have underpredicted outturns by 1.6 percentage points on average over 2015–25. However, this result is driven entirely by the post-Covid period, which saw unusually large errors in these variables, as discussed further in Section 5. Pre-Covid, inflation forecasts were largely unbiased, although wage growth had been slightly overpredicted. Persistent downside surprises to wage forecasts were a well-documented feature of the 2010’s (IEO (2015)), alongside a persistent overestimation of productivity trends.

…though observed biases can partly reflect the distribution of underlying shocks over the period in question.

Because shocks can hit after the forecast was made, care is needed when interpreting unbiasedness results. If the realised distribution of such shocks happens to be skewed over a period – for example, due to some exceptionally large shocks – resulting forecast errors can appear biased in hindsight. That factor is consistent with forecasts from the full range of alternative models exhibiting a similar pattern of biasedness for wages and inflation in the post-2020 period.^{footnote [19]} The analysis in Section 5 points to a sizeable role for energy price and other shocks driving the inflation errors in 2022.

The unemployment rate has been overpredicted on average, continuing a pattern observed in the IEO’s forecast evaluation, and while GDP growth forecasts appear generally unbiased, they are sensitive to data revisions.

The IEO forecast evaluation in 2015 also identified a persistent downward bias in the Bank’s unemployment forecasts (IEO (2015)), which has continued into 2015–25. In this period, the unemployment rate has been overpredicted on average at every forecasting horizon, and especially for one-year ahead forecasts.

The Bank’s GDP growth forecasts meanwhile have generally remained unbiased. While Year 0 GDP growth forecasts at first sight appear statistically biased, this has been largely the result of some exceptionally large revisions to GDP data since Covid (Section 5).

4.3: Efficiency

A forecast is said to be efficient if it makes good use of available information.

Efficiency tests help determine whether forecasts made full use of available information (Nordhaus (1987)). Some ‘weak’ efficiency tests do this by checking whether forecasts could have been made more accurate by rescaling, or if forecast revisions could have been predicted with past information of the same variable. Table 4.D shows the results for a range of these tests, with figures in bold those that show statistically significant evidence of weak inefficiency.

Tests suggest the Bank’s forecasts for inflation are ‘weakly efficient’, but those for GDP, wage growth and the unemployment rate are not.

There is limited evidence to suggest that inflation forecasts are not ‘weakly efficient’, in other words these appear to utilise past information on inflation well. But each of the Bank’s forecasts for GDP growth, wage growth, and unemployment fails a range of these tests, suggesting scope for better information processing in those variables, in particular for unemployment.

Some key economic relationships may have been underestimated.

A second type of ‘strong’ efficiency tests considers whether forecasts for a certain variable could have made better use of information contained in forecasts of other variables, rather than own-variable information. A Blanchard-Leigh methodology outlined in Kanngiesser and Willems (2024) can be used specifically to test whether certain economic relationships are mis-calibrated. For example, if a correctly forecasted movement in one variable, eg a rise in wage growth, was found to be systematically associated with errors in another, eg higher-than-forecasted inflation, then forecasts would be said to underestimate the link between wage growth and inflation.

As discussed in greater detail in Kanngiesser and Willems (2024), results from an application of this Blanchard-Leigh methodology to key economic relationships within Bank forecasts from 2011 to 2024, suggest that the Bank may have:

• underestimated the strength of pass-through from (higher) wage growth to (higher) CPI inflation at longer forecast horizons (two to three years). This could be a factor in explaining persistent upside surprises in inflation over recent years, where both wage growth and inflation have been unusually elevated; and
• underestimated the strength of the Phillips curve relationship between (lower) unemployment and (higher) inflation. This could likewise point to a greater role for labour market tightness in explaining the strength and persistence of inflation in recent years, as unemployment reached historical lows in the second half of 2022, before starting to rise gently.

But, perhaps partly offsetting that, the analysis suggests the Bank may also have:

• underestimated the strength and speed of monetary transmission linking (higher) Bank Rate to (lower) inflation, which all else equal might have weighed on inflation by more than anticipated as Bank Rate climbed rapidly to a peak of 5.25% in August 2023, before stabilising at that level and beginning a more gradual descent from mid-2024.

These tests can be highly sensitive to the choices for sample period and outlier exclusions, however. They also rely on simple bivariate relationships, which can inadvertently pick up the effects of third (omitted) variables. Some inefficiencies are also evident in the alternative benchmark models.^{footnote [20]} As such, these results should be interpreted as more indicative than definitive, and flagging potential areas for further investigation, some of which are picked up again in Section 5.

5: Key findings from recent forecast errors

This section unpacks some of the key findings and lessons learned from evaluating the Bank’s forecast performance over the past five years. It starts by outlining some of the key economic drivers and forecast errors over this period. It then presents seven key findings from evaluating the Bank’s forecasts since the August 2021 Monetary Policy Report. This period was chosen as it was a particularly challenging time to produce accurate forecasts given the series of large shocks to which the economy was exposed.

5.1: Economic context

A series of significant shocks have affected the UK economy since mid-2021 that made forecasting particularly challenging.

The UK economy has faced a number of historically large shocks over recent years (Bailey (2023), Pill (2024), Mann (2025), Taylor (2025)). For example, the pandemic simultaneously hit demand and supply, with sectoral shutdowns and global supply‑chain disruptions. While the most acute effects of the Covid pandemic restrictions in 2020 are not in scope for this section of the Report, the after-effects of that disruption are. They include the supply-chain bottlenecks that pushed up goods prices and fed into inflation pressures in the UK. Russia’s invasion of Ukraine led to a further, historically large rise in global energy and food prices, significantly adding to UK inflation pressures and dampening real activity. In addition, the UK’s departure from the EU introduced a new set of trading relationships, with implications for trade flows, productivity and GDP (Section 3 of the February 2023 Monetary Policy Report).

The various shocks are reflected by an application of the model described in Bernanke and Blanchard (2025) to the UK economy (Haskel, Martin and Brandt (2025)), which estimates the contribution of different economic drivers to CPI inflation (Chart 5.1).

Chart 5.1: Several significant shocks have pushed inflation away from the 2% target in recent years

Estimated decomposition of contributions to UK CPI inflation since 2020 ^(a)

• Sources: Bank of England, Federal Reserve Bank of New York, ONS and Bank calculations (based on Haskel, Martin and Brandt (2025)).
• (a) Decomposition is based on dynamic simulations, initiated with data up to 2019 Q4, after which only data on the exogenous variables (shocks) is shown to the model and predictions for the endogenous variables (wages, prices and inflation expectations) are used iteratively. The estimated contributions capture the dynamic effect of shocks on inflation, including both their direct effects and second-round effects via inflation expectations and wages. ‘Initial conditions’ shows the estimated level of inflation had the economy remained at conditions prevailing in 2019.

That sequence of shocks and structural changes has made forecasting over this period particularly challenging. And, as noted by Bernanke (2024), these have meant that central banks across the world have made larger forecast errors than in previous years, with the experience of the Bank’s forecasts comparable to other major central banks. Chart 5.2a and chart 5.2b show the realised data outturns against successive Bank forecasts for CPI inflation, GDP growth, wage growth and the unemployment rate.

Lags in the transmission of monetary policy mean that unforeseen circumstances – such as the large rise in energy prices described above – can push inflation temporarily away from target, even if inflation is then expected to return to target in the medium term. And in the presence of supply shocks, policymakers can also face a trade-off between the speed with which inflation returns to target, and the degree of economic volatility required to achieve that (Broadbent (2024)). Together, this means that some deviations in inflation from the 2% target are unavoidable.

Some caution should be applied in drawing conclusions from this set of forecasts.

Data for this period remain subject to revision, so the passage of time may reveal new information about forecast performance. Further, for longer forecast horizons there are a relatively small number of observations. The latest three-year forecast that can be analysed is from August 2022 and the latest two-year forecast is from August 2023. Future Forecast Evaluation Reports will continue to learn from forecasts in this period, as more data become available and existing data are revised.

Forecasts are one of a number of important inputs to support the MPC’s policy discussions. While the forecast evaluation outlined here reveals some areas to improve performance, the results should be taken in context. The conditional nature of the Bank’s forecasts means they are not designed to maximise forecast accuracy. The weight that individual MPC members place on the central forecast, versus the risks around it (including in the form of scenarios), will vary over time. Forecast errors can only be assessed with the benefit of hindsight, when more information is available. These findings should not, therefore, be taken to imply that policymakers set monetary policy incorrectly, given the information that was available to them at the time.

Chart 5.2a: On average, data outturns have been higher than forecast for inflation and GDP growth…

Forecasts and outturns for CPI inflation and GDP growth ^(a)

• Sources: ONS and Bank calculations.
• (a) Latest datapoints shown are for 2025 Q2 and show data as at the time of the 2025 Q2 ONS Quarterly National Accounts. Lines in light orange show the successive Bank forecasts for each variable, including those published alongside the August 2021 Monetary Policy Report up to the August 2025 Monetary Policy Report. Forecasts shown are for year-on-year growth in GDP and CPI inflation.

Chart 5.2b: …and outturns exceeded expectations, on average, for wage growth, but were lower than forecast for the unemployment rate

Forecasts and outturns for wage growth and unemployment ^(a)

• Sources: ONS and Bank calculations.
• (a) Refer to Chart 5.2a. Forecasts shown are for year-on-year growth in whole-economy total average weekly earnings and for the level of the unemployment rate.

5.2: Accuracy of forecasts published from August 2021 to August 2025

From mid-2021 onwards, there were repeated upside surprises in inflation, with the annual rate peaking in October 2022. The subsequent disinflation process, which started in late-2022 onwards, has developed a little more in line with the Bank’s one-year ahead forecasts (orange line in left panel of Chart 5.3a). But inflation data have more systematically continued to overshoot two- and three-years ahead projections (purple and gold lines in left panel of Chart 5.3a). In part, this is likely to reflect strength in wage growth, which has been consistently higher than forecasts through that period (left panel of Chart 5.3b).

GDP growth has, on average, also been stronger than projected in forecasts made since 2022 (orange line in right panel of Chart 5.3a). Activity did not contract as much as expected in 2023 and 2024 following the rise in energy prices and tightening in monetary policy. The unemployment rate was also lower than projected, on average, over this period (right panel of Chart 5.3b). Thereafter, GDP growth picked up by more than successive forecasts predicted. Since 2024 Q4 the unemployment rate has been higher than expected, however, as the labour market has weakened more than expected (aqua line in right panel of Chart 5.3b).

Chart 5.3a: Forecast errors have generally fallen since 2021, but have remained positive for CPI inflation and GDP growth…

Time series of forecast errors by forecast horizon ^(a)

• Sources: ONS and Bank calculations.
• (a) Differences between data outturns and projections in Monetary Policy Reports since August 2021. Data outturns are recorded after three years (or the latest available data vintage) to incorporate subsequent revisions to the initially published data. Forecast errors shown are shown for year-on-year growth in GDP and CPI inflation.

Chart 5.3b: …and have also remained positive for wage growth

Time series of forecast errors by forecast horizon ^(a)

• Sources: ONS and Bank calculations.
• (a) Refer to Chart 5.3a. Forecast errors shown are shown for whole-economy total average weekly earnings and for the level of the unemployment rate.

5.3: Key findings

Key finding 1: Inflation rose sharply over 2021–22, with outturns notably higher than forecast. One-year ahead inflation forecast errors peaked in mid-2022, with the increase in energy prices the single most important factor.

Inflation turned out materially higher than projected at the time of the August 2021 Monetary Policy Report.

The August 2021 Monetary Policy Report forecasts contained a period of above-target inflation, with a projected peak of 4.0% in 2021 Q4 followed by a gradual return towards the 2% target by end-2023. Relative to that projection, inflation was materially higher over most of the three-year forecast period. The rate of quarterly annual inflation peaked at 10.8% in 2022 Q4 and only returned close to the August 2021 Monetary Policy Report forecast in mid-2024 (Chart 5.4).

Unanticipated changes in energy prices can account for around half of the inflation forecast error in the August 2021 forecast…

The peak inflation forecast error from the August 2021 Monetary Policy Report was 8.2 percentage points in 2022 Q4. In large part that reflects the sharp rise in energy prices following the Russian invasion of Ukraine in early 2022, with gas prices rising over 600% in the year to January 2022 and fluctuating widely after that (Chart 5.5).

To help estimate how much of a role news in energy prices played in the Bank’s inflation forecast errors, Bank staff have performed a ‘counterfactual’ exercise. In this exercise, the DSGE model described in Albuquerque et al (2025) is used to update the August 2021 Monetary Policy Report forecast to take account of the realised data outturns for energy prices. Those estimates suggest that taking account of the direct and indirect (through supply chains and wider demand channels) effects of energy can account for around 4 percentage points, or half, of the Bank’s peak inflation forecast error (Chart 5.4).

Chart 5.4: News in energy prices can account for around half of the peak forecast error in inflation

CPI inflation, August 2021 Monetary Policy Report forecast and counterfactual ^(a)

• Sources: ONS and Bank calculations.
• (a) Year-on-year CPI inflation. The August 2021 Monetary Policy Report forecast incorporates backdata revisions up to 2021 Q1. The August 2021 Monetary Policy Report forecast including energy price outturns incorporates the modelled impact of the actual path for the direct energy price contribution to CPI inflation using the Bank’s main DSGE model for forecasting (Albuquerque et al (2025)).

…as wholesale energy price conditioning assumptions initially underestimated the increase in energy prices.

Wholesale gas prices rose by over 600% to record levels over 2021 and 2022. The gas conditioning path that fed into the Bank’s August 2021 forecasts was therefore 77% lower than the realised gas spot price in 2022 Q3 (gold line in Chart 5.5). Given the direct effect of the spike in energy prices was not reflected in forecasts from this period, this contributed to forecast errors.

Once energy prices had peaked, there remained considerable uncertainty over their subsequent evolution.

The Bank’s August 2022 forecast captured the peak in wholesale gas prices, at historically unprecedented levels (Chart 5.5). Given the nature of the uncertainty over the outlook for gas prices, multiple scenarios were used to support the MPC’s policy discussions at that time (Box A of the August 2022 Monetary Policy Report). In particular, the central forecast then reflected the assumption that energy prices would remain persistently elevated (as set out in Box 5 in the August 2019 Inflation Report and shown by the orange line in Chart 5.5). However, given the uncertainty around that assumption, the MPC also explored an alternative scenario in which energy prices fell back in line with financial market energy futures curves.

Based on that alternative approach, from November 2022 onwards, the conditioning assumption for the central forecast was changed back to one based on financial market prices (purple line in Chart 5.5). This change resolved what would otherwise have been an inconsistency in the forecast, as government policy measures capped the retail price of energy based on the futures path. Wholesale energy prices subsequently fell back even more quickly than that (aqua line in Chart 5.5).

Chart 5.5: Wholesale gas prices rose by over 600% to record levels over 2021 and 2022

Outturn and conditioning assumption used in August 2021 Monetary Policy Report forecast for wholesale gas prices ^(a)

• Sources: Bloomberg Finance L.P. and Bank calculations.
• (a) Spot gas prices are quarterly averages of Bloomberg UK NBP Natural Gas Forward Day prices. Other lines refer to respective futures curves using one-month forward prices based on asset price conditioning assumption window in each forecast noted.

Key finding 2: Other global factors can account for most of the remaining peak inflation forecast error.

Higher-than-expected global export prices contributed nearly 1 percentage point to the peak inflation forecast error under standard assumptions…

Stronger-than-expected global export prices also contributed to the peak inflation forecast error. Disruption to global supply chains as economies reopened following Covid restrictions and strong demand for consumer goods, along with the indirect effects of rising energy prices, led to a sharp rise in export prices among the UK’s trading partners. Those in turn pushed up on UK import prices and CPI inflation. This was particularly evident for tradable goods prices.

Bank staff’s forecasts for global export price growth were repeatedly revised upwards over the course of 2021–22 (bottom panel of Chart 5.6). Higher-than-expected global export prices contributed almost 1 percentage point to the peak August 2021 inflation forecast error, according to the Bank’s standard assumptions on the pass-through from global export prices to UK CPI inflation (green bar in the top panel of Chart 5.6).

…but the scale and nature of the increase in global export prices may mean they contributed a further 2 percentage points.

Bank staff estimate that global export prices contributed more to higher inflation over 2021–23 than the Bank’s standard forecast assumptions would suggest (gold bar in the top panel of Chart 5.6 and Broadbent (2024)). Usually, pass-through of global export prices to inflation would be expected to be incomplete with protracted lags, due to complex global supply chains.^{footnote [21]} There is evidence, however, that the extent and speed of pass-through were both much greater than expected over this period (Box D of the November 2024 Monetary Policy Report). This may reflect the nature of the shocks, which were large, facing almost all companies in some form, and associated with shortages of some products. It also may have reflected domestic conditions, as strength in demand alongside supply constraints may have afforded companies greater than usual power to pass on cost increases to their prices.

Global export price growth slowed sharply from late-2022, although Bank staff forecasts projected a prolonged period of global export deflation that proved briefer and less pronounced. This further contributed to medium-term UK inflation forecast errors, contributing on average 0.5 percentage points to two-years ahead forecast errors between August 2021 and May 2023. In seeking to learn the lessons from this period, Bank staff have developed a new forecasting model for global export prices.

Chart 5.6: Global factors contribute most of the remaining peak inflation forecast error, including from persistent strength in global export prices

Contributions to the peak inflation forecast error (top panel) ^(a) and successive forecasts for global export prices (bottom panel) ^(b)

• Sources: Bank of England, BEA, Bloomberg Finance L.P., Eurostat, LSEG Workspace, OBR, ONS, regional central banks, regional statistical agencies and Bank calculations.
• (a) Chart breaks down contribution of different factors to the peak UK inflation forecast error in 2022 Q4 (compared to the forecast made at the time of the August 2021 Monetary Policy Report). ‘Conditioning assumptions’ bar is calculated by taking news to the conditioning paths for yield curve, exchange rate, energy prices, and fiscal policy – the difference between the realised path and the August 2021 Monetary Policy Report forecast assumption – and running it through the set of models used to produce the August 2021 forecast. Similarly, the ‘Standard impact of news to global export prices’ bar is calculated by running news to global export prices excluding fuel through the models used to produce the August 2021 forecast. The ‘Additional impact of global export prices in this episode’ bar is a Bank staff estimate based on examining the breakdown of CPI inflation over the period, particularly movements in its goods component.
• (b) Global export prices (excluding fuel) is a series produced by the Bank of England, which aggregates price data for 31 individual country series plus the euro area. The list excludes any large oil and gas exporters (except for the US). Where available, it relies on seasonally adjusted implicit price deflators for exports of goods and services for the particular country or region, as these allow for the changing composition of export baskets over time. However, for some countries where those are not available, it uses export price indices. The underlying price series are provided by individual country central banks, statistical agencies or Eurostat. These 32 series are then aggregated into the total world export price (excluding direct oil) series based on their relative share in UK imports. A forecast for world export prices (excluding direct oil) is alongside each Monetary Policy Report. The chart shows the percentage change on a year earlier.

Other factors, including non-linearities in the face of large deviations from the inflation target, may have also played a part in the peak inflation forecast error.

Firms and households may behave differently in high inflation environments, affecting their expectations for future price rises and their behaviour when setting prices or bargaining for wages (Box C of the November 2025 Monetary Policy Report). The sharp rise in energy and other global prices may have triggered changes in domestic price setting that contributed to the peak rise in inflation, and subsequent slower-than-expected decline in inflation (key finding 4). The core models used to construct the central forecast are linear and so do not capture non-linearities like these. The Bank has developed models that can capture such non-linear effects which can be used to inform judgements layered on to the central forecast or to produce scenarios (for example Buckmann et al (2025)).

Key finding 3: In response to inflationary pressures, Bank Rate increased by more than the market-implied paths upon which forecasts were conditioned. There is also evidence that monetary policy transmission may have been faster than previously assumed.

The financial market conditioning path for Bank Rate from August 2021 substantially underpredicted the subsequent rise in Bank Rate.

The MPC increased Bank Rate by 5.15 percentage points between December 2021 and September 2023. That was a significantly higher path for Bank Rate than the 0.45 percentage point rise implied by financial markets at the time of the August 2021 Monetary Policy Report (Chart 5.7). This tightening in interest rates played a role in mitigating inflationary pressures and weighed on activity. The market rate conditioning path shifted upwards repeatedly, before peaking in the August 2023 Monetary Policy Report. Bank Rate has followed a path broadly in line with market prices since, declining gradually from its peak of 5.25%.

Chart 5.7: Bank Rate increased to a peak of 5.25%

Market-based conditioning paths for Bank Rate and outturn ^(a)

• Sources: Bloomberg Finance L.P. and Bank calculations.
• (a) The chart shows the evolution of Bank Rate and the market implied path from the August 2021 Monetary Policy Report to the November 2025 Monetary Policy Report. Bank Rate paths over forecast periods reflect overnight index swap-implied expectations at the time of each Monetary Policy Report and represent the conditioning paths for each specific forecast.

Shorter transmission lags and a smaller peak impact of the rise in Bank Rate may have biased downwards the medium-term growth outlook as interest rates increased.

As Bank Rate and the market conditioning path increased, that was expected to gradually weigh on economic activity, dampening growth over the full three-year forecast horizon. However, the average bias in GDP growth forecasts over this period prompted an evaluation of those assumptions (key finding 5). This evaluation found, in this tightening cycle, the monetary transmission mechanism appears to have acted more quickly and to a lesser degree than previously, with the effect of biasing downwards medium-term growth forecasts.

An evaluation of the transmission of the tightening in monetary policy has provided evidence that there may have been shorter lags in this episode than previously.

Several factors point to this tightening cycle having a more front-loaded impact on activity and inflation than historical experience would have implied. One example is the significant build-up of liquid deposits during the pandemic, which may have increased the sensitivity of household spending to changes in interest rates relative to historical experience (Box C of the August 2024 Monetary Policy Report). Additional evidence on the relationship between Bank Rate and inflation in the Bank’s forecasts is presented in Kanngiesser and Willems (2024). As a result, the transmission of policy rates assumed in the Bank’s optimal policy toolkit used to support the MPC’s policy strategy deliberations has been updated to take account of the empirical evidence from this period (Chart 5.8 and Alati et al (2025)).

Chart 5.8: The transmission of changes in Bank Rate to GDP has been updated in the Bank’s toolkit for policy strategy deliberations

The response of the level of GDP to a persistent 100 basis points cut in Bank Rate in the Bank’s toolkit ^(a)

• Source: Bank calculations.
• (a) Chart shows the impact assumed in the Bank’s forecast of a Bank Rate cut. It shows the impact on the level of GDP in per cent relative to the baseline for a persistent 100 basis points cut in Bank Rate over three years, which then unwinds back to the baseline over the subsequent three years. The estimates show the GDP responses from two versions of the Bank’s main model for policy analysis, with the previous monetary transmission mechanism used at the start of the tightening cycle in 2021, compared to the latest estimates (Alati et al (2025)).

Key finding 4: Since mid-2021, medium-term inflation forecasts have been too low on average, as inflation has exhibited unexpected persistence.

Since mid-2021, both wage and inflation forecasts have not been statistically unbiased.

Since the Bank’s August 2021 projections, both wage growth and inflation data have exceeded their forecast on average across most horizons (Chart 5.9). For two-years ahead forecasts, inflation has on average been nearly 2 percentage points higher than forecasts and wage growth nearly 3 percentage points higher. Considering the role of wages as a driver of domestic inflation, the emergence of these biases at the same time could be indicative of a role for wage-price interactions in accounting for persistent inflationary pressures.

Chart 5.9: Wage growth and inflation forecasts show signs of bias since 2021

Estimated biasedness coefficients for wage growth (left panel) and CPI inflation (right panel) ^(a)

• Source: Bank calculations.
• (a) Bias is calculated as outturn minus forecast. A positive number indicates that outturns have been higher than forecast. Sample of forecasts included is from the August 2021 to August 2025 Monetary Policy Reports. Shaded swathe denotes 95% confidence interval. Quarters refers to the number of quarters ahead for forecasts. The number of observations in the estimation sample decreases as forecast horizon increases. For example, there are 13 observations of one-year ahead forecast errors versus five three-years ahead forecast errors underpinning these estimates.

Energy prices and other conditioning assumptions do not account for much of the average two-years ahead inflation forecast error.

Medium-term inflation forecast errors, after the peak in late-2022, cannot be explained by changes in energy prices or other forecast conditioning assumptions (Chart 5.10). In fact, wholesale energy prices fell more than expected relative to forecasts made between November 2021 and August 2022 (aqua bars in Chart 5.10). All else equal that would have contributed to inflation being lower than expected two years later.

Chart 5.10: News to conditioning paths – including for energy prices – explain little of the Bank’s medium-term inflation forecast errors since mid-2021

Contribution of conditioning path news to year two CPI inflation forecast errors ^(a)

• Sources: Bloomberg Finance L.P., LSEG Workspace, OBR, ONS and Bank calculations.
• (a) The contributions of the conditioning paths to the inflation forecast errors are calculated by taking the news in each conditioning path – the difference between the realised path and the assumption made at the time of the forecast – and running it through the set of models in use at the time that forecast was produced.

Rather, second-round effects from the sharp rise in inflation are likely to have been a significant driver of unexpected persistence…

The rise in inflation in 2021 and 2022 led to ‘second-round effects’, whereby higher inflation contributed to higher inflation expectations, which fed back to wage growth and added to inflation pressure. Judging the size and speed of these second-round effects was challenging (Broadbent (2022)). The size of the external shock to inflation, and the tightness in the labour market over 2021 and 2022, meant that estimates from historical episodes were unlikely to be a good guide to behaviour this time. Given this uncertainty, Bank staff have utilised a range of approaches to estimate second-round effects and the MPC has applied judgements to the Bank’s inflation forecasts to account for them. These judgements have acted to reduce the size of forecast errors on average over this period.

One approach that Bank staff have drawn on is a UK application of the Bernanke-Blanchard model (Bernanke and Blanchard (2025) and Haskel, Martin and Brandt (2025)). This model^{footnote [22]} suggests second-round effects added nearly 1 percentage point to inflation in late-2023, before beginning to unwind gradually. Estimates from the Bernanke-Blanchard model have been used as a cross-check on MPC judgements for the persistence of inflation (Chart 5.11).^{footnote [23]} The estimated impact of second-round effects from the Bernanke-Blanchard model would have also helped to better explain wage growth over 2024 (Haskel (2024)).

There remains uncertainty around the strength of some of these mechanisms and how quickly they might fade. Inflation persistence scenarios were developed to help assess risks around the central projections in the May 2025 Monetary Policy Report and November 2025 Monetary Policy Report.

Chart 5.11: The Bernanke-Blanchard model estimates of second-round effects have informed the MPC’s judgements on inflation persistence

Estimated second-round effects from rise in inflation since 2021 in the Bernanke-Blanchard model and MPC judgements for inflation persistence in the forecast ^(a)

• Sources: Bank of England, Federal Reserve Bank of New York, ONS and Bank calculations (based on Haskel, Martin and Brandt (2025)).
• (a) The second round effects estimated in the Bernanke-Blanchard model are constructed by contrasting dynamic simulations for inflation since 2019 Q4 with and without price-to-wage feedback. Estimates are based on data for exogenous variables in the model up to 2025 Q3, and conditioning paths for exogenous variables consistent with the November 2025 central projection thereafter. MPC inflation persistence judgements are consistent with the November 2025 central projection.

…and it is also possible that the persistent strength in wage and price growth may reflect structural changes.

In addition to the lagged effects of previous shocks and labour market tightness, a more persistent or even structural change in wage-setting behaviour may have been at play, which could have kept wage growth elevated for longer even as the impact of recent shocks fade. Workers may have become more resistant to reductions in real wages for example (Pill (2025)), or there may have been long-lasting changes in the formation and propagation of inflation expectations following the inflationary shocks in recent years (Section 3.3 of the November 2025 Monetary Policy Report). There may also have been a structural change in firms’ pricing behaviour, such that inflation for a given rate of wage growth is stronger than in the past (Breeden (2025)).

Given the challenges with explaining the persistent strength of wage growth and inflation over recent years, and as recommended by Bernanke (2024), better understanding labour market dynamics and wage-price interactions remain key priorities for the Bank.

Key finding 5: On average, since mid-2021, GDP growth has been stronger than the Bank’s forecasts, partly reflecting news in conditioning paths.

GDP growth forecasts since August 2021 have been around 0.5 percentage points stronger than expected at each of the one-, two- and three-year horizons. In recent years, positive GDP surprises have also been associated with less spare capacity, in turn as potential supply is judged to have been slightly weaker than expected (key finding 6).

The unanticipated rise in energy prices, along with higher interest rates, initially contributed to GDP growth being weaker than expected.

Higher energy prices act like a supply shock, weakening GDP growth and pushing up inflation, creating a trade-off for policymakers. Relative to the August 2021 forecast, before wholesale prices started to rise rapidly, the latest data show GDP growth in 2022 and 2023 was notably weaker than forecast. The rise in Bank Rate also contributed to weaker-than-expected GDP growth in this period, as discussed in key finding 3.

However, the prolonged downturn in activity that was forecast in late-2022 did not materialise, as wholesale energy prices fell back from their peak.

Forecasts in August and November 2022 forecasted a prolonged fall in GDP, as wholesale energy prices reached a peak (Chart 5.5). Subsequent data showed GDP was more resilient. This partly reflected changes in the conditioning assumption for wholesale gas prices and partly reflected that realised wholesale prices fell more significantly than forecasts had assumed, which supported economic activity more than initially thought. At the same time, the Government’s Energy Price Guarantee and other support measures helped to limit the fall in household incomes.

The impact of a given rise in energy prices on activity is also judged to have been smaller than Bank staff initially estimated.

Bank staff’s latest assessment is that a rise in energy prices has a smaller impact on activity than embedded in earlier Bank staff models. This partly reflects households’ and firms’ greater willingness and ability to substitute away from energy consumption over recent years (Breeden (2025)). This finding, in part, motivated Bank staff’s improved modelling of energy impacts within its central forecasting model COMPASS (Albuquerque et al (2025)).

Overall, conditioning assumptions can explain most medium-term GDP errors since mid-2021, in particular from energy prices, Bank Rate expectations and fiscal policy.

Chart 5.12 shows the contribution from different conditioning assumptions to two-years ahead GDP forecast errors. In addition to energy prices, other conditioning assumptions can explain a significant share of medium-term forecast errors since mid-2021. As noted in key finding 3, market paths for Bank Rate were repeatedly higher than expected, bearing down on GDP forecasts (purple bars in Chart 5.12). There has also been a positive contribution from fiscal policy (orange bars in Chart 5.12). This reflects outturns for public sector borrowing in the medium term generally being higher than expected, consistent with the findings of the OBR’s forecast evaluation (OBR (2025)).

Chart 5.12: News to conditioning paths account for most of the MPC’s medium-term GDP forecast errors

Estimated contribution of conditioning paths to year two GDP level forecast errors ^(a)

• Sources: Bloomberg Finance L.P., LSEG Workspace, OBR, ONS and Bank calculations.
• (a) The contributions of the conditioning paths to the GDP forecast errors are calculated by taking the news in each conditioning path – the difference between the realised path and the assumption made at the time of the forecast – and running it through the set of models in use at the time that forecast was produced.

Key finding 6: Potential supply is judged to have been slightly weaker since 2021, and alongside stronger GDP growth has meant the economy has operated with less spare capacity than initially expected.

In assessing the outlook for activity and inflation, forecasts need to make judgements for the potential supply capacity of the economy. This underpins the rate the economy can grow – and is one factor affecting the level of interest rates – that is consistent with meeting the 2% inflation target in the medium term.

The level of potential supply capacity cannot be observed directly and is likely to fluctuate over time. Instead, signals about the balance between demand and supply potential can be inferred from other sources, such as changes in the unemployment rate, and the rates of wage and price growth. The presence of measurement uncertainties and the varying timeliness of data availability, coupled with lags in the inflation process, mean that estimation of supply potential is inherently uncertain and new evidence can lead to revised assessments over time (Orphanides (2001)).

Overall, potential supply is judged to be slightly weaker in the latest estimates compared to those made in late-2021 (Chart 5.13). A series of significant shocks have weighed on potential supply, such as the after-effects of the Covid pandemic, the energy price shock and the change in trading relationships with the EU (Bailey (2023)).

Chart 5.13: Estimates of potential supply are slightly weaker than in 2021

Estimates of potential supply average annual growth (left panel) and level of potential supply (right panel) made in 2021 and 2025 ^(a)

• Source: Bank calculations.
• (a) ‘November 2021’ refers to the estimate published in Section 3 of the November 2021 Monetary Policy Report, which was the first supply stocktake conducted in the period considered in this section since August 2021. The November 2021 estimate extended to 2024 Q4. ‘February 2025’ refers to the estimate published in Box E of the February 2025 Monetary Policy Report, which is the latest supply stocktake in the period this Report covers up to August 2025.

Stronger-than-expected GDP and slightly lower potential supply have meant there is judged to have been less spare capacity than initially thought.

The signal that monetary policy makers take from GDP forecast errors tends to depend on their implications for spare capacity in the economy. As spare capacity reduces, inflationary pressures tend to rise, and vice versa. Spare capacity is unobservable and so estimates of this are inherently uncertain, but they have been revised such that MPC has judged that there has been less spare capacity in the economy (left panel of Chart 5.14).

One reason is that Bank staff now estimate higher energy prices weighed less on activity and spare capacity than initially thought. The right panel of Chart 5.14 shows the successive estimates for how the rise in energy prices since mid-2021 would affect the output gap, with the latest profile sitting near the top of the range. Looser-than-expected fiscal policy has also contributed to there being less spare capacity. Judgements on spare capacity have been taken looking across developments in activity and inflation. Inflation persistence and stronger-than-expected GDP growth since 2021 are both consistent with less spare capacity than had been initially judged.

Chart 5.14: It is judged that the economy has operated with less spare capacity, partly reflecting judgements on the impact of higher energy prices

Range of output gap estimates since August 2020 (left panel) ^(a) and the impact of energy prices on output gap estimates under different assumptions (right panel) ^(b)

• Source: Bank calculations.
• (a) The left panel shows successive Bank staff estimates of the output gap on a calendar-year basis from the Projections Databank published in the November 2025 Monetary Policy Report. The latest estimate is for August 2025.
• (b) The right panel shows the estimated impact of energy prices on the level of the output gap from 2021 to 2025 using different model relationships, and a comparison between realised and conditional energy price paths. The August 2025 energy paths use the realised path of energy prices at the time of the August 2025 Monetary Policy Report, whereas the August 2022 energy path uses the energy price conditioning path in the August 2022 Monetary Policy Report. The estimates labelled ‘previous treatment’ refer to the 2021 model relationships. Box B of the November 2022 Monetary Policy Report contains further discussion of the treatment of energy price shocks.

Key finding 7: Data revisions have reduced the performance of the Bank’s near-term forecasts for GDP growth and wage growth when evaluated using later vintages of data.

Challenges with data measurement have been particularly acute in the post-pandemic period, with the ONS taking steps to improve data quality.

Uncertainty over the starting point of the economy has been exacerbated in the post-pandemic period by challenges with data measurement, including challenges with Labour Force Survey sample sizes, which the ONS is taking steps to address (Box D of the May 2024 Monetary Policy Report and Box B of the November 2023 Monetary Policy Report). According to the ONS (2025) revisions to GDP over the pandemic period have been large by historical standards, albeit proportional to the scale of economic volatility over that period.

Initial GDP and wage growth estimates have tended to be revised upwards in subsequent data releases since 2021.

Some series are subject to revision after their initial release, and the performance of the Bank’s forecasts can change depending on the ‘vintage’ of data the forecasts are evaluated against. This is particularly relevant for the GDP and wage data, where evaluating forecast performance against later data vintages reduces near-term forecast accuracy (Table 5.A). Revisions can lead to forecast errors in two ways. Directly, by changing data for past quarters which Bank staff take as given and which directly affect annual growth rates at the Year 0 forecast horizon. But indirectly as well, as they may mean Bank staff did not have access in real time to the most accurate steer for those economic variables.

The steer taken for the economic outlook from data revisions depends on their nature.

The signal to take from these revisions depends on the nature of them and the economic circumstances at the time. Given known lags in some of the key economic relationships, revisions can provide a signal about current and future inflationary pressures in the economy. As inflation data are not revised, however, revisions to GDP and wage data for periods that are beyond those assumed lags in pass-through are typically considered to contain little signal for inflationary pressures.

Bank staff will continue to monitor and assess data revisions. Determining whether the observed pattern of revisions is simply due to the identified data‑quality issues after the pandemic, or whether it reflects something more systematic will be a key consideration. In the case of the latter, approaches such as the Bank’s former ‘backcasting’ method, which was used previously to attempt to predict revisions to early estimates of GDP (Cunningham et al (2007)), may prove useful.

6: Looking ahead

Several lessons emerge from the forecast evaluation in this Report.

This Report finds the Bank’s forecasts have been at least as accurate as the average of external forecasters or a range of model-based alternatives over the past decade. That is consistent with findings from the Bernanke Review, which found similar forecast performance between the Bank and external forecasters, as well as when compared to a range of other central banks.

However, forecast errors over recent years have proved unusually large and persistent, particularly for inflation and wage growth. In a world with larger and more novel shocks hitting the economy, addressing lessons from forecast evaluations will likely require a response across a range of inputs to the MPC’s policy deliberations.

Some changes have already been made to the Bank’s models and processes in response to challenges of recent years.

To best support effective policymaking, the Bank’s models and processes should be adapted to lessons from real world developments in a timely manner. After a long period of relatively little variation in energy prices and policy rates, the recent inflationary episode provided new empirical evidence that the Bank can learn from. This has already led to a number of changes being implemented as Bank staff sought to incorporate some of the lessons in real time.

On models, a more sophisticated treatment of energy prices was added to the Bank’s central forecasting framework, COMPASS (Albuquerque et al (2025)), with changes also made to how energy-price conditioning paths are determined. A wider programme of continuous improvement to this model has also been established. Reflecting the latest Bank staff analysis, the speed of monetary transmission within the modelling toolkit used to support the MPC’s policy strategy discussions has also been made faster (Alati et al (2025)). And a number of other ad-hoc models have been developed to help the MPC explore specific mechanisms or ‘non-linearities’ that may have been more prominent than usual over recent years, but that are not fully captured within the Bank’s core machinery (Haskel, Martin and Brandt (2025) and Buckmann et al (2025)).

Given the increased role of supply factors as drivers of economic fluctuations over recent years, the frequency of Bank staff’s internal supply-side monitoring processes has also increased, compared to the previous annual ‘supply stocktake’ approach. And Bank staff have made more systematic use of scenario analysis to supplement the central projections in support of monetary policy deliberations (Dhami et al (2025)). This has enabled greater emphasis in the policy process on risks and uncertainty (Haberis et al (2025)).

Further changes are in progress, reflecting the recommendations of the Bernanke Review.

In 2024, the Bernanke Review provided several recommendations to improve the Bank’s forecasting and policymaking more generally (Bernanke (2024)). The Bank has embraced these recommendations (Lombardelli (2024)) and is implementing several changes to the MPC’s policymaking and communications framework, widening the range of inputs provided by staff to include more systematic analysis of risks, scenarios, and alternative policy paths, alongside the baseline forecast (Dhami et al (2025)). The Review further recommended an increased focus on forecast evaluation to help Bank staff and the MPC draw lessons from forecast errors closer to real time, which this Report is responding to.

Results from this evaluation point to the labour market and wage-price interactions as potential areas for further model development.

This Report has helped to identify further modelling avenues for Bank staff to explore, particularly around the Bank’s labour market modelling. The statistical analysis of the Bank’s forecasts has flagged labour market variables as a relative weak spot. More recently, the surprising persistence of inflation alongside elevated wage growth since the 2022 inflation surge has also underscored the importance of better modelling various potential interactions between labour market tightness, wage growth, CPI inflation and inflation expectations.

Alongside these issues, the Bernanke Review had highlighted some other potential areas of focus for the Bank’s modelling. These include a richer representation of the monetary transmission mechanism, and greater attention to supply-side elements and their role in the determination of inflation. These, and a range of other potential further areas for development, are being considered by Bank staff as part of a wider programme of modelling investments made under the Bank’s ongoing Monetary Policy Transformation project (Lombardelli (2024)).

There are inevitably limits to the extent to which model development can mitigate future forecast errors.

It is important to recognise, however, that there is no single model that can effectively encompass every potential variable or mechanism of relevance to monetary policy. And by their nature, economic shocks are unpredictable implying a degree of forecast error is inevitable. Notwithstanding this, it will be important for Bank staff to continue to develop and maintain complementary models that can be deployed as appropriate in the light of evolving economic circumstances. Forecast evaluation can help to inform this.

Particularly at times of elevated uncertainty, forecast evaluation can also help inform the selection and use of scenarios to explore risks as an important complement to central projections in informing the MPC’s policy discussions. And there will always remain an important role for judgement in combining the signals from a range of these models alongside other data and analysis when forming views about the economic outlook.

This Report is the culmination of material investment in the Bank’s data, technology and forecast evaluation methods, which will support continuous improvement.

The analytical framework for forecast evaluation laid out in this report is underpinned by a material investment in the Bank’s forecast evaluation toolkit. This toolkit has been made available for public use alongside the data underpinning this Report in the interests of openness and transparency (Abiry et al (2026)). The toolkit employs data science best practices to underpin a flexible and streamlined approach to monitoring forecast performance at the Bank of England going forward.